Vol. 65.] GLACIAL EROSION IN NORTH WALES. 323 



have modified them ; for, in these valley-heads, the first-formed 

 protective glaciers must have had their seats. But it is not 

 possible to reconcile the cwms with any reasonable pre-Glacial 

 forms. If the pre-Glacial valleys had heads as steep as the heads 

 of the existing cwms, then the pre-Glacial valley-floors, near the 

 valley-heads, must have been narrow ; but the cwm-floors are 

 broad. If the pre-Glacial valley-floors near their heads were as- 

 broad as the cwm-floors, then the valley-heads must have been 

 weathered back to moderate declivity ; but the head-slopes of the 

 cwms fully deserve to be called cliffs. So far as I have read tho 

 explanations of those who accept the theory of glacial protection, 

 there is no indication of a way of escape from this quandary ; 

 indeed, the quandary is not always recognized as such, because the 

 theory of glacial protection has not been consciously and carefully 

 carried out to the detailed consideration of its consequences. But 

 there are other difficulties. Let it be accepted for the moment that 

 the pre-Glacial valley-heads had essentially the forms of the cwms 

 of to-day, and that when they first came to be occupied by glaciers 

 the lower end of each glacier stood, as required by the ice protec- 

 tion theory, where the top of the cwm front step is now seen ; then,, 

 as has already been shown, the upper margin of the protective ice- 

 and-neve mass must have stood at the line now marked by the 

 sudden change of slope from cwm-head cliffs to dome-like summit- 

 slopes. As has already been pointed out, this requires a systematic 

 relation between the height of the cwm-cliffs and the forward 

 distance to the front rock-step. The actual location of the front 

 rock-step in each valley-head cwm, in relation to the valley-head 

 cliffs around Snowdon, does not appear to be such as to satisfy this 

 requirement. If any of one of the four great cwms that head in 

 against the summit of Snowdon were filled up to the cliff-top with 

 ice and neve, the resulting glacier would, as well as I can estimate 

 it, extend distinctly beyond the rock-step at the cwm front. 



Another point : — Some of the sharp ridges or cribs between the 

 Snowdon cwms are strongly serrated ; they are typical Alpine aretes. 

 in everything but altitude. Part of such a ridge is shown in fig. 22 

 (p. 324), a sketch of part of the spur (Crib Goch) that runs eastwards 

 from the northern part of the Snowdon mass. Another ridge, fig. 23,. 

 running south-eastwards from Snowdon to Lliwedd, is also sharp 

 and somewhat serrate, but more unsymmetrical, as is often the case 

 in glaciated mountains : the sunny slope being less steep than the 

 shady slope. Some part of the form of such serrated ridges is due 

 to post-Glacial weathering : the rest must, under the ice-protection 

 theory, be referred to pre-Glacial weathering, for, in the absence of 

 any sudden change of slope in the face of the cwm-head cliffs, we 

 are constrained (as has already been shown) to suppose that the 

 first glaciers which occupied the cwms filled them at least as high 

 as the present cliff-tops, and thus preserved the cliffs essentially 

 unchanged. But to suppose that the pre-Glacial spurs of Snowdon 

 were so narrow, sharp, and steep-sided as the present cribs, is to- 



