Vol. 65.] GLACIAL EROSION IN NORTH WALES. 311 



the protecting local glaciers will have retreated. Hence under the 

 theory of the protective action of glaciers, it must be assumed that 

 all the rock-steps in the valley-floors were made during the phase 

 of increasing glacial conditions, unless indeed it be assumed that, 

 during retreat, pauses occurred at the same points as during the 

 advance, and that at each pause the land resumed the altitude which 

 it had in the corresponding stage of glacial growth ; or that no pauses 

 occurred during the phase of retreat. The consequences of the 

 latter supposition are illustrated in fig. 20 (p. 317), where the 

 disappearance of the glaciers of fig. 12 reveals the valley-floor steps 

 previously developed. 



XIX. Special Features oe the Theory oe 

 Glacial Protection. 



Certain special consequences of the theory of glaciers as protec- 

 tive agencies need mention, particularly as applied to pre-Glacially 

 subdued mountains. 



(a) Protective glaciers occupying the upper parts of mature 

 valleys in a group of monadnocks cannot produce cliffs by under- 

 cutting the valley-head slopes; the glaciers can only preserve the 

 moderate slopes of the pre-Glacial valley-heads without significant 

 change. The mountain-tops, in so far as they rise above the pro- 

 tective neve-and-ice cover, must be reduced to very dull shape 

 during the successive episodes of glacial growth. 



(6) Even if a glacier be protective, the water-stream that is 

 given forth by the ice would cut back a narrow slit in the steep 

 face of the rock-step down which it cascades below the glacier-end. 

 The streams that are to-day rushing down such rock-steps in 

 glaciated mountains are actively at work trenching their courses 

 in cleft-like gorges, but they have as yet made very little progress ; 

 had the streams been there during all the long interval of time 

 required, not only for the deepening, but also for the widening of 

 the lower valleys, one would expect that the stream-gorges would 

 have attained a great length and depth. 



(c) The number of rock-steps should correspond in each of the 

 several radiating valleys of a mountain-group like Snowdon ; for 

 each step is only the local result of a well-maintained climatic 

 episode which must have been of essentially uniform value all 

 around the mountain-group. Furthermore, the height and spacing 

 of the corresponding steps should be systematically related, for 

 each set of steps is the result of a single climatic change and of 

 a single uplift. 



(d) It may be noted that the necessary postulate of increasing 

 uplift with increasing glaciation runs counter to the evidence which 

 in many regions associates increasing glaciation with depression. 



(e) Each enlarged portion of a valley down-stream from an ice- 

 covered rock-step should be of normal pattern ; the several side- 

 valleys should be well opened if the main valley is well opened ; 



t2 



