With regard to the mechanism of vertical erosion by glaciers the Horunger-Fanaråken area of central Norway proved instructive. Ice derived both from the mountains and plateau once converged on the deep glacial valley plunging down past Turtagrø into Fortundalen and the Sogne Fjord. The evidence here presented seemed to confirm the suggestion long held in many quartersFootnote ² that whereas relatively thin ice masses do little vertical erosion, as they thicken the vigour of down-cutting greatly increases. Such a process of over-deepening would clearly be a cumulative one provided the surface gradient were adequate. This would help to account for steps in the long profiles, and “U-in-U” forms in the cross profiles of valleys without the necessity of invoking separate glaciations, though the latter might well assist in the process.Footnote ³

A “U-in-U” section seems to be in active formation to-day in the valley occupied by the upper Aletsch Glacier, the glacier sculpturing the lower “U,” and bench-ice, hanging well above the main glacier, sapping the higher mountain sides and so forming the upper “U.”

But heavily glaciated valleys can have a markedly “V” cross-section. The Rhone Valley below Gletsch is of the steep narrow “V” form although ice several thousand feet thick must have passed through it in the Quaternary Ice Age. The valley below Turtagrø is another example. The best “V” section I have yet seen—still occupied by ice—breaches the rocky head-wall of the U-shaped Stadardalur in S.E. Iceland (see 1:50,000 map, Kalfafellstadur N.A.).

These “V” sections always occur when the valley gradient is steep and can frequently be recognized in glaciated valleys in Britain, as in the upper part of the Tal-y-llyn valley due east of Cader Idris, and in the Lake District.Footnote ¹

Though we might not accept Garwood's explanation of these anomalous features he deserves full credit for focusing our attention on them. They are mostly not post-glacial in origin, yet their likeness to deep river-cut valleys tempts one to attribute them to the action of streams. Might they not be due in a large measure to melt-water streams, often flowing violently under great hydrostatic pressure, when the glaciers were still present? Recent work by Ahlmann, Seligman and others on the temperature regime within glaciers would seem to have established that such streams usually originate far back in the névé regions, and of this there is much direct evidence. What is doubtful is the depth at which such streams could maintain their channels open against plastic collapse of the ice, but this might well be several hundred feet.

Glaciers are generally assumed to be relatively thin where they break into a complex of crevasses and seracs and flow down a steep stretch of their valleys. Their capacity to erode is assumed to be much less at such points than in the deeply filled basins above and below, whereas with streams it is quite the reverse. Thus towards the end of a glacier, on steep stretches, the sub-glacial streams might do a substantial proportion of the valley cutting. In view of the countless major and minor climatic cycles within the Quaternary Ice Age, considerable stretches of the Alpine and other glaciated valleys must at one time or another have been occupied by the relatively thin terminal portions of a valley glacier, thus giving full opportunity for this form of erosion.

Recently exposed steps towards the heads of glaciated valleys in Norway and elsewhere are frequently not scarred by water-worn notches. This seems to be due to the fact that the steps are so broad and smooth that the streams, when under the ice, were not allowed to remain for any length of time in one position. With narrower valleys, as at Grindelwald, the chances of a sub-glacial stream incising itself sufficiently to become locked in one position are greater and so gorges and V notches would be more likely to develop.

The fact that steps are usually more abrupt and less notched towards the head of glaciated valleys than farther downstream, might lend support to this hypothesis. But whether such assumptions prove right or wrong it seems evident that great volumes of ice can, in certain circumstances, flow through V-sectioned valleys without turning them into the somewhat over-emphasized “U” shape.