and became corrie glaciers. The process of erosion begun by the snow was continued by the ice, but on different lines. The walls and sides were snapped back and the basinis were enlarged till the glaciers they held became small ones of the valley type. As this went on, the floor was eroded deeper and deeper at the head, while little erosion went on near the lip, so that when the ice disappeared the hollow left behind was usually occupied by a lake. This seems to be the ordinary course of the development of a corrie glacier and the hollow usually associated with it, according to the most advanced school of physiographers; but there is a weighty body of opinion totally opposed to the idea that corrie glaciers are potent agents in modifying landscapes. Although the present writer is somewhat chary of expressing a dogmatic opinion on a subject which has lead to so much controversy, his experience in the glaciated districts of this country, notably in the Sounds region, leads him to think that the course of development outlined above accounts best for the phenomena that occur. By granting the capacity of glaciers of this type to sap back their-heads one can explain the formation of the jagged and razor-backed ridges which so frequently separate the head of one tributary glacier from that of its neighbour across a divide. The usual sequence of events in such a case appears to be as follows: First there is a ridge, more or less rounded, with two shell-like hollows containing snow on either side. In process of time a corrie glacier forms. Then sapping goes on, and the divide becomes narrower and narrower; then it is a mere wall, and finally this collapses and a saddle results. At the head of the Rakaia there are existent glaciers which show all stages of this development and furnish some idea of the modification which results from their action on mountain-ridges; but in those parts from which the glaciers have retreated the landscape-forms resulting from this action can be readily noted and studied in detail. We see here all the stages from the shell-like hollows, through the razor-backed ridges, to the final “pass” form. The latter are usually U-shaped in cross-section, but they tend to become parabolic by the accumulation on their floors of detritus shed from the walls. It seems possible that the isolated ridges which so frequently occur in all the great valleys of Canterbury owe their dissection to this process, especially as the main valleys run across the strike and the cross valleys are developed in the soft beds parallel with it, these beds furnishing an opportunity for snow to weather rapidly the shell-like hollow where thicker and more persistent drifts can gather. When the lowering of the divide has been accomplished by this action the main glaciers occasionally pass through them, especially if they are on the increase; they are cut down further still by the usual methods of glacier erosion. The occurrence of parallel or subparallel ridges in the mountainous district of Canterbury are of such a frequent occurrence, and such an important feature of the landscape, that their formation appears to be connected with the former glaciation, and the explanation I have given seems to me to be the most satisfactory way of accounting for their existence. Note.—Since writing the above I have seen an interesting paper by Professor W. H. Hobbs, of Michigan University, published in the Geographical Journal for February, 1910, which emphasizes the important effect produced on the mountain topography by the sapping-back of the walls of corrie glaciers. Most of the landscape-features mentioned by Professor Hobbs are exemplified in that part of the mountain district of