It has for some time been obvious that far-reaching results could be obtained by penetrating right through a glacier to its bed. The expense however has always been prohibitive. Recently however Dr. H. Carol, of Zürich, had the ingenious but simple idea of going down through a glacier by means of a glacier pot-hole and has described his experiences in Les Alpes, 1945 (pp. 180–84).

The typical U-shaped cross-section of glaciated valleys is well known but little understood. If a climatic change results in a glacier occupying a deep river valley, and some erosion is assumed to occur wherever the ice makes contact with the bedrock, then a partial change from the “V” to “U” section is to be expected. A more potent agency in such a change is probably the sapping of the valley sides by the alternate freezing and thawing of melt-water flowing down to, and under, the glacier. This has been briefly suggested by de Martonne and worked out somewhat more fully in the case of cirques1