In geological text-books the explanation is generally given that pot-holes are formed by semipermanent moulins which occur where the glacier breaks over a sharp edge. Consequently there will be a semi-permanent crevasse through which water falls more or less vertically down on to the glacier bed.

In a recent issue of the Journal of Glaciology (Vol. 1, No. 9, p. 488–90) Dr. R. Streiff-Becker contends that pot-holes are not produced in this manner.Footnote * I shall not try to prove that the old hypothesis covers all cases—in fact, very much seems to indicate that pot-holes may be formed in different ways, even if we restrict ourselves to holes with a vertical axis—but I should like to draw attention to one particular group of pot-holes where I think the old explanation fits in better with observations than any other hypothesis.

The locality is a small, exposed ocean island west of Bergen, Norway, called Kvannholmen (parish of Herdla), near the very top of which (height 23 m.) there is a group of quite good potholes. The one illustrated in Fig. 1 (p. 20) is 3.20 m. deep with a diameter of 1 m. Fig. 2 (p. 20) shows the location on the island, and Fig. 3 (p. 24) the position in relation to the surrounding country. Comment is hardly necessary; it will be seen that the location of the pot-holes is very easily explained by the old theory, but is in hopeless antagonism to Streiff-Becker’s assumption. No sub-glacial stream could run across this island as long as there were excellent outlets just north and south of it—and besides, the main outlet goes north-west on the other side of the chain of islands.

Fig. 3. Kvannholmen and surroundings. Depth contour 100 m. indicated in the main map. The black area outside the frame of the lower inset map is the town of Bergen. Reproduced with permission from Topografisk kart over Norge, Herdla, edited by Norges geografiske Oppmåling

The assumption that these pot-holes have been formed during an earlier glaciation and have remained while the drainage pattern has changed fundamentally, must be rejected as totally improbable. In late-glacial times the island was submerged and emerged again. We must assume that the erosion that is seen in Fig. 1 (front part of pot-hole broken away), and which has cut another hole in half lengthwise, took place during this period.