24 



IKIDTJOF NANSKN. 



M.-.\. Kl. 



---Tf°S&fc=s:--^^^ 



Fig. 6. Cirquc-gUiciers on the west coast of Cross Bay, Spitsbergen [from Nansen, 1920]. 



Fig. 7. Future picture of the coast in Fig. 6 [from Nansen, 1920]. 



would here move with consirably greater velocity along the fjords and 

 sounds then over the islands between them, and it would go on deepening 

 them at a very much greater rate than it could erode the surface of the 

 islands. The velocity of its movement along these channels may have gone 

 on increasing with increasing depth, and its erosive effect would increase 

 very much more. How then would it be possible that the inland ice could 

 have planed the surface of the islands and peninsulas to practically the 

 same level, while it has excavated the sounds between them to very 

 different depths? 



The inevitable effect of the erosion of the inland ice in this region 

 must be that, not considering its smoothing and rounding of the local rock 

 surfaces into roches moutonés, it will make the wliole lanrl surface much 

 more uneven than it was before. 



For this reason, according to my view, and as has also so clearly been 

 pointed out by Hogbom [19 13, p. 57], it is impossible that the deep sounds 

 and channels, traversing the strandfiat and dissecting it into its many 

 islands, can have been formed after this plane had been levelled. If the 

 sounds were deepened, to some considerable degree, by the glaciers of the 

 inland ice, this deepening process could not in that case have left the plane 

 between the sounds as undisturbed as it actually has. A glacial erosion, 

 to that degree selective, is not conceivable on a flat plain. 



If the sounds and channels have been formed chiefly by atmospheric 

 and fluvial erosion, then as Hogbom has pointed out, it becomes still 

 more absurd to think that they could have been eroded without the inter- 

 vening plain being dissected by deep valleys, sloping towards the floors 

 of these sounds. But this is not the case. 



