134 THE JOURNAL OF GEOLOGY. 



will be carried with considerable speed by steep glaciers to the 

 sea ; the localized erosion will be very energetic ; and we will 

 get the fiords and the lakes — while the eroded matter will only 

 be deposited before the glaciers as terminal moraines and terraces. 



As the ice shed will be rather near the coastward margin of 

 the inland ice, it follows that it is for a great part the enormous 

 snowfall on the great western glacier's own surface which is to be 

 transported. Hence these glaciers can follow more irregular 

 lines from the first, only deepening and widening the preexist- 

 ing valleys, and so we get the complicated fjord systems ; while, 

 on the continental side, where a more uniform vis a tergo pushes 

 the ice plate forward, the eroded depressions must be more regu- 

 lar. It is quite necessary to keep in view this great contrast 

 between erosion and deposition on the part of the continental 

 and the coast sides of the great ice sheets, respectively, in order 

 to be able to understand many of the complicated glacial phe- 

 nomena. 



We have seen that the deuteroglacial and especially the 

 epiglacial inland ice had its ice shed far to the east from the 

 watershed. Across this divide the ice must move and with 

 somewhat accelerated speed in the narrow defiles. There must 

 here originate passes or gaps (skar in Norwegian) across the 

 watershed. Of these, we really have very many in Norway with 

 a development in distinct relation to the distance between the 

 ice-shed and the watershed and to the greatness of the epiglacial 

 lakes and valleys on both sides. In not a few we have lakes 

 with outlets on either side. 



When the ice stream through such a gap came out to the 

 western edge of the high plateau, there resulted a sort of ice 

 cascade, which, like a waterfall, receded with rather great speed. 

 These receding icefalls evidently gave origin to most of our 

 fjord- valleys, sack-valleys, ctds de sac interior to the fjord heads. 

 Also these must generally be of deuteroglacial origin, as only 

 then the ice flowed in great streams across the watershed. 



The epiglacial stage of the deuteroglacial period must have 

 had a very abrupt termination. The glaciers retired from the 



