132 THE JOURNAL OF GEOLOGY. 



tudinal section in the best accordance with the probable distribu- 

 tion of the erosive power in a glacier. 



The rock matter which was scooped out by the epiglacial 

 glaciers was deposited partially as terminal moraines, but mostly 

 as terraces in the fjords. We therefore now find the epiglacial 

 lakes everywhere separated from the fjord heads by some kilo- 

 meters of terrace land. These great upper terraces mark the 

 level of the epiglacial sea. Their height differs very much in 

 the different districts, and it was first by the study of the corre- 

 sponding ancient sea beaches that I was able to find the correla- 

 tion between them. In western and northern Norway, these 

 ancient sea margins are very distinctly marked, often in the 

 living rock. As might be expected, in accordance with the 

 theory of isostasy of the earth's crust, these lines are now raised 

 towards the former center of the ice sheet, from which the max- 

 imum ice load was taken away. Nearest this point the epigla- 

 cial sea beach and terraces now reach 200 meters above the sea, 

 with a gradual fall towards the outer coast, where their height 

 is only some 20 meters or less. The gradient in the fall is much 

 greater than in Lake Agassiz, reaching i.2 meters (4') per kilo- 

 meter, against o.io meters (20"), but it is clearly the same 

 cause that has been at work in both cases. 



The height of the deuteroglacial ice sheets seems to have 

 been almost the same as in the proteroglacial — less than 2,000 

 meters maximum above the present sea level. 



The epiglacial terraces contain in some places banks of 

 shells with quite an arctic aspect. The deeper clay has Leda 

 artica as its leading fossil. The climate must have been very 

 like that at present in South Greenland, and the topographical 

 physiognomy must also have been very much the same, with a par- 

 tially alpine foreland (which constituted nunatak forms as in the 

 proteroglaciated epoch), and with glaciers at the heads of the 

 fjords, with great clay bars or deltas before them, and with 

 small floating icebergs to score out the strandlinjer — the sea 

 beaches in the rock. 



It is altogether probable that a meteorological map of Nor- 



