THE GEOLOGY OF THE FZ ROE ISLANDS. 261 
that cloaked the valley-slopes,—in the bevelling and rounding-off of basalt-cliffs 
and ledges,—in the deepening and widening of cirques and valleys, the levelling 
of valley-bottoms, the reduction of low-lying water-partings or cols, and the 
excavation of rock-basins. It is in accordance with all that we know of the 
glacial phenomena of Scotland, Norway, and Switzerland, that the material 
produced by glacial abrasion should not have collected in any great quantity 
under the ice. The gradients, as a rule, are too steep, and comparatively little 
till, therefore, was accumulated in the valleys, the great mass having doubtless 
been rolled out to sea, and spread over the sea-bottom—part of it probably 
being carried away by the icebergs that broke off from the terminal front of the 
ice-sheet. 
10. Direction of Ice-flow and Extent of Ice-sheet.—The undulating lines which 
I have described as indicating the primeval water-shed of the old table-land 
also mark out the centres from which the mer de glace flowed. The long sound 
that separates Osterée from Stromée brimmed with ice which flowed in two 
directions. North of Nordskaale the movement was northerly, while south of 
the shallow part of that sound the ice held on a southerly course. So thick 
was the mer de glace that its upper strata flowed across Kollefiord and 
Kalbaksfiord, and even overwhelmed Naalsée. All Stromée south of Kalbaks- 
fiord appears to have been smothered in ice flowing to south-east, and I believe 
that the direction of the flow in Hestéefiord was the same. Sanddée was also 
overwhelmed, nor can there be any doubt that the ice which covered it was 
continuous with that which cloaked all the islands to the north. Of these last 
it is enough to say that so far. as our observations went, they appear to have 
been glaciated invariably in the direction of the principal fiords. A glance at 
the map, indeed, will show that the ice streamed outwards everywhere from the 
dominant high grounds. 
The’ appearances in Suderée are extremely interesting, inasmuch as they 
prove that the ice of that island, although for the most part strictly local, and 
flowing east and west from the chief heights, was yet connected with the mer de 
glace of the northern islands. This is shown by the fact that glacier-ice has 
passed up Qvalbéefiord across the island to the west coast. That ice could not 
have come from Suderée itself, but from a thick glacier-mass occupying the bed 
of the sea between Suderée and Sandée. In other words, the ice-sheet of the 
northern islands must have coalesced with that which covered Suderée. This is 
not astonishing when we remember that the mer de glace of Stromée must have 
reached a thickness, at what is now the sea-level, of 1500 or 1600 feet. So 
thick a mass could not have floated off in the shallow water (20 to 40 fathoms) 
that separates Suderée from the northern islands. But the ice that streamed 
south from Stromée and Osterée was thicker even than these depths imply. 
To the north of Naalsde we get a depth of 120 fathoms which must have been 
