Survey of Northeast Greenland. 421 
and have in the main to fall back upon conjectures. Jokelbugten 
and Nioghalvfjerdsfjorden in reality belong to the marginal zone of - 
the inland ice. Channels of melting water will consequently hardly 
be met with in that place, but on the other hand surface rivers as 
on the inland ice. Our limited knowledge of the conditions in Dan- 
marks Fjord, Independence Fjord and Hagens Fjord only shows us 
that in this place the ice is generally, as far as great portions are 
concerned, several years old; we know that during the summer a 
channel of open water forms along the shores, and it is a reasonable 
supposition that the interior of the fjords likewise become free of 
ice in normal summers; but we know nothing of the manner in 
Fig. 106. Elongated lakes on the fjord ice. Skærfjorden, August. 
In the background Rekvedgen. 
which the last stages of the ice-melting of the summer appear on 
the great mass of ice out on the fjords. 
In Frederick E. Hyde Fjord rather peculiar conditions seem to 
obtain. In the outer part of the fjord there had, as mentioned above, 
in 1907 accumulated an immense, fairly evenly distributed layer of 
snow, presumably several metres thick; but already 25 km inside 
the mouth the conditions were quite different. The snow here lay 
in downs which were one to two metres high and elongated in the 
direction of the fjord. In the hollows between the downs it is true 
that we generally met with snow, but very frequently we here found 
quite smooth ice, which showed that in this place there had in the 
previous autumn been a small lake on the ice, which was thus 
perennial. I am inclined to think that here we had to do with 
sikosak (for fuller details as to this see below), and that the “downs” 
owed their existence to well-developed “melting knolls” (see the 
