712 Holsť’s Studies in Glacial Geology. 
are, among others, first, that the water-divides on the inland-ice 
shall be sufficiently far apart so that the water supply may be 
sufficiently great to form larger rivers, and, secondly, that the ice 
shall be sufficiently free from crevasses, which would otherwise 
drain off the water beneath the ice instead of on its surface. 
It is then evident that, in Sweden, the broad valleys and low- 
lands with gently rising sides must have offered particularly favor- 
able conditions for the formation of vast gravel-osar, whilst such 
osar can occur only as local formations of smaller dimensions in 
the mountain regions of the country. It has long been supposed 
that gravel-osar were entirely absent from Norway. This is, how- 
ever, not altogether true, but they are of rare occurrence, which 
fact fully agrees with the above theory. On the plains of the 
extreme south of Sweden, as also of Denmark and Germany, the 
absence of large drainage basins has hindered the formation of 
greater osar, although they are not altogether absent from Skane, 
and equivalent formations have been observed by Dr. Holst at 
Neustadt-Eberswalde. It is equally evident, that the topograph- 
ical conditions in those parts of southern Greenland above described 
(page 705) do not admit of any formation of larger osar. In 
a country so broken and mountainous, the inland ice must be full 
of cracks, preventing the water from gathering to any great extent 
over its surface. Such cracks do not necessarily exist in a moving 
inland ice, and Dr. Holst mentions a smaller tract of ice between 
Tasek Atdlek and Kangarassuk, which was entirely free from 
cracks, and, as a consequence, was covered with water, which gath- 
ered into a channel five feet wide and five feet deep, in one place 
separating into two branches, enclosing an island of ice, before it 
finally rushed into a jékel-well. Also Nordenskiöld and the Dan- 
ish explorers of the inland-ice met with water flooding its surface. 
If the above-given reasons for the absence of gravel-osar from 
the mountainous part of Greenland are correct, there could have 
been nothing to prevent such osar from forming in the less broken 
tracts, f. i. the district of Holsteinborg. Dr. Holst found no oppo 
tunity of visiting that district, but after returning home he learnt 
from A. Kornerup’s report of his travels in 1879 (published ™ 
1881) that he had found in the Arsalik valley, N. E. of Holstein- 
borg, a typical gravel-ose about four miles long, parallel to the 
