A survey of the Phytogeography of the Arctic American Archipelago 155 
a colitinuous icesheet (Russell, 1. c, p. 138). The air thus dried woiild have very 
little precipitation left for tlie plains behiiid the mountain range, neither would they 
get any from the always frozen Polar Sea to the north. The sigDs of glaciatioii 
detected on the asiatic side of the Bering Sea (Geikie, 1. c, p. 697, and Plate XIII) 
do not contend against this view, as a wide bay of deep water stretclied into the 
present Bering Sea from Kamtshatka north wards. The land occupying the area 
north of the »Aleutian Mountains», as well as the present northern Alaska and its 
now subtnarine northern continuation, had doubtless had a verv dry and eold chmate, 
hut so far south the summei- must have been warm enough to meh away the 
snow of the plains, and it seems highly probable that plauts have hved there during 
the whole glacial period mider coiiditions rather nearly approching those of the 
Arctic Archipelago at jiresent. The importance of this unglaciated area as a refuge 
and centre of reimmigrations seems, however, only to have occurred to a single 
author, viz., C. C. Adams, who writes (Biota, p. 62): »It also appears that some of 
the mountain biota of the Canadian Rockies were driven north into unglaciated 
Alaska as the ice spread from the Cordilleran centre of ice accumulation». He 
also speaks of this region as a refuge of the siberian types which had arrived 
when the land-connection with Asia was formed (p. 58), as well as of an earlier 
return of biota living south of the borderline of glaciation to the west thau to 
the east (p. 55). 
Farther northeast, over the northernmost extremity of the Continent, now 
forming the Archipelago, and over Greenland, a still more severe climate must have 
ruled, and in the firstnamed region it must have been extremely dry too, as tlie 
winds blowing there had to pass either over wide expauces of land, mostly high 
and icecovered, or over the icebound Polar Sea. This accounts for the almost 
entire absence of marks of glaciation in the Arctic Islands; the amount of preci- 
pitation, small as it is now, was still smaller then, and the summer north of the 
immense icesheet of America and in the vicinity of an almost ahvays and entirely 
icebound sea must have been very cold, even if the hypothesis of a general lowering 
of temperature is not resorted to. That any plants, even the most hardy, could 
live there under such conditions, is hardly to be thought of, and thus the geological 
evidence leads to the same conclusion as the study of the flora, namely that plant- 
hfe has been extinct in the Arctic Archipelago during the cliraax of the iceage. 
The same holds good for the icefree areas in northwestern Greenland, where the 
same canses alone prevented the forming of an icesheet. The marks of glaciation 
found by M'Millan in Melville Island (App. Arctic, p. 411, 461) certainly are formed 
at a låter period. 
The only part of the Arctic Archipelago that shows signs of a considerable 
former glaciation is southern Baffin Land. Its icesheet has, however, not been 
confluent with that of Labrador. Tliat the glaciation has here lasted up to a very 
late date, is generally supposed by geologist who have studied the region in question, 
and when Upham, Glac. lakes, p. 271, speaks of glaciation of northern lands after 
