200 H. P. STEENSBY. 
America: on the prairie for example. It has not been possible to transfer 
such an earth-house to the tundra and the coasts of the Archipelago 
without more ado, and this for several reasons, of which the most im- 
portant is the loose consistency of the soil compared with the solid sods 
of the prairie. Added to this is the great scarcity of wood, and, also, 
the fact that the ground is almost always frozen and impenetrable. On 
the prairie it is possible to raise a mound of turf, but this cannot be 
done on the coast of the Arctic Ocean. There the soil can only be used 
to cover a sloping surface, where it can lie firmly. If one wishes to have 
a firmer mound, or wall, it must be of stone, or at any rate must have 
a firm core of stone. 
But then, on the other hand, there is the snow, which on the tundra 
can take the place of the firm sod of the prairie. In reality the snow 
is an ideal material to use. One can easily dig out a round hollow in 
the snow, and with the snow build a wall round the hollow, in a similar 
way as one sets up a wall of turf round the hollow of the house on the 
prairie?. Nowadays the opening above is closed with a vault. Origin- 
ally, of course, there must have been times when this was not known, so 
the snow house was covered with skins in a similar way as it still is 
towards spring, when there is a risk of the snow-vault melting and 
collapsing, a practica still in common not on the Alaskan coast south 
of Point Barrow through the winter. 
I have tried also to correlate the snow house with the North Ameri- 
can round earth-house, because, inter alia, a round earth-house and, 
besides this, the dome-shaped tent evidently represent a couple of the 
oldest forms of dwellings, not only in these districts, but also in Northern 
Asia. Probably they must both be reckoned as typical adjuncts to the 
oldest Palæasiatic-American culture in the northern regions with the cold 
winter. 
In its oldest form, such as it must be supposed to have been here, 
the snow house, therefore, must be apprehended as a snow-tent — or 
a parallel with the earth-tent. Not until later — perhaps even relatively 
late — was the art of making an arch acquired. But, when this was 
acquired, the snow house gradually became so greatly improved that 
finally it was able almost entirely to supersede other forms of winter 
houses from the central districts. 
How the art of building an arch was acquired is a difficult question. 
There are two possible explanations: either the Eskimo did not learn 
1 Cf. STEFANSSON, I, p. 346. 
2 A geographical condition must also be emphasized, which stipulates that it 
is on the tundra and the sea ice that the snow house has arisen, and not 
in the forest. The fact of the matter is that in the forest the snow falls 
evenly like a homogeneous but comparatively thin and loose layer. In the 
open districts north of the forest boundary, however, large masses of snow 
collect in places which is a contingency with the building of snow houses. 
