An Anthropogeographical Study of the Origin of the Eskimo Culture. 195 
refound at Kotzebue Sound on the coast of Alaska, and finally in the 
Mackenzie district! (cf. figs. 5 and 6 on the plate). 
When these northern regions are reached, there are two facts, how- 
ever, which assert themselves. Firstly the wood-fire in the middle of the 
floor of the house has to be relinquished and replaced by the blubber- 
lamp, secondly the cold becomes more perceptible. These two facts en- 
force the avoidance of any superfluous space which would make the 
warming up of the house more difficult. The problem is solved by filling 
up the platformless space between the platforms, the outer walls being 
drawn in to the platforms themselves, which thereby come to lie in 
niches. Finally the construction is thus so altered that the roof becomes 
flatter and lower, and the space becomes restricted by this also. 
By these means a house is attained which can comparatively easily 
be warmed up, and which can nevertheless hold a relatively large number 
of inhabitants. The geographical distribution of this house in America 
enters like a wedge into the West Eskimo domain from Kotzebue Sound 
to the Mackenzie region (in this connection one must remember that 
the most used route of communication from Kotzebue Sound to the 
Mackenzie does not go round the coast past Point Barrow but straight 
across country along Colville River to the coast of the Arctic), and this 
condition of distribution may perhaps indicate that this form of house 
has arrived at a later date than the rectangular house. Like the latter, 
however, one must suppose that it originates from Asia, and that it has 
been introduced along the coast route; but it seems that not until the 
Mackenzie-domain is reached do we find it finally adapted to the con- 
ditions of Arctic winter dwelling. This explanation is more probable 
than SARFERT's assumption that the whole development from an ordi- 
nary rectangular house into a Mackenzie house must have taken place 
in America amongst the Eskimo. To this must be added that SIRELIUS 
has now shown how the Gilyak earth-house, which in my opinion repre- 
sents the primitive form of the Mackenzie house, has eventuated from the 
conical tent, in that this, as a support for the roof, is provided with four, 
or sometimes more, inside wooden pillars. So far, therefore, the Mackenzie 
house originates from a circular building; but the point of origin is in 
Аза and not in America. Thus I agree with THALBITZER?” as regards 
the possibility of the Mackenzie house finally being traced back to the 
Siberian “earth tent”; but I do not agree with him when he assumes 
that it is one of the most ancient Eskimo domiciliary forms. Necessarily 
it must have found its way into Alaska at a secondary period, no matter 
whether it came later than the rectangular house, or even if it perhaps 
came somewhat before this. 
In this way the forms of the Eskimo houses which have been 
1 Cf. SARFERT, pp. 56 sqq. 
2 № 0. @. Vol. 39, р. 361. 
13* 
