194 H. P. STEENSBY. 
Different forms of earth-houses occur however in North-America 
besides: a rectangular earth-house is also known from Northern Cali- 
fornia. There cannot, however, be any question of its being one of these 
forms which has been transferred to the Eskimo domain. The type 
is quite different. There is no doubt that JocHErson! is right when, 
on the basis of his investigations, he says “we may draw the conclusion’ 
that the underground houses of the Palæasiatic tribes bear more simi- 
larity to those of the Aleuts and Alaskan Eskimos than to the under- 
ground dwellings of the North-Western Indians.” 
In this way nothing remains but North-East Asia. We must assume 
that the rectangular earth-house is due to Asiatic influence 
exerted along the Pacific Coast. SARFERT? also assumes something 
of the same sort when he emphasizes some features which he thinks are 
to be found both with Aleuts and Alaskan Eskimo, and in the older houses 
which tormerly were in use with the Kamchadales, Koryaks and Gilyaks. 
Still further south we find semi-underground rectangular houses which 
from outside appear like round hills. В. Нттенсоск? has observed such on 
a small island Thikotan off the east coast of Jesso, and he supposes that 
similar houses formerly played a röle on Jesso, where he could see the 
hollows where the houses had been. 
There still remains to be mentioned that form of an apparently 
rectangular house which we have called the Mackenzie-house. 
Formerly I assumed that it was a further development of the earth- 
covered prairie-house, which, through Carrın’s and the PRINCE OF 
Wrep’s descriptions for example, we know so well from the Mandans. 
Thus I thought that the Mackenzie house was an original round house 
like this prairie-house, and that it had adjusted itself to this peculiar 
cross-form in the ground-plan. I supposed that this form of house might 
have struck root at the mouth of the Mackenzie on account of the great 
richness in drift-wood. Now, however, I must admit that THALBITZER 
is right when he, on the basis of SIRELIUS’S work, points to Asia as the 
place of origin for this form of house. 
It seems to me that in reality the history of the development and 
origin of the Mackenzie house is now manifest. The source of origin 
must be a form like that which, according to SCHRENCK, we find with 
the Gilyaks. It is an earth-covered, approximately square house with 
four posts in the centre which carry the obliquely placed poles or beams 
which form the walls and the roof. Inside, out from the three sides, 
rectangular platforms are placed, while the entrance is found in the 
fourth side. The fire place is in the middle of the floor. A form of house 
quite corresponding with this is found with the Koryaks*. Next it is 
1 JOCHELSON, р. 446. 
? SARFERT, pp. 61—62. 
3 Нитснсосн, р. 425, Fig. 65 and Plate LX XIX. 
+ JOCHELSON, p. 454. 
