420 BUREAtJ OF AMERICAN ETHNOLOGY [Bull. 187 



mats ( Joutel in Margry, 1875-86, vol. 3, p. 345 ; Swanton, 1942, pp. 

 148-154). 



This was, in short, a house without eaves, resembling in that respect 

 most of the Algonquian houses and the upper portions of the lower 

 Mississippi houses. The walled house seems not to have been described 

 by any early writer, but is figured in the sketch above mentioned 

 (Bolton, 1915, frontispiece). This, as Harrington well remarks, was 

 evidently constructed like the lower Mississippi houses, i. e., like the 

 round houses of the Yazoo, Tunica, and Acolapissa rather than those 

 of the Natchez. During his explorations in the Caddo country, Har- 

 rington found "fragments of these Avattle-and-daub walls, acciden- 

 tally preserved by burning, which turned the clay into terracotta . . . 

 on most of the sites explored" (Harrington, 1920, p. 252). 



It is also interesting to note that Harrington (1920, pp. 256-258) 

 found in the same Caddo country numerous sites of earth lodges like 

 those used in historic times by the Pawnee, Mandan, and other north- 

 ern tribes. And just as the Wichita grass house may be described as 

 the lower Mississippi dwelling with the roof structure carried down 

 to the ground, so the earth lodge may be defined as the same with the 

 wall structure carried to the top. But in the lower Mississippi struc- 

 ture a mud covering was generally extended over the roof under the 

 grass and mats so that it is perhaps nearer the earth lodge than the 

 Wichita building. However, the relation between all three seems 

 pretty clear. Larger houses of the two first-mentioned types were 

 employed as town houses or ceremonial houses, one of them in par- 

 ticular being devoted to the uses of the great chenesi or high priest of 

 the Hasinai Confederation. Such buildings were usually placed upon 

 mounds. In 1916 Alanson Skinner explored a square house site in 

 southwestern Arkansas which he regarded, no doubt rightly, as the 

 house of a chief (Harrington, 1920, pp. 291-297). Although this was 

 in the country later occupied by the Caddo, it is highly probable that 

 it belonged to some tribe culturally connected with the Natchez and 

 Taensa. 



Outside arbors were used by the Caddo and probably by most of the 

 other Southeastern tribes, as we find them mentioned in Louisiana, 

 Florida, and the Carolinas, and the descendants of many of these tribes 

 make use of them today in Oklahoma. During summer, fires for 

 cooking were often made out of doors (Dumont, 1753, vol. 1, p. 144; 

 Swanton, 1911, p. 59). 



Adair (1775, p. 418) mentions a partition in the summer house of 

 the Chickasaw of his time, and this is important, for from it prob- 

 ably evolved the later house of both Indians and whites which was 

 made either with a central partition or by setting two small 1-room 

 houses end to end, usually with a platform between them. There is 

 reason to think that this type was of purely Indian origin. 



