390 



BUREAU OF AMERICAN ETHNOLOGY 



[Bull. 137 



entirely with earth — between the pillars and the extreme exterior, a space of 

 several feet, are seats of mats, like those of the sheds. The manner of con- 

 structing the roof is very remarkable for Indian work. Upon the alternate 

 couples of the twelve pillars are first placed horizontal pieces — then upon the 

 ends of these are placed other horizontal pieces between tlie other couples of 

 pillars then another series of horizontal pieces resting upon the second set, 

 but drawn within towards the centre of the circle a few inches. Upon these 

 again are other pieces still more drawn in. 



There are four tiers of horizontal pieces thus placed upon each other. [Fig. 

 1.] a. 6. c. d. are four of the twelve pillars ; pieces are first laid upon a. b. and 

 upon c. d. then a piece upon these and between &. c, etc. These horizontal pieces 

 are strongly bound together by leather thongs of green hide ; it is evident they are 

 of the nature of an arch. They are only carried up to the number of four suffi- 

 cient for giving a direction and a foundation for the rafters, which are laid upon 

 these extending up to a point in one direction and in the other direction over 

 outside to near the ground. The rafters are strongly bound by thongs and 

 covered with ordinary rived boards for shingles. There is but one small entrance 

 to the house which is next towards the angle of the square adjacent to which 

 the round house stands. (Hitchcock, 1930, pp. 114-115.) 



Figure 1. — Structural detail of the roof of a Creek tcokofa from beneath (after Hitchcock). 



The hot house described by Swan, evidently that belonging to the 

 Upper Creek town of Otciapofa, was 25 feet in diameter, and 25 in 

 height, the outer wall being 6 feet high. There was but a single bed 

 running round the wall inside. The house was covered with bark like 

 the hot house described by Hawkins, and seems to have been the samo 

 as that in all other essentials. 



Bartram has not left us a minute description of any of the Creek 

 tcokofas, but he tells us that the one attached to the Upper Creek 

 town of Atasi was identical in construction with a Cherokee council 

 house, which he had dwelt upon earlier in his Travels, except that it 

 was much larger — larger he seems to imply than any other known to 

 him. That it was larger than any we have considered is indicated by 

 the fact that the roof was supported by a central post and in addition 

 3 concentric rows of wooden pillars (Bartram, 1791, pp. 452-455) . In 

 the plan of a rotunda given in his Observations on the Creek and 

 Cherokee Indians (fig. 2) (Bartram, 1909, p. 64) , he evidently has this 

 rotunda in mind. This shows 3 beds, or tiers of beds around the wall, 



