170 BUREAU OF AMERICAN' ETHNOLOGY f bull. 76 



looked that the tops of the houses would have even more need of 

 such protection than the sides. The marks indicating that the clay 

 was " stamped apparently with an implement made of split reeds " 

 are only the impressions of the reeds or saplings by which the clay 

 was supported; the "brick like" or "clinker like" condition of the 

 clay being due, of course, to the destruction of the house by fire. 



Adair, in his History of the Southern Indians, says they daub their 

 houses with tough mortar mixed with dry grass; that they build 

 winter or hot houses after the manner of Dutch ovens, covered with 

 clay. Again : 



They are lathed with cane and plastered with mud from bottom to top, 

 within and without, with a good covering of straw. 



This seems to mean that the entire building was plastered with 

 mud, and then covered with grass to shed the rainfall. 



In a mound in Arkansas County, Arkansas (Twelfth Ann. Kept. 

 Bur. Ethn., p. 231)— 



About 2 feet under the surface was a thiclv layer of burnt clay, Avhich prob- 

 ably formed the roof. In tracing out the circumference a hard clay floor was 

 found beneath,- and between the two several inches of ashes, but no skeletons. 

 There were a great many pieces of broken dishes so situated as to lead one to 

 believe they were on top of the house at the time it was burned. 



The fact that no skeletons or utensils were discovered on the floor 

 finds its most reasonable explanation in the supposition that the in- 

 mates, finding their abode to be unsafe, moved out and took their 

 possessions with them. This would account, also, for the absence 

 of such remains in similar mounds farther north. The abundance of 

 pottery fragments found in this case, and in many others, may mean 

 only that these were worked in as a part of the clay roofing. They 

 would be of some service in holding the clay in place in wet weather. 



It is quite probable that the continuous, though fragmentary, 

 layer of burned clay on the floor so often noted is due in part at least 

 to the material forming the roof. The walls would be more apt to 

 fall outward than inward, and would be more liable to crumble than 

 to fall as an intact mass. In fact, this is clearly shown by the 

 statement (p. 229) that in certain house sites in St. Francis County, 

 Arkansas, 



The edges are all higher and have a thicker layer of this [burned] material 

 than the inner areas. 



Further, in describing explorations of certain " hut rings " at 

 "Beckwith's Fort" in Mississippi County, Missouri (p. 187), the 

 report states that they are 



from 30 to 50 feet in diameter, measuring to the tops of their rims, which are 

 raised slightly above the natural level. The depth of the depression at the 

 center is from 2 to 3 feet. Near the center, somewhat covered with earth, 

 are usually found the baked earth, charcoal, and ashes of ancient fires, and 



