foWKE] ARCHEOLOGICAL INVESTIGATIONS 169 



gin has been removed ; in others the erosion has progressed to such an 

 extent that observations were possible at varying distances, to and 

 beyond the center. In every instance a monotonous uniformity of 

 appearance prevails from the top of the mound into the underlying 

 gravel. At no level is there a sign of a floor, fire bed, or other evi- 

 dence of human work; and no difference can be detected between the 

 earth upon which the mound rests and that on either side. Yet the 

 mounds are indubitably artificial. 



Exactly the same remarks apply to several mounds on the County 

 Farm, near Salem. A little creek and a drainage ditch have cut 

 away varying portions of them, and they merge insensibly into the 

 soil and gravel on either side. 



In further support of the theory that these mounds are the remains 

 of earth-covered houses, a few extracts relating to the area under 

 discussion will be given from Dr. Cyrus Thomas in the Twelfth An- 

 nual Report of the Bureau of Ethnology : 



Near " Beckwith's Fort," in Mississippi County, Missouri, are 

 (p. 189)- 



Low, flattish, circular mounds * * * [which] appear to belong to two 

 classes, those used for dwelling sites and those used for burial purposes, the 

 former being the higher and the color of the surface layer darker than that 

 of the other class. This darker color of the surface layer is probably due to 

 the fact that immediately below it are found fire-beds with burnt earth, char- 

 coal, ashes, and the bones of animals, (mostly split). There are seldom any 

 human skeletons or entire vessels of pottery in the mounds of this class though 

 the earth is tilled with fragments of broken vessels. 



In describing mound excavations in Crittenden County, Arkansas, 

 the explorer states (p. 227) : 



As an almost universal rule, after removing a foot or two of top soil, a 

 layer of burnt clay in a broken or fragmentary condition would be found, 

 sometimes with impressions of grass or twigs, which easily crumbled but 

 was often hard and stamped apparently with an implement made of split reeds 

 of comparatively large size. This layer Avas in places a foot thick and fre- 

 quently burned to a brick red or even to clinkers. 



Below this, at a depth of 3 to 5 feet from the surface, were more or less 

 ashes, and often 6 inches of charred grass, immediately covering skeletons. 

 The latter were found lying in all directions, some with the face up, others 

 with it down, and others on the side. With these were vessels of clay, in 

 some cases one, sometimes more. 



The positions of the skeletons in this mound would indicate that 

 while the inmates of the house were asleep the roof fell and killed 

 them. It was customary among some southern Indians to bury 

 the dead under the floors of the houses; but the text clearly shows 

 that these skeletons were lying on the floor. It would be supposed 

 from most reports, not only in the volume quoted, but from various 

 other, sources as well, that only the walls of these houses were plas- 

 tered with mud, the roof being of thatch alone. It seems to be over- 



