FOWKE] AECHEOLOGICAL INVESTIGATIONS 171 



around these and beneath the rims [that is, the surrounding ring or embank- 

 ment] split bones and fresh-water shells. Often mingled with this refuse 

 material are rude stone implements and fragments of pottery. 



Note is made of 

 the similarity in the size, form, and general appearance of these depressions 

 and earthen rings to those of the earth lodges of the abandoned Mandan towns 

 along the Missouri River. 



It appears, too, that certain sites were occupied for long periods, 

 new houses being constructed when necessary. In describing 

 mounds in Poinsett County, Arkansas, the same writer says (p. 

 205) that 



The positions and relations of these beds * * * make it evident that 

 upon the site of one burned dwelling another was usually constructed, not in- 

 frequently a third, and sometimes even a fourth, the remains of each being 

 underlaid and usually overlaid in part by very dark, adhesive clay or muck. 



The peculiar black color of these beds is chiefly in consequence of the large 

 proportion of charcoal with which they are mixed, some of it doubtless the 

 fine particles of burned graSs and reed matting with which the cabins appear 

 to have been thatched. 



These layers of " very dark " material undoubtedly are remains 

 of mud from the adjacent swamps, which was mixed with or plas- 

 tered over the grass roofs. It is difficult to understand how they 

 could have become mixed after the burning. 



As showing the extent to which this prolonged occupancy was 

 carried, we are informed (p. 254) that in Coahoma County, Missis- 

 sippi, a mound was — 



oval and rounded on top, 210 feet long, 150 broad at the base, and 16 feet high. 

 This mound and several smaller ones near it are so nearly masses of fire beds, 

 burnt clay, fragments of stone and pottery, together with more or less charcoal 

 and ashes, as to indicate clearly that they are the sites of ancient dweUings 

 thus elevated by accumulation of material during long continued occupancy. 



In still other portions of the country besides those already men- 

 tioned are evidences of similar houses whose sites are now marked 

 by mounds. In southern Ohio, especially, records of excavations con- 

 tain numerous references to post holes under mounds both large and 

 small. In the case of the former, so far as we may judge from the 

 reports, the houses were destroyed before the mounds were built, and 

 it does not appear that they were ever covered with earth. In the 

 small, low, flat mounds, under which such holes existed, no thought 

 was taken that these may mark the position of posts used to support 

 a roof ; all mounds were explored with the idea that they were for 

 burial purposes, consequently no attention was paid to these features. 



The Mandan houses, as described by Lewis and Clark, Catlin, and 

 others, when fallen into ruins would leave exactly such mounds or 

 hut rings as those found in Missouri and Arkansas. 



