234 BUREAU OF AMERICAN ETHNOLOGY [Bull. 182 



Some pits were hemispherical in cross section while others were 

 formed by a series of connecting pits, giving the whole a sort of amoebic 

 outline. Most pits had sloping sides and rounded bottoms. Those dug 

 by white man had vertical sides and either rounded or square bottoms. 



Throughout the area worked, which measured 110 feet along the 

 southern limits and 45 feet along the eastern limits, we located two 

 burial areas (fig. 53). Burial types have been listed as fully flexed, 

 partially flexed, bundle, extended, and what must have been an attempt 

 to cremate. The latter, exemplified in a single instance, concerned a 

 number of individuals who were submitted to the fire but 

 were only partially consimied. These remains were further reduced 

 in size by pounding and the whole heaped into a common pile and 

 covered over with midden material. Single and multiple burials also 

 occurred. 



Again we found a number of turtle carapaces associated with some 

 of the burials. Most of these were so placed that they rested upon 

 their natural base, which excludes their possible use as vessels to re- 

 tain food or other materials. Their position, in relation to the skeletal 

 remains, would strengthen our premise that they functioned solely 

 as indicators of clan affiliation or totemic badges. We got our first 

 inkling as to their status while excavating a portion of the Clarksville 

 site, MMcl4. Now we found this same situation occurring here. 



In our second burial area, we located the remains of several small 

 dogs, each in its own special grave. Most of the dogs were coiled up 

 into small bundles, with their extremities folded close to their bodies, 

 and placed into small round graves. Some of them were well pre- 

 served while others were badly disintegrated, and still others were 

 minus their skulls. 



In the eastern limits of our diggings we found the remains of a 

 fairly large rectangular structure. Feature 24 (fig. 54 and pi. 82). 

 This was represented by a well-packed floor. The corners were 

 rounded. Within its western limits were two square post molds. We 

 traced their origin up to the base of the present humus and determined 

 that they formed part of the garden fence put up by the occupants 

 of the old Tollifero house; hence they are of white man's origin. In 

 the western sector of this rectangular structure were three round post 

 molds. These were not found until the floor of the aboriginal struc- 

 ture was completely uncovered. No other molds were found either 

 inside of the structure or the area surrounding it. The floor of the 

 house now rests between 18 to 20 inches lower than the present ground 

 level. We still do not know what type of walls made up the sides or 

 anything about roof supports or the shape of the final roof. We have 

 supposed that wall and roof supports must have originated on the 

 shelf outside of the floor pit. Whether this was a religious structure 



