pLp N^f' IsY' McNARY RESERVOIR — SHINER 181 



showed a surface depression that was saucer shaped. The older houses 

 often had no surface indications at all. Most of the dwellings in the 

 region had traces of floors that could be picked up only near the 

 center of the houses. The floor level, when present, was found as a 

 charcoal and ash stain that did not extend to the edges of the house. 

 (Map 3.) 



House pit 3, at the Cold Springs site, showed a depression on the 

 surface of the ground and was carefully excavated (fig. 25). Al- 

 though there was little direct evidence of it, there were apparently two 

 or more occupations of the house site. 



During the excavation, no trace of a floor was found until a depth 

 of 3 feet was reached. At first it was believed that the deep floor be- 

 longed to the surface depression. However, the contour of the depres- 

 sion plotted before excavation did not match the contour of the deep 

 floor. The floor center proved to be more than 7 feet southwest of the 

 center of the depression. Furthermore, artifacts found just above 

 the deep floor were of a type associated with the earliest occupation of 

 the site ; they included two of the large basalt knives. It follows there- 

 fore, that the depression that plainly showed on the surface belonged 

 to a house that had been occupied at a later time. Another reason for 

 this conclusion was that the thick layer of discarded mussel shells that 

 blanketed this portion of the site was not present in the fill of the 

 house pit. This showed that the house had been occupied during or 

 after the period in which the shells accumulated. Since an average 

 of about 3 feet of midden had accumulated before the deposition of the 

 shells, the shell layer was stratigraphically later than the early arti- 

 fact complex found both in the midden and in the earliest house. In 

 brief, the house was occupied at a time when the earliest material cul- 

 ture at the site was being used. Later, the thick layer of shells accumu- 

 lated. The final occupation of the house came after the shells had 

 accumulated, and the occupants apparently cleaned the trash out of 

 the house pit before living in it. 



Time did not permit the full excavation of the deep floor but a good 

 profile was obtained. It showed the house to be bowl shaped, sloping 

 up toward the surface with no vertical walls. No features were ob- 

 served. This negative evidence that was to be repeated in nearly every 

 house excavated in the region was actually of some significance. It 

 limited the type of superstructure to something that may be visualized. 

 Obviously, there was no heavy structure of wood, for there were no 

 postholes. A roof entrance would hardly have been possible without 

 interior posts. A flat roof flush with the ground would not have left 

 more than 2 or 3 feet of space under it, even in the center of the house. 

 "Wliat is indicated here is light framework, perhaps of small poles, 

 covered with mats. Mats are mentioned in all of the early travelers' 



