198 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179 



(east) of the McNary Dam (map 2). The terrace is on a semide- 

 tached island on the north (Washington) side of the Colmnbia River 

 and is nearly opposite the Hat Creek site. The local physiography is 

 almost identical with that of Techumtas Island — sandy loess soil, 

 steep shingle beach, and some shifting sand dunes. 



ARCHITECTUEB 



The village itself was approximately 300 yards long, parallel to the 

 river, and 35 yards wide. It consisted of 182 houses or house depres- 

 sions. Since only a few were excavated, almost nothing can be said 

 about the relative ages of the houses or the duration of the occupa- 

 tion. The evidence tends to indicate that it did not extend over a long 

 period of time, but that evidence is not conclusive. 



The house pits of various sizes were scattered along the terrace in 

 no particular pattern. No streets or significant groupings of houses 

 could be detected, although some information might be forthcoming 

 if the ages of most of the houses were known. (Map 5.) 



Six of the house pits at 45-BN-53 were excavated or trenched. 

 Houses Nos. 11, 49, 70, and 168 were completely excavated, while Nos. 

 52 and 169 v/ere only trenched. The results of the excavations were 

 nearly identical to those at Cold Springs and Techumtas Island. Oc- 

 cupation levels were found, but no postholes or traces of superstruc- 

 ture could be detected. East-west diameters were generally greater 

 than north-south diameters, making the long axis in line with the 

 prevailing wind. Median diameters were north-south, 16 feet, and 

 east-west, 19 feet. These are smaller than the averages for the houses 

 at Techumtas Island and Cold Springs. 



Generally, there were several occupational levels in each house, 

 and apparently the houses were not cleaned out from one occupation 

 to the next. Fireplaces were usually in the center of the saucer- 

 shaped floors. They were informal fireplaces, a few rocks and a 

 concentration of charcoal and ash. Entrances were not detected, and 

 the only trace of building material was in the form of mat fragments 

 found on the floor. Occasionally pairs of houses were side by side so 

 as to suggest contemporaneity, but the poor preservation of floors and 

 structural details prevents their reconstruction. 



MATERIAL CULTURE 



Most of the work at 45-BN-53 was concerned with the excavation 

 of house pits. About 12 working days with a crew of 12 men were 

 spent at the site, so the list of materials recovered is small. Of the 

 209 artifacts recovered, 129 were of the large coarse type such as 

 hammers, choppers, or net weights. The others were so similar to 

 the materials from Techumtas Island (35-UM-17) that there is little 



