170 BUREAU OF AMERICAN ETHNOLOGY [BuU. 179 



pletely excavated ; at site 45-BN-53, a pit house village was partially 

 excavated ; and at site 35-UM-8, an early occupation site was tested. 

 In 1949, site 35-UM-8 received additional testing, and at site 35- 

 IIM-7 (Cold Springs), a pit house and midden area were partially 

 excavated. In 1950, two pit house villages were partially excavated. 

 These were 45-BN-6 and 35-UM-17 (Techumtas Island). In 

 1951 an early site 35-UM-5 (Hat Creek) and a late site 45-WW-6 

 (Wallula) were each partially excavated. In 1952 further testing 

 was done at sites 35-UM-7 and 35-UM-8. Sites 35-UM-3, 35-UM- 

 10, and 45-BN-55 were scenes of minor excavations. 



After each summer in the field, the excavated materials and data 

 were taken to the laboratory in Eugene for study. As soon as pos- 

 sible, reports were written on the results of each excavation. These 

 reports were little more than a factual account of the material recov- 

 ered with an attempt to relate it to other cultures in time and space. 

 Because of their limited distribution, illustrations were kept at a 

 minimum. 



Techniques in the field were kept as flexible as possible and only a 

 few were standardized. One of these was the grid system. A site 

 was staked with sections of reinforcing iron with wooden caps which 

 could be lettered. A single row of stakes 10 feet apart were set across 

 a site in a north-south direction and numbered consecutively from 

 south to north. Facing north for directions right and left, rows of 

 stakes parallel to the first row were given the additional designation R 

 or L. The first stake to the right of stake 1 would be 1-R-l. The 

 second stake to the left of stake 3 would be 3-L-2, and so forth. Since 

 engineering surveying equipment was provided, and it was calibrated 

 in feet and tenths of feet, these units were used in the field. 



Artifact locations were recorded in three dimensions, with hori- 

 zontal measurements taken from the nearest stake. Vertical measure- 

 ments were taken from both the surface of the ground and by alidade. 

 Animal bones and other specimens were normally recorded by arbi- 

 trary levels within a grid square. 



Sites were mapped and contoured with excavations, houses, and 

 other features shown. It was not deemed necessary to relate each 

 map to sea level but only to a single benchmark arbitrarily chosen 

 within the site. Each house was excavated as a unit, and was con- 

 trolled by an interior grid system of 5-foot squares. Normally, a 

 house was trenched from north to south to determine what condition 

 the floor or floors were in. Further excavation could be made by 

 parallel trenches or by quartering. A few houses were excavated 

 by horizontal plane sections 1 foot apart so as to distinguish the floor 

 from the fill. It was not feasible to follow floors, for they were far 

 too thin and nebulous. It proved to be easier to find them in profile 

 by the changes in soil color. 



