2 BUREAU OF AMERICAN ETHNOLOGY [Bull. 189 



a preliminary classification of the pottery from the excavations, with 

 notes on extra-site distributions. In addition, there are references 

 to Paul Brave in the report on the Thomas Riggs site, 39HU1 (Hurt, 

 1953). In Hurt's report Paul Brave is referred to as the "Fort 

 Yates site." In the present study, site 32SI4 is designated as the 

 Paul Brave site to avoid duplication in terms, since the term "Fort 

 Yates" has been used in other contexts. 



ARCHEOLOGY OF THE SITE 



SITE DESCRIPTION 



The Paul Brave site is located in the N1/2N1/2NE14 of sec. 8, T. 129 

 N., R. 79 W., Sioux County, N. Dak. It is on the west bank of the 

 Missouri River 6 miles downstream from the town of Fort Yates, and 

 about 6 miles upstream from the North and South Dakota boundary. 

 The site is on a rolling terrace that overlooks the Missouri River flood 

 plain, this terrace averaging 20 feet above the flood plain. To the 

 south, the land rises in a series of low hills that abut against the 

 Missouri River bluffs, about one-half of a mile distant (map 1 ; also 

 pi. 1, a) . Fire Heart Butte, a high, flat-topped butte, is about 4 miles 

 to the southeast. 



At the time the site was investigated in 1955, it was found that at 

 least 14 oval house depressions were distributed over an area of about 

 4 acres. These depressions were as much as 65 feet long and were 2 

 to 3 feet deep. They were filled with a lush vegetation that contrasted 

 with the short, dry grass on higher points in the village area. The 

 site is said to have been cultivated (Will and Hecker, 1944, p. 89), 

 but activity does not seem to have obscured the surface indications of 

 the larger subsurface features. The village may have been larger 

 originally, because Will and Hecker (1944, p. 89) also noted house 

 floors exposed in the terrace edge facing the river. This terrace was 

 checked in 1955. A number of artifacts were recovered from the sur- 

 face, but there was no indication of house floors. There is no evidence 

 that the site was fortified, but it is partially isolated from the re- 

 mainder of the terrace by a narrow swale south of the village (map 2 ; 

 also pi. 1, a). 



EXCAVATION METHODS 



Preliminary testing determined that the houses were built in shal- 

 low pits. Three houses were chosen for excavation after the depth 

 of the house floors and the house walls had been determined. These 

 houses were evenly distributed over the site, and none of them had 

 been tested previously. The entire site was mantled by a deposit of 

 light-buff aeolian soil. This mantle was as much as 4 feet deep over 

 some of the houses, but high points within the village were buried 



