52 BUREAU OF AMERICAN ETHNOLOGY [Bull. 189 



Details of the house superstructure are rare. In House 3, timbers 

 along one wall indicate that the wall posts were at least 5 feet high. 

 There was no evidence of leaners on the bench of earth outside the 

 house wall. These facts, together w^ith the presence of small branches 

 and twigs along the house walls, suggest that the walls were interlaced 

 with branches. Some form of a wattle-and-daub wall may be rep- 

 resented in this architectural form rather than the earth-covered 

 lodge of later, historic tribes. Each of the houses had a line of center 

 posts which apparently supported a ridge pole. In House 2, the roof 

 may have been A-shaped or gabled, but in houses 1 and 3 the roof form 

 is complicated by the presence of the intermediate row of posts. 



Bell-shaped pits predominate in the houses, but basin-shaped pits 

 are also frequent. The bell-shaped pits in the house floors are not par- 

 ticularly large, nor are they deep. A maximum depth of about 2 feet 

 prevails. Exterior cache pits are larger ; some of them are 6 feet in 

 diameter and attain a depth of 5 feet below the present surface. Orig- 

 inally they were probably no more than 4 feet deep. The additional 

 depth is due to the soil accumulation over the site since it was aban- 

 doned. All the pits contained refuse, and even the bell-shaped pits 

 were used for rubbish disposal after their primary function as food 

 storage pits was fulfilled. 



A long shallow trench in House 3, F70, may have an analogy in Ff)7 

 in House 2, although this latter pit is irregularly shaped. These pits 

 are similar to features in houses at Thomas Kiggs (Hurt, 1953, p. 8, 

 figs. 7-8, 10, 12). The pits at Thomas Eiggs contained very little 

 refuse, and some of them were lined with wooden slabs. Although the 

 pits at Paul Brave contained no wood, their refuse content was low. 

 These pits are also similar in form to a pit in the midline of the long 

 rectangular house at 32ME59, the site of "Grandmother's Lodge," al- 

 though here the pit contained small stones (Woolworth, 1956, pi. 2). 



ARTIFACT COMPLEX 



The Thomas Riggs site, in central South Dakota, is the only ade- 

 quately excavated site that compares closely with Paul Brave, although 

 limited comparisons are possible with "Grandmother's Lodge." The 

 pottery and other artifacts from Paul Brave refer to the "Archaic 

 Mandan" period, the earliest Imown village culture on the Missouri 

 River in present-day North Dakota. Comparisons with other sites 

 are possible, but are rather restricted because of a lack of data. Ex- 

 cavation in long rectangular house sites has been restricted primarily 

 lo testing in North Dakota, and space does not permit a detailed com- 

 parison of the artifacts from Paul Brave with the numerous but lim- 

 ited samples from sites listed by Will and Hecker (1944, pp. 118-121) 

 as "Archaic Mandan." An inspection of the various collections in the 



