210 BULLETIN 183, UNITED STATES NATIONAL MUSIEUM 



pit 4 at Steed-Kisker. Moorehead (1928, pi. 25, fig. 13) jQgures another 

 example from the Saw Mill mound at Cahokia.^^ 



Between the house type evidenced at Steed-Kisker and the little 

 known Middle Mississippi dwellings elsewhere there seems to be only 

 a general similarity. Middle Mississippi houses are characterized by 

 Phillips (1940, p. 351) as simple, rectangular, 1-room affairs, lacking 

 specialized entrances or other features, and constructed of bent poles 

 covered with cane, thatch, or clay daub; occasionally interior posts 

 suggest a more substantial structure. At the Fonts site, in Fulton 

 County, 111., Cole and Deuel (1937, pp. 112-117) found rectangular 

 sunken floors outlined by shallow trenches into which walls of bark, 

 wattlework, or other material are thought to have been set. Fire- 

 places were usually located near the center of the floor; no evidence 

 was noted of wattle-and-daub roof or walls, or of openings for doors, 

 nor do the ground plans show any systematic arrangement of interior 

 roof supports. At Aztalan, Barrett (1933, p. 88) suspects that the 

 houses, both circular and rectangular in form, were built of wood and 

 plastered with mud. Farther to the south, at the Gordon site in Ten- 

 nessee, Myer uncovered circular semisubterranean house sites, which 

 he compares with the historic Omaha earthlodge (Myer, 1928, pp. 

 514^518, 527, 535). The house plans here, however, show no interior 

 supporting posts or any evidence of an entrance passage, two features 

 that occur almost universally in the historic and prehistoric earthlodges 

 of the Great Plains and the Missouri Valley. Wliat little information 

 is available on details of Middle Mississippi house construction sug- 

 gests that beyond their prevailingly rectangular form and semisub- 

 terranean floors there are virtually no specific resemblances to the 

 houses at Steed-Kisker. 



The single house unit fully excavated by us at Steed-Kisker, with 

 its four primary interior roof supports, vestibule entrance, straight 

 wall lines with rounded corners, and the pocket cache (?) between 

 hearth and doorway, is essentially identical with the prehistoric earth- 

 lodge sites of the Nebraska Culture farther up the Missouri (cf. 

 Strong, 1935, pp. 262-266; Gilmore and Bell, 1936, p. 308; Cooper, 

 1940) . This same basic type, usually subrectangular in outline, occurs 

 widely throughout the prehistoric Great Plains and probably is di- 

 rectly antecedent to the circular earthlodge characteristic of the semi- 

 sedentary tribes of the Missouri Valley in historic times. The genesis 

 of this rather specialized Plains earthlodge is not altogether clear, 

 though it presumably derives from a southeastern prototype (Strong, 



^^ The use to which thesse objects were possibly put is sugrrested by a "reaping hook" from 

 the Caddo Indians, figured by Rau (1876, p. 95, fig. S35). This consists of "the right lower 

 Jaw of an antelope, around which is bent a sapling forming the handle." The jawbone is 

 broken and polished at tiie diastema ; it lacks the worn groove noted on the Stoed-Kisker 

 specimen immediately behind the last molar. How the implement was used is not indicated. 



