SEVENTY-SIXTH ANNUAL REPORT 13 



the Stallings Island type and fragments from steatite vessels. This 

 stratigraphic evidence augments that f omid in other locations along 

 the river. Sergeant Chase turned over to the River Basin Surveys 

 party extensive notes and collections resulting from his previous 

 work at both sites. He also assisted Mr. Huscher in making detailed 

 plane-table maps of the sites and plans of the excavations. 



G. Hubert Smith excavated in two historic sites in the Walter F. 

 George Reservoir Basin. One of them on the Georgia side of the 

 river was the location of the village of Roanoke, a colonial settlement 

 that had originally been occupied by Creek Indians but was subse- 

 quently taken over by the vvliites who lived tliere from 1831 until the 

 community was destroyed by Indians in May 1836. Because of the 

 long period in which the area was under heavy cultivation, Mr. Smith 

 was unable to determine the settlement pattern or to obtain outlines 

 for any of the village structures. He did, however, obtain an ex- 

 tensive collection of specimens attributable both to the white occupa- 

 tion and that by the Indians. Careful study of the material may 

 provide information that will be useful in dating some of the other 

 late Indian sites along the river. From the Roanoke site Mr. Smith 

 went to one on the Alabama side in Russell County, which was the 

 location of a fort built and occupied by the Spaniards from 1689 to 

 1691. The fort known as Apalachicola was probably the most north- 

 ern outpost of the Spanish occupation in the Southeast and was 

 erected for the purpose of stemming the southward expansion of 

 the English. The Spaniards possibly did not occupy the fort con- 

 tinuously, but lived at times in an adjacent Indian village. The fort 

 was destroyed by the Spaniards to prevent its falling into the hands 

 of English traders from the Carolinas who were operating among 

 the Creek Indians. Mr. Smith did not dig in the fort proper but 

 confined his investigations to the area immediately surrounding it in 

 order to delimit the extent of the fortifications and to determine the 

 proximity of Indian occupation. The fort remains will not be sub- 

 jected to flooding by the Walter F. George Reservoir, but the maxi- 

 mum pool level will not be far distant and may damage the remains 

 to some extent as a result of seepage. Consequently it is thought that 

 a thorough study should be made of the site at a later date. Further- 

 more, associations between Spanish and Indian objects will provide 

 a helpful checking point in establishing chronology of the area, par- 

 ticularly since the exact dates for the fort are Imown. After complet- 

 ing the investigations at the two sites, Mr. Smith assisted Mr. Huscher 

 in making detailed plane-table maps and trench plans for both. 



In addition to the test excavations described above, Mr. Huscher 

 located and recorded 10 new sites in the Walter F. George and Co- 

 lumbia areas and made collections from 46 sites. At tlie end of the 



