114 SMITHSONIAN INSTITUTION 



town sites by Moreau B. Chambers of the Mississippi State Depart- 

 ment of Archives and History, who was then engaged in attempting 

 to locate the site of the old Bienville battlefield. I also had an oppor- 

 tunity to visit that section of the Tombigbee south of Tupelo. 



The second expedition was directed primarily to the investigation 

 of a site in Clarke County, Ala., on the banks of Choctaw Lake near 

 the Alabama River, which it was thought might prove to be the long- 

 looked-for location of the town of Mabila at which De Soto's most 

 famous battle with the Indians took place. James Y. Brame, Jr., of 

 Montgomery, Ala., the discoverer of this site, one of the foremost 

 students of the De Soto narratives and the De Soto route, organized 

 and promoted this investigation and obtained for its purposes access 

 to the land and the use of a camp maintained by the Choctaw Hunt- 

 ing and Fishing Club at Choctaw Bluff. The site is known locally as 

 Lower James Hammock and its position agrees well with what the 

 documents of the De Soto expedition would lead us to expect. Num- 

 bers of test holes on this site failed, however, to show evidence that 

 it had been surrounded by a stockade or traces of European materials 

 or of skeletons, and while there is still a possibility that this is the 

 site in question, it remains unproven. The potsherds collected here 

 show certain interesting features, particularly in the employment of 

 the edge of a corrugated shell in decorative designs and indications 

 that the Indians living here made pots with long legs comparable 

 in length to some used in Costa Rica, though there is no cultural 

 relation whatever between them. After work was suspended at this 

 site, investigations were extended to other parts of the county, but 

 with the exception of a small site near Gainestown previously located 

 by Mr. Brame, and the old salt works along Salt Creek (fig. 115), no 

 Indian remains of consequence were found in the sections visited. 



After leaving Clarke County, I went to Tuscaloosa, and from there 

 David L. De Jarnette took me to Scottsboro, from which point as a 

 center I was enabled to examine the middle course of the Tennessee 

 River between Guntersville and the Tennessee State line, the object 

 being to determine the probable route pursued by De Soto's army in 

 descending this river in the summer of 1540 (fig. 116). It seems 

 evident that they crossed and recrossed more times than have been 

 supposed. 



