Chickamauga Reservoir Mussels 81 



nessee River until ca. AD 100. 



In evaluating similarities and differences among mussel species 

 assemblages collected by people of various cultural groups, factors such 

 as individual preference or selection (e.g. large vs. small specimens), 

 location of the habitation site in relation to available mussel beds, sea- 

 son of the year collected, and probable differences in species comprising 

 the beds along a section of river, must be taken into account. On the 

 basis of samples collected during this study it is apparent that certain 

 species, such as A. ligamentina and D. dromas, or an assemblage of 

 closely related forms (e.g. Pleurobema spp.), were common and gener- 

 ally distributed throughout this section of the river during the past 2000 

 years. Considering the total sample from all cultural units, valves of 

 Pleurobema spp., A. ligamentina, and D. dromas comprised 75% of all 

 valves collected. 



Valves of D. dromas varied from 24% in the Middle Woodland site 

 sample to 45% in the multicomponent Middle Woodland/ Mississippian 

 sites, averaging 35% of the total sample. Of the 60,350 valves studied by 

 Warren (1975) from the Widows Creek site shell midden (Tennessee 

 River, northeastern Alabama: Early Woodland to Late Middle Wood- 

 land), D. dromas comprised 23% of all shells. Of the seven shell mounds 

 sampled in the Pickwick Landing Basin (Tennessee River, northwestern 

 Alabama: Archaic, Woodland), D. dromas was "One of the most 

 abundant species in these shell deposits . . . and made up a major part 

 of the total mussel fauna gathered for food" (Morrison 1942:359). 

 Although a form of D. dromas is still fairly common in unimpounded 

 reaches of the upper Powell and Clinch rivers in East Tennessee, the 

 large river form appears to be surviving at only one restricted locality in 

 the Tennessee River (Chickamauga Reservoir) (Pardue 1981). 



Ortmann (1918:556) stated that E. crassidens is "In the Tennessee 

 below Knoxville, and down to Chattanooga, it is extremely abundant"; 

 it is reported to now be "by far the most abundant species in the upper 

 Tennessee River" (Pardue 1981:44). Valves of this large, big-river mussel 

 occurred in all middens sampled and accounted for about 6% of the 

 total shell sample. Specimens of the related species, Elliptio dilatatus 

 (Raf. 1820), also occurred at all sites but had been collected in greater 

 numbers (ca. 12% of the total) than E. crassidens, possibly because it 

 was more numerous in shallower riffles and shoals and consequently 

 more accessible to the Indian. Today, populations of E. dilatatus have 

 been greatly reduced in the Tennessee River, probably as a result of 

 impoundment and its concomitant adverse factors. Actinonaias liga- 

 mentina is another species that occurred consistently in all middens; 

 valves of this species varied from ca. 4.5% of the total from Late Wood- 

 land/ Mississippian sites to nearly 9.5% at the Middle Woodland site 

 (average for all sites, 7.5%). At the Widows Creek site (Warren 1975) its 



