A prehistoric unionid assemblage from the Big Black River drainage in Hinds County, Mississippi.

Evan Peacock (1)

Archaeological remains can provide baseline biogeographical data for plant and animal species. Freshwater mussel shells from the Pocahontas site, a ca 500-year old Native American mound and village in Hinds County, Mississippi, were analyzed. A minimum of 20 taxa were represented in over 1,800 mussel valves obtained from a large pit feature. Frequency data for these taxa are presented. The known historical ranges of three species--Gyprogenia aberti, Plethobasus cyphyus, and Quadrula quadrula--can be extended into the Big Black River drainage based upon their presence at the site.

Shell deposits left by prehistoric humans have been found throughout the world, on every continent except Antarctica. Most known shell-bearing sites date to the Holocene epoch, probably because melting glacial sheets caused sea levels to rise at the end of the Pleistocene, destroying many coastal and estuarine sites before around 10,000 years ago (Claassen, 1998:1, 2). Throughout the Holocene, prehistoric Native Americans intensively harvested freshwater mussels from the river systems of the southeastern United States. Evidence of this activity exists in abundance at archaeological sites in the region, where archaeologists frequently recover thousands of mussel valves from excavations (Peacock, 2002).

In late 1974 and early 1975, archaeologists with the Mississippi State Highway Department (now the Mississippi Department of Transportation) conducted excavations at the Pocahontas site (state site number 22HI500) in Hinds County, Mississippi (Figure 1). The site is a prehistoric village with a large, flat-topped mound characteristic of the Mississippian period (A.D. 1000-1540). Among other things, the excavations produced large amounts of freshwater mussel shell, analyzed for the first time in this report.

MATERIALS AND METHODS

Natural setting--The site is located near the end of a broad, low ridge separating Limekiln Creek, which lies about 240 meters to the east, and a small tributary of Limekiln located about 270 meters to the west/northwest. Limekiln Creek flows into Bogue Chitto Creek, itself a tributary of the Big Black River. The site lies within the thick loess physiographic zone as displayed in Moran et al. (1997).

Archaeological work--Excavations were done prior to the four-laning of U.S. Highway 49 and attendant construction of a parking facility at the mound (Rucker, 1976:4). Most of the mussel shell recovered was found in "Area B," an area of cultural deposits near the eastern edge of the mound. Rucker (197625-27) describes "a large basin shaped pit" approximately 12 meters long and 4 meters wide. Its original depth is unknown, as the pit had been truncated during construction of Old Highway 49. Rucker interpreted this feature as a borrow pit for mound fill, and described the contents and excavation procedures as follows:

Because preservation of bone and shell remains appeared to be excellent, arrangements were made to waterscreen virtually all of the pit fill in an attempt to secure nearly 100 percent recovery of both large and small remains... The large quantity of mussel valves present in several lenses made excavation with shovels most difficult, and resort was had [sic] to four-pronged gardening forks. The fork tines more easily penetrated the shall [sic] layers and resulted in less damage to mussel valves, bone fragments, and pottery sherds. Pit fill was initially broken loose using garden forks, and all visible sherds and large bone pieces picked out and placed in level bags for hand washing and laboratory sorting. Fill dirt containing all shell and other small cultural remains was then shoveled into large cotton bags and removed to a water source for water-screening through one-sixteenth inch aluminum mesh screen (Rucker, 1976:28, 29).

Based on stylistic analysis of the recovered ceramics, Rucker (1976:35) believed the pit to date between A.D. 1400 and 1600, although overall occupation of the site probably encompassed a greater span (Rucker, 1976:90). Mussel shell occurred in discrete lenses within the pit and was tentatively interpreted by Rucker as representing seasonal deposits. We are assuming that the mussel remains excavated at Pocahontas originated in nearby Limekiln creek, as most prehistoric mussel assemblages in the Southeast seem to have been gathered from the nearest waterway capable of sustaining a mussel population (Peacock, 2002). It is possible, however, that mussels from elsewhere in the drainage are represented.

RESULTS

Considerations of bias--According to archaeologists currently working at the Mississippi Department of Transportation, much of the shell recovered from the site was discarded without being analyzed. Unfortunately, no tabulation of shell by valve count or weight is given in Rucker's report. It is thus difficult to say how representative of the overall site assemblage the shell sample reported herein actually is. Two shell beads were the only modified shell found in the excavations, so it is unlikely that much shell was being introduced for reasons other than subsistence (it should be noted, however, that crushed mussel shell was a common pottery tempering agent at the site). The range of taxa identified does not suggest cultural bias in the selection of species.

Following Peacock (2000), the left valves of three species (Fusconaia flava, Pleurobema rubrum, and Amblema plicata) were analyzed for size using the pallial line--lateral tooth (PL-L) measurement suggested by Warren (1975). Peacock (2000) has demonstrated that interior measurements such as PLL are strongly correlated with more typical size measurements such as length. A lateral interior measurement (e.g., between the anterior and posterior adductor scars) would perhaps provide a better proxy for shell length, commonly used by biologists studying living mussel populations. Unfortunately, many valves from Pocohontas are broken along the posterior and/or ventral margins, making measurement of shell length difficult or impossible. The measurements were taken only on the left valve in order to prevent one bivalve from being counted twice and thus skewing the data. Left valves were used because they were more abundant for the three species chosen for measurement. As measured by PL-L, the valves are small overall ( F flava--average 16.72 mm, range 10.15-23.80 mm; P rubrum--average 15.82 mm, range 8.8--25.33 mm; A. plicata-average 19.67 mm, range 13.34-25.53 mm), a common characteristic of archaeological shell in the eastern United States (Peacock, 2000,2002). These data indicate that collection was not biased against small mussels. Large mussels either were not present at that time or were avoided by the prehistoric shell fishers. A recent literature review and analysis of shell from three archaeological sites on the Tombigbee River suggests that prehistoric human bias in the selection of species was the exception, rather than the rule, in the Southeast (Peacock 2000).

As can be seen in Table 1, 454 valves (about 24 percent of the total assemblage) were unidentifiable even to genus. These were mostly small and/or fragmentary valves with eroded shell features. Such a high percentage of unidentifiable valves is not uncommon in archaeological assemblages from the Southeast (Peacock, 2000; Peacock and Chapman, 2001). Overall, the shell from Pocahontas is in moderately good condition in terms of preservation. The major source of bias that likely is structuring the assemblage is the limited contexts from which the shell was retrieved. As demonstrated by Peacock (2000), the accuracy of taxonomic frequency data improves when multiple contexts at an archaeological site are sampled.

DISCUSSION

Discussion of taxa--The mussel taxa identified at Pocahontas are presented in Table 1. The unidentifiable valves were not used in calculating the percentages. These data provide an interesting comparison to modem mussel surveys in the Big Black River drainage.

Hartfield and Rummel (1985) compared the results of their 1980-81 mussel survey along the Big Black River to earlier data reported by Hinkley (1906) and Grantham (1969). Following Hartfield and Rummel, we are assigning specimens to Fusconaia flava rather than to F. cerina (cf. Bogan 1987). The two are indistinguishable in archaeological material, but F cerina is considered to be an endemic of the Tombigbee River drainage. Also following Hartfield and Rummel, we have assigned our specimens of Obovaria to O. subrotunda, although it is possible that other species of this genus may be present in the Pocahontas collection.

Of the 31 species of unionids collected by Hartfield and Rummel, 15 are present in the prehistoric shell from Pocahontas. Some of the taxa listed by Hartfield and Rummel but not found at Pocahontas are silt-tolerant species (e.g., Plectomerus dombeyanus; Miller et al., 1992:10) that may have extended their ranges in historic times. Others (Uniomerus tetralasmus, Pyganodon grandis, Anodonta imbecillis, Ligumia subrostrata) were collected by Hartfield and Rummel "from oxbows or [were] found dead below oxbows draining into the river" (Hartfield and Rummel, 1985:119). Other differences between their list and ours may simply reflect differences between small- and large-stream faunas. Surprisingly absent at Pocahontas in this regard is Tritogonia verrucosa, found in the Big Black by Hinkley (1906), Grantham (1969), and Hartfield and Rummel (1985). Its absence at Pocahontas may simply be due to sampling error.

Four species reported by Hinldey (1906) and/or Grantham (1969) (Pleurobema curtum, Obovaria unicolor, Lampsilis hydiana, L. ornata) were not found by Hartfield and Rummel. We did not find any of these taxa in the archaeological shell assemblage. Distinguishing these species in archaeological material would be difficult, however (e.g., Peacock and Chapman, 2001). Hartfield and Rummel reported three species (Quadrula nodulata, Q. cylindrica cylindrica, and Pleurobema rubrum) "for the first time in the eastern drainage of the Mississippi River south of the Tennessee River" (1985:117, 118).

Hartfield (1987) notes that all of these species have been found in prehistoric shell middens in the Delta. Pleurobema rubrum is the dominant mussel in our sample, making up about 27 percent of the total identified shell by valve count. Quadrula cylindrica cylindrica also is fairly common, at about five percent Quadrula nodulata is present, although it apparently was rare in Limekiln Creek.

To our knowledge, the western fan shell, Cyprogenia aberti, has not been reported in the Big Black drainage before. It apparently was common in prehistoric times (assuming prehistoric human collection bias was negligible), being the fourth most abundant taxon in the assemblage at about ten and one-half percent. Peacock and Chapman (2001) recently reported a single valve of this species from a prehistoric shell midden on the Ouachita River in Ouachita Parish, Louisiana (cf. Vidrine et al., 1999). The known historical range of C. aberti was pushed considerably southward and eastward by its presence at two prehistoric (A.D. 300-600) sites located on O'Neil Creek, a tributary of the Yazoo River in Yazoo County, Mississippi (Bogan 1987). The Yazoo is the first major drainage north of the Big Black. The range of C. aberti can again be extended based on its presence at Pocahontas. The archaeological specimens are generally triangular in outline and have a well-developed sulcus adjacent to the posterior ridge. Sample valves are shown in Figure 2.

Other species not previously reported from the Big Black drainage are Plethobasus cyphyas and Quadrula quadrula. The presence of the P cyphyus at prehistoric sites on O'Neil Creek (Bogan 1987) represented a major extension of its range as shown, for example, in Parmalee and Bogan (1998: Range Map 78). The Pocahontas site specimens represent a further southern extension for this species. The presence of the Q. quadrula in the Big Black drainage is not as much of a surprise, given its occurrence in the Mississippi River drainage south into Louisiana (Vidrine, 1993). Bogan (1987) also found this species from archaeological sites on O'Neil Creek. Sample valves of both species are shown in Figure 2.

CONCLUSIONS

The archaeological shell assemblage from Pocahontas has provided valuable data on the historical ranges of freshwater mussels. The natural ranges of three species (Cyprogenia aberti, Plethobasus cyphyus, Quadrula quadrula) can be extended based upon their presence at the site. Although not discussed in this article, the site also produced large numbers of aquatic gastropods of the genus Campeloma. Apparently these gastropods were a food source for the prehistoric inhabitants of the Pocahontas site.

It would be valuable to have a larger sample of mussel valves from the site, to confirm the taxonomic frequencies and to provide identifications for poorly represented taxa such as those of the genus Lampsilis. Future archaeological work there should be conducted with this need in mind.

ACKNOWLEDGEMENTS

This article had its roots in a paper prepared by the junior author for an undergraduate class in Archaeological Method and Theory, taught at Mississippi State University by Dr. Janet Rafferty. We would like to thank the archaeologists with the Mississippi Department of Transportation for making the shell available for analysis and for providing information regarding the site. We would like to thank Art Bogan for bringing the O'Neil Creek archaeological shell assemblages to our attention. Bob Jones and Paul Hartfield were kind enough to check our identifications on a large sample of valves. The work was conducted at the Cobb Institute of Archaeology Mississippi State University, where the senior author is a Senior Research Associate. We would like to thank the reviewers for their positive comments and helpful suggestions.

Table 1

Freshwater mussel taxa identified from the Pocahontas site
(22HI500), Hinds Co., Mississippi

Taxon                    # of   Percent of total
                        valves     identified
                                     valves

Pleurobema rubrum        385         27.07
Fusconaia flava          301         21.17
Obovaria subrotunda      162         11.39
Cyprogenia aberti        149         10.48
Quadrula pustulosa        83          5.84
Amblema plicata           82          5.77
Quadrula cylindrica       70          4.92
cylindrica
Fusconaia ebena           50          3.52
Elliptio dilatata         31          2.18
Obliquaria reflexa        29          2.04
Plethobasus cyphyus       28          1.97
Truncilla truncata        11          0.77
Lampsilis sp.             10          0.70
Quadrula nodulata          9          0.63
Quadrula sp.               8          0.56
Quadrula quadrula          6          0.42
Ellipsaria lineolata       3          0.21
Elliptio crassidens        2          0.14
Truncilla donaciformis     1          0.07
Glebula rotundata          1          0.07
Megalonaias nervosa        1          0.07
unidentifiable           454

Total:                  1876        100.00

(1.) Author for correspondence.

LITERATURE CITED

Bogan, A. E. 1987. Molluscan Remains from the Milner Site (22YZ515) and the O'Neil Site (22YZ624), Yazoo County, Mississippi. Appendix D in Hearfield, Price and Greene, Inc., Data Recovery at the Milner (22YZ515) and O'Neil Creek (22YZ624) Sites, Yazoo County, Mississippi. Report submitted to the U.S. Army Corps of Engineers, Vicksburg District, by Heartfield, Price and Green, Inc., Monroe, Louisiana.

Claassen, C. 1998. Shells. Cambridge University Press. 266 pp.

Grantham, B.J. 1969. The Freshwater Pelecypod Fauna of Mississippi. Ph.D. Dissertation. Hattiesburg: University of Southern Mississippi.

Hartfield, P. 1987. Status of Selected Freshwater Mussels (Unionidae) in Mississippi. Journal of the Mississippi Academy of Sciences 32:133-141.

Hartfield, P.D., and R.G. Rummel. 1985. Freshwater Mussels (Unionidae) of the Big Black River, Mississippi. The Nautilus 99(4): 116-119.

Miller, A.C., B.S. Payne, and P.D. Hartfield. 1992. Characterization of a Freshwater Mussel (Unionidae) Community in the Big Sunflower River, Sharkey County, Mississippi. Journal of the Mississippi Academy of Sciences 37(3):8-11.

Moran, L.P., D.E. Pettry, RE. Switzer, S.T. McDaniel, and R.G. Wieland. 1997. Soils of Native Prairie Remnants in the Jackson Prairie Region of Mississippi. Mississippi State, Mississippi: Mississippi Agricultural and Forestry Experiment Station, Bulletin 1067.

Parmalee, P.W., and A.E. Bogan. 1998. The Freshwater Mussels of Tennessee. Knoxville: The University of Tennessee Press. 384 pp.

Peacock, E. 2000. Assessing Bias in Archaeological Shell Assemblages. Journal of Field Archaeology 27(2): 183-196.

Peacock, E. 2002. Shellfish Use During the Woodland Period in the Middle South. In David G. Anderson and Robert C. Mainfort, Jr., eds. The Woodland Southeast. Tuscaloosa: The University of Alabama Press In Press.

Peacock, E., and S. Chapman. 2001. Taphonomic and Biogeographic Data from a Plaquemine Shell Midden on the Quachita River, Louisiana. Southeastern Archaeology 20(1):44-55.

Rucker, M.D. 1976. Archaeological Investigations at Pocahontas Mound A, Hinds County, Mississippi. Jackson: Mississippi State Highway Department Archaeological Excavation Report No. 3.

Vidrine, M.F. 1993. The Historical Distributions of Freshwater Mussels in Louisiana. Eunice, Louisiana: Gaul Q. Vidrine Collectibles. 225 pp.

Vidrine, M.F., J.W. Saunders, and C.M. Allen. 1999. Freshwater Mussel Remains from Watson Brake Mound Complex. Abstract in the Louisiana Archaeological Society program for the 1999 annual meeting in Monroe, Louisiana.

Warren, R.E. 1975. Prehistoric Unionacean (Freshwater Mussel) Utilization at the Widows Creek Site (1JA305), Northeast Alabama. M.A. Thesis. University of Nebraska, Lincoln.