Wando Series Ceramics: Behavioral Implications of a Local Ceramic Type Eric C. Poplin HISTORY AND DESCRIPTION Archaeological investigations in and around Charleston Harbor, South Carolina, over the last 40 years encountered a type of Pre-Contact Native American pottery that required nearly three decades for archaeologists to recognize as a component of the regional ceramic sequence. Many investigators recognized the identifying characteristic of this pottery, namely limestone tempering, but no one ventured to define a new series or type. Figure 1 displays the Charleston Harbor area where most of the limestone-tempered ceramics have been reported. There was even some confusion over the nature of the temper, much like the recently defined Hamps Landing limestone tempered wares of coastal North Carolina (Herbert and Mathis n.d.). Limestone tempered pottery frequently displays voids where the limestone leached or burned away very similar to the voids created by shell tempering. Figure 2 displays an example of Wando Plain sherds with the voids highly visible. Like the Hamps Landing series in North Carolina, the South Carolina limestone tempered wares were often called shell tempered. Early European explorers even noted the use of shell tempering for ceramics. Thus, a misinterpretation of limestone tempering by twentieth century archaeologists is not so glaring an error as one might expect. In the late 1980s, archaeologists working in the Charleston area began to recognize that the material that created the voids in this pottery was indeed limestone, presumably derived from local outcrops or gravel beds. Figure 4 displays a limestone nodule collected from the marsh surface along the Wando River. By the early 1990s, enough information had been gathered on the limestone-tempered wares to permit the definition of a series. Adams and Trinkley (1993) provided the first descriptions of the series, calling it Wando after the coastal river that flows into Charleston Harbor along whose banks the limestone tempered pottery most frequently appeared. Adams and Trinkley (1993) defined two types- Check Stamped and Cord Marked. Investigations at other sites in the region recovered brushed, fabric impressed, incised, simple stamped, and punctated sherds as well as undecorated ones. Decorations, with the exception of shell scraping occur only on the exterior of the vessels. Rims are usually straight, reflecting various sized straight-sided bowls and jars; bowl forms dominate. Vessel thicknesses are 6-8 mm. The limestone temper fragments, when present, are sub-round and range in size from medium to granular. Sand is a secondary tempering agent in many vessels; iron concretions and other local lithic fragments occasionally are present rather than limestone. Figures 5 and 6 display examples of Wando Cord Marked and Wando Simple Stamped ceramics, respectively. Wando surface decorations are identical to those that occur on Middle Woodland Deptford and Wilmington series ceramics from the central coast of South Carolina. Thus, most researchers assumed that the Wando series was manufactured at the same time. Recent investigations at two sites, 38CH1025 and 38BK1603, produced four radiocarbon dates from charcoal in association with Wando ceramics that indicate a Late Woodland association. Dates, with 2 sigma calibrations, range from AD 600-1220, although three of the four dates coincide in the AD 600-1000 range. Figure 7 displays date ranges from these two sites. Table 1 summarizes the dates as generated by Beta Analytic, Inc. 57
Figure 1. The Charleston Harbor Area. 58
Figure 2. Example of Wando Plain pottery. Note the characteristic voids in the surfaces. Figure 3. Native Americans using pottery (from Dumont 1753). 59
Figure 4. Limestone pebble from the marshes of the Wando River. Figure 5. Example of Wando Cord Marked pottery. 60
Figure 6. Example of Wando Simple Stamped pottery. Figure 7. Range of radiocarbon dates from 38CH1025. 61
Table 1. Radiocarbon Dates from 38CH1025 and 38BK1603. Sample 38CH1025 Level (horizon) Radiocarbon Age 2 Sigma Calibration Intercept Date 1 Sigma Calibration Beta 1144360 2 (A/shell midden) 950 ± 60 BP AD 990 to 1220 AD 1040 AD1015 to 1170 Beta 144361 3 (shell midden) 1340 ± 70 BP AD 600 to 815 AD 840 to 855 AD 6670 AD 645 to 720 AD 745 to 760 Beta 144362 4 (sub-midden) 1460 ± 60 BP AD 445 to 670 AD 615 AD 550 to 650 38BK1633 Beta 172218 2 (Ab horizon) 960 ± 80 BP AD 910 to 920 AD 960 to 1250 AD 1030 AD 1000 to 1180 Beta 172219 Feature 601 (pit) 1190 ± 60 BP AD 690 to 990 AD 870 AD 770 to 900 The Late Woodland association for Wando ceramics, particularly the Check Stamped type, runs counter to our previous interpretations of the differences between Middle and Late Woodland ceramics in coastal South Carolina. For decades, archaeologists defined the Late Woodland period by the lack of check stamping, the hallmark of the Early and Middle Woodland Deptford series (see Poplin et al. 1993 or Trinkley 1989, 1990 among many others). The presence of check stamping in the Late Woodland Wando series appears to invalidate this assumption. DISTRIBUTION Recent excavations at site 38CH1025 by Eubanks et al. 1995 and Poplin et al. 2002), that contains predominantly Wando ceramics, permits a preliminary analysis of this locally manufactured series to the other ceramic series so commonly found around Charleston Harbor. Site 38CH1025, in the Pointe at RiverTowne Country Club (a residential subdivision on the east bank of the Wando River), contains three small shell middens with extensive disturbed areas in between. The significance of this site lies in the apparent layering in two of the shell middens that reflect occupations by the makers of Wando ceramics. Figures 8-10 displays the midden profiles from the excavations at 38CH1025. Thus, very limited excavations in these middens provided the material that generated the first radiocarbon dates associated with Wando ceramics, and an opportunity to examine a relatively unadulterated assemblage of limestone tempered ceramics. Also, there is information from excavations at site 38BK1603, in the O.L. Thompson Mine, a component of the Charleston Regional Business Center light industrial park on the Cainhoy Peninsula between the Wando and Cooper Rivers. A buried occupation horizon containing predominantly Wando ceramics lies within this site. Site 38BK1603 generated the second set of radiocarbon dates for Wando ceramics, and a much larger collection of sherds than the one recovered from 38CH1025 (Poplin et al. n.d.). 62
Figure 8. Profile of large shell midden at 38CH1025 (after Eubanks et al. 1995). 63
Figure 9. Profile of large shell midden at 38CH1025 (after Poplin et al. 2002). Figure 10. View of the profile of the small shell midden at 38CH1025 (after Poplin et al. 2002). 64
These two sites provide a limited sample for exploring the characteristics of Wando ceramics. While Wando components have been identified at over 50 sites in and around Charleston Harbor (see Figure 11), only a few of the sites that contain predominantly Wando ceramics have witnessed more than survey level investigations to date. Data recovery investigations occurred at four other sites (38CH843, 39CH908, 38CH949, and 38CH950); however, contractual, financial, and logistical constraints prevent the completion of reports for any of these investigations to date. Poplin et al. (2002) extracted information from the artifact catalogs for 38CH949 and 38CH950 to include in our analyses. Also, Fletcher and Poplin s (2002) extensive testing of site 38BK1803/1804, in the O.L. Thompson Mine, generated a large sample of Wando ceramics. Figure 11 displays the locations of all of these sites. Thus, we have data from five sites, two on the Cainhoy Peninsula between the Wando and Cooper Rivers and three on the east side of the Wando, to examine in greater detail. Obviously, these analyses and the following discussions represent a preliminary assessment, awaiting evaluations employing information from future investigations. The apparent limited geographic range and unique temper of Wando ceramics creates an opportunity to examine a small segment of the Late Woodland Native American population of coastal South Carolina. Poplin et al. (2002) explored several assumptions or hypotheses regarding Wando ceramics and their relationships with other regional ceramic series. Why did some Native Americans decide to use limestone as temper rather sand or grog, the more common tempering agents? Did limestone provide an improvement over sand or grog for some function? Why does limestone tempering occur only along the Wando River? Was limestone only available in this portion of the Coastal Plain? Or, was there a group of Native Americans who consciously selected limestone to include in their pottery? Did this set them apart from other groups who made sand and/or grog tempered pottery? CO-OCCURRENCE WITH OTHER TYPES Poplin et al. (2002) examined the distribution and co-occurrence of Wando ceramics with other types, and technological aspects of the ceramics themselves to attempt to address these questions, particularly the association of Wando ceramics with a distinct group of Late Woodland Native Americans. They reviewed 80+ reports of archaeological investigations in the Charleston Harbor region; 44 of these reports provided information on 162 sites that contain Middle Woodland, Late Woodland, or Wando components. Figure 11 displays the locations of these sites. Time constraints precluded the use of information from the Francis Marion National Forest. However, Robert Morgan, Forest Archaeologist (personal communication, 2002), notes that there are only a few sites in the Forest, which lies to the north and east of the Wando River basin, that contain Wando ceramics. For this analysis, sites with Deptford and Wilmington series ceramics reflect Middle Woodland occupations despite the continued manufacture of Deptford and Wilmington Cord Marked and Fabric Impressed types during the Late Woodland period; sites with McClellanville, Santee, and St. Catherine s ceramics reflect Late Woodland occupations. There are 99 sites with Middle Woodland occupations only. Three sites contain evidence of Late Woodland occupations only. Five sites contain only Wando occupations. Figure 12 displays the frequencies of each component among the sample of sites There are 54 sites in the sample that contain Wando ceramics. Only five sites produced Wando ceramics exclusively. Wando ceramics co-occur with Middle Woodland types at 49 sites, and with Late Woodland ceramics 65
Figure 11. The location of Wando pottery sites in Charleston Harbor region. 66
Figure 12. The frequency of components at the selected sites. Figure 13. The frequency of the co-occurrence of components at the selected sites. 67
Figure 14. The relative frequency of the co-occurrence of components at the selected sites. at 19 sites. Interestingly, all 19 of these sites also produced Middle Woodland ceramics. Like the presence of check stamped decorations, this could imply that Wando ceramics are a Middle Woodland phenomenon. The radiocarbon dates from 38CH1025 and 39BK1603 indicate otherwise. This also could be an Wilmington Cord Marked and Fabric Impressed types as Middle Woodland components, resulting in the under-representation of Late Woodland components in the sample. We will return to this apparent association later. Figure 13 displays the frequencies of component co-occurrences among the sampled sites. Poplin et al. (2002) also examined the relative frequency of each ceramic series at the sites where Wando ceramics co-occur with other ceramics. Only 50 of the 54 Wando sites had sufficient data to calculate these frequencies. Of these 50 sites, 13 have relative frequencies of Middle Woodland ceramics that are at least twice the frequencies of the next highest group (Late Woodland or Wando). Three sites have higher relative frequencies of Late Woodland ceramics. Fourteen sites have higher relative frequencies of Wando ceramics. The remaining 20 sites have relatively equal relative frequencies of Middle Woodland and Wando ceramics; that is, the frequency of Middle Woodland or Wando ceramics is less than two times the frequency of the other group. Figure 14 displays the frequencies of the sites within each of these groups. Like the presence/absence analysis above, this suggests that Wando ceramics co-occur more frequently with Middle Woodland occupations than Late Woodland occupations. Again, the association of Deptford and Wilmington types 68
Figure 15. Seriation of Middle and Late Woodland ceramic series recovered from selected sites. 69
that span both the Middle and Late Woodland periods with the Middle Woodland period in this analysis could create this apparent association. We will revisit this association below. Using the vertical distributions of Wando ceramics from the five sites noted above, Poplin et al. (2002) seriated all Wando, Deptford, Wilmington, and Late Woodland types by excavation level. Figure 15 displays these seriations. This analysis reveals that Wando ceramics occur in a pattern that is most similar to that evidenced by Late Woodland types. That is, Wando and Late Woodland types occur most frequently in the upper excavation levels while Middle Woodland Deptford types occur most frequently in lower excavation levels and decline in the uppermost excavation levels. Interestingly, Wilmington ceramics increase in frequency from the lower levels to the middle levels and again in the uppermost excavation levels. This could reflect the typological confusion of Wilmington and St. Catherines grog tempered wares. Many researchers do not attempt to separate these series. MANUFACTURE TECHNOLOGY AND CERAMIC ATTIBUTES Review of technological aspects of ceramic manufacture shows a greater similarity between Wando and Late Woodland types than between Wando and Middle Woodland types. We used information from Poplin et al. s (1993) excavations at the Buck Hall site (38CH644) in the Francis Marion National Forest for these comparisons. Detailed ceramic analyses of the ceramics from this site were most comparable to the analyses conducted for 38CH1025; similar analyses have not been completed for the other Wando ceramic sites that witnessed data recovery investigations to date. Sufficient data existed for three aspects of manufacturing technology to permit comparisons between the ceramic series. These aspects include carbon core retention, temper size, and interior finishing. Poplin et al. (1993) provide information for Deptford Check Stamped, Cord Marked, and Fabric Impressed vessels; McClellanville Cord Marked/Fabric Impressed vessels; and Santee Simple Stamped vessels. We combined all of the Wando vessels identified at 38CH1025 to generate similar data. Figure 16. The relative frequency of carbon core retention in selected ceramic types. 70
Figure 17. The relative frequency of temper size in selected ceramic types. Figure 18. The relative frequency of surface treatments evident on ceramic types from the selected sites. Comparison of these data indicate that vessels identified as the three Deptford types all display higher average frequencies of carbon core retention (60-69 percent) than the Late Woodland McClellanville (20 percent), Santee (54 percent), or Wando (44 percent) vessels. This suggests that Wando ceramics are technologically more similar to the Late Woodland series. Figure 16 displays the average frequencies of carbon core retention among the selected ceramic types/series. Temper size is more ambiguous. Deptford Check Stamped vessels display high frequencies of coarse and very coarse sand temper inclusions. Deptford Cord Marked vessels from38ch644 all had coarse sand temper. Deptford 71
Fabric Impressed vessels displayed higher frequencies of fine-medium sand temper. The Late Woodland McClellanville and Santee types all display high frequencies of fine-medium sand temper. Wando vessels from 38CH1025 display high frequencies of coarse and very coarse limestone fragments. This suggests that Wando ceramics are more similar to the Middle Woodland types although this could be an artifact of the size of the limestone fragments available to the makers of Wando pottery, or the differences between sand and limestone. Figure 17 displays the frequencies of temper size for each of the selected ceramic types/series. Comparisons of the interior finishes of the analyzed vessels shows that the Deptford types have predominantly (94-100 percent) smoothed interiors; only six percent of the check stamped vessels displayed scraped interiors. In comparison, the Late Woodland types display primarily smoothed interiors (50-75 percent) but also have low to moderate frequencies (12.5-50 percent) of other finishes, including scraped, semi-burnished, and burnished. Wando vessels possess primarily smoothed interiors (62 percent) with scraped, wiped, tooled, and roughened finishes accounting for the remaining diversity, with 1.7-16.7 percent per category. The greater diversity of finishing approaches suggests a greater similarity between Wando and the Late Woodland types. Figure 18 displays the relative frequencies of each finishing attribute among the selected ceramic types/series. Poplin et al. (2002) also examined the vertical distributions of decorative types within the Wando series. The current reconstructions of the Middle and Late Woodland ceramic sequences in coastal South Carolina suggest that check stamping and simple stamping precede cord marking and fabric impressing during the Middle Woodland. During the Late Woodland, cord marked and fabric impressed decorations appear before simple stamped decorations; check stamping does not appear again until the Early Mississippian period. Again, we seriated the Wando types recovered from the five sites noted above. Wando ceramics display a trend more similar to Late Woodland types. Cord marked and fabric impressed decorations occur more frequently in the lower (or earlier) excavation levels. Simple stamped and check stamped decorations occur more frequently in the upper or later excavation levels. Again, this suggests a greater similarity between Wando and Late Woodland ceramic series. Figure 19 displays the seriation of Wando series ceramics by type from the selected sites. FUTURE RESEARCH AND CONCLUSIONS Do these similarities and differences in specific ceramic attributes indicate a distinct population of Native Americans who lived along the Wando River during the Late Woodland period? The use of limestone tempering sets Wando ceramics apart from all other ceramic series found in the central coast of South Carolina. Use of limestone appears to be a conscious selection by the makers of Wando pottery. Although no one has made a comprehensive survey of possible outcrops or sources for the limestone fragments, the parent material underlies all of the Coastal Plain of South Carolina. It seems highly unlikely that only one small area of the Coastal Plain along the Wando and Cooper Rivers contains or exposes this parent material given the nature of the many other rivers (like the Waccamaw, Black, Ashley, Ashepoo, and Combahee) that also rise on the Coastal Plain, and in some instances contain greater volumes of water and greater erosional potential. Perhaps the limestone outcrops or deposits are associated with older beach ridges and higher elements of the modern topography like the Cainhoy Peninsula. Similar landforms exist in Georgetown, Dorchester, Colleton, and Beaufort Counties but we have yet to recover limestone-tempered pottery from these areas. Therefore, for the present, it seems reasonable to assume that there is a local Native American population who selected limestone fragments as the tempering agent for their pottery in and around Charleston Harbor during the Late Woodland period. 72
Figure 19. Seriation of Wando types recovered from the selected sites. Does the limestone provide some technological or functional characteristic to the vessels that sand or grog does not? Several efforts to examine possible functional differences between contemporary sand tempered Deptford vessels and grog tempered Wilmington vessels that often co-occur on Middle Woodland sites in the central coast of South Carolina have yet to demonstrate any major differences. Bridgman (2001) hypothesized that grog tempered vessels would be more efficient for cooking since the use of clays throughout the vessel promotes more even expansion and contraction during heating and cooling, and concomitantly less vessel failure. Sand tempered vessels served primarily as storage vessels. In the Middle Woodland assemblage they examined from 38CH949 and 38CH950, sooting occurred only a little more frequently on the grog-tempered vessels, providing very limited support to this hypothesis. Limestone tempering usually results in voids in the vessel surface and walls. These voids give the appearance of the reduction in the overall stability of the vessel. Also, given the continued breakdown of the 73
limestone that would result through repeated firing, it seems that limestone would make a poor choice as an element of a cooking vessel. While most vessels identified in the ceramic assemblage from 38CH1025 were likely used for storage, several specimens also display sooting that appears to be from use as a cooking vessel and not a result of the initial firing. Thus, Wando vessels serve all of the roles associated with the vessels from other Middle and Late Woodland ceramic series. The final aspect of the above analyses that supports the association of Wando pottery with a distinct local population is its unique and predominant occurrence at a number of sites in the region. If Wando ceramics were merely another element of the Late Woodland ceramic assemblage of the central coast, we would expect them to co-occur with the other series. Instead, we see sites, albeit less than 10 percent of all Wando sites in the sample, where Wando ceramics occur exclusively as well as sites where they dominate the ceramic assemblage. This implies that there is a distinct element of the population who make and use Wando pottery at specific sites in the region. Interestingly, Wando ceramics co-occur more frequently with Middle Woodland ceramics than Late Woodland ceramic series. While this may be an artifact of the definition of Middle and Late Woodland associations employed in the above analyses, it could imply that groups along the Wando and Cooper Rivers were revisiting sites that were in use 100-500 years earlier and not utilizing sites visited for the first time by their Late Woodland contemporaries who made sand and grog tempered pottery. This implies that the makers of Wando pottery may have been more familiar with the local conditions and sought locales where they and their forebears found the resources they needed to survive. Perhaps, the makers of Wando pottery continued using the same settlement and subsistence strategies of their Middle Woodland predecessors rather than adopting newer systems associated with other Late Woodland groups. The continued use of check stamped pottery in the Wando series argues strongly for some connection to the Middle Woodland ceramic traditions if our earlier assumptions regarding the temporal association of check stamping are accurate. In conclusion, the recently defined Wando ceramic series appears to reflect a small, local population of Native Americans who possessed a distinctive technological assemblage. The presence of such an indicator provides an opportunity to examine this population, address specific questions about their adaptations and behaviors, and to examine in more detail the evolution of ceramic technologies within cultural groups. The discussions above are only the initial efforts to approach this problem and the makers of Wando pottery. Undoubtedly, we will discover other aspects, some supportive and possibly some that could discount our present assumptions, as future researchers examine the Wando ceramic series in greater detail. ACKNOWLEDGMENTS Originally presented at the 60 th Annual Southeastern Archaeological Conference Charlotte, North Carolina, 9-12 November 2003. Acknowledgments to Associated Developers, Inc., of Newport News, Virginia, and the Beach Company of Charleston, South Carolina, who sponsored the majority of the work reported in this paper. REFERENCES CITED Adams, Natalie and Michael Trinkley 1993 Archaeological Survey of the Seaside Farms Tract, Charleston County, South Carolina. Chicora Foundation Research Series 35. Columbia, South Carolina. 74
Bridgman, Kara 2001 A Technofunctional Analysis of Prehistoric Ceramics at 38CH949. Manuscript on file at Brockington and Associates, Inc., Mt. Pleasant, South Carolina. Dumont, de Montigny Lieutenant 1753 Memories historiques sur la Louisiane, contenant ce qui y est arrive de plus memorable depuis l annee 1868 jusqu a present. C.J.B. Bauche, Paris. Eubanks, Elsie I., James R. Hill III and Eric C. Poplin 1995 Archaeological Testing of 38CH1025 Parker Island Development Tract, Charleston County, South Carolina. Prepared for Joe Griffith Realty, Inc., Charleston, South Carolina. Fletcher, Joshua and Eric C. Poplin 2002 Archaeological Testing of 38BK1800 and 38BK1803/1804, Charleston Regional Business Center, Berkeley County, South Carolina. Prepared for Thomas and Hutton Engineering Company, Mt. Pleasant, South Carolina. Herbert, Joseph M., and Mark A. Mathis n.d. An Appraisal and Re-evaluaion of the Prehistoric Pottery Sequence of Southern Coastal North Carolina. In Indian Pottery of the Carolinas, edited by David G. Anderson, pp. 136-189. In preparation for the Council of South Carolina Professional Archaeologists, Columbia. Poplin, Eric C., Christopher T. Espenshade and David C. Jones 1993 Archaeological Investigations at the Buck Hall Site (38CH644), Francis Marion National Forest, South Carolina. Prepared for the US Department of Agriculture, Forest Service, Columbia, South Carolina. Poplin, Eric C., Kara Bridgman and Patrick Sieverts 2002 Archaeological Investigation of 38CH1025 at the Pointe at RiverTowne Country Club, Mount Pleasant, South Carolina. Prepared for Associated Developers, Inc., Newport News, Virginia. Poplin, Eric C., Todd A. McMakin, Patrick Morgan, and Patrick Sieverts n.d. Archaeological Data Recovery at 38BK1603, O.L. Thompson Mine, Berkeley County, South Carolina. In preparation for the Beach Company, Charleston, South Carolina. Trinkley, Michael 1989 An Archaeological Overview of the South Carolina Woodland Period: It=s the Same Old Riddle. In Studies in South Carolina Archaeology, edited by Albert C. Goodyear III and Glen T. Hanson, pp. 73-90. South Carolina Institute of Archaeology and Anthropology Anthropological Studies 9. Columbia. 1990 An Archaeological Context for the South Carolina Woodland Period. Chicora Foundation Research Series 22. Columbia, South Carolina. 75