Archaeological Excavations at Silver Bluff Plantation (38AK7)
Salem Asleep: A Discoursive Archaeology of God's Acre, 1771-1815
Bettis Academy and Junior College, SC National Heritage Corridor Site, Trenton
SRS Historic Land Plat Project
Mound Town Scale and Structure
With funding from the South Carolina Army National Guard at Ft. Jackson, two more AMS radiocarbon dates (formerly 11 dates) were obtained from the 133-cm long, 18 ka RCYBP sediment core recovered from the stream-head basin immediately downslope of 38RD628 near Colonel's Creek (Taylor et al. 2003; Brooks and Taylor 2004; Brooks and Taylor 2005; Brooks et al. 2005). The dates, which were on charcoal obtained immediately below two stratigraphic boundaries and complement dates obtained earlier from above those boundaries, help constrain better the timing of erosional and depositional events in the basin; the boundaries represent erosional unconformities, and the dates that bracket them indicate hiatuses, or gaps, in the lithostratigraphic record. In one instance, a hiatus of -300 radiocarbon years is indicated by dates of 3200 +/- 40 and 2910 +/- 40 RCYBP. More significant, a hiatus of -2200 radiocarbon years is indicated by dates of 8870 +/- 40 and 6600 +/- 40 RCYBP.
Because much of the mineral (i.e., sand) component of the stream-head basin fill was probably derived via slopewash from the 38RD628 landform immediately upslope (Brooks et al. 2005), work also was conducted at 38RD628 to link better the depositional and erosional processes across the two landforms. This was accomplished by reopening one of the units excavated during the mitigation of 38RD628 (Clement et al. 2005). The unit selected was one that had had a good archaeological sequence for age control; it was also one from which Brooks et al. (2005) obtained grain-size data to examine depositional processes across the site, and to use as an independent data set for reconstructing buried occupation surfaces (e.g.. Brooks and Sassaman 1990; Brooks et al. 1996; Brooks et al. 2001). With the unit reopened, Andrew Ivester (Department of Geosciences, University of West Georgia) described the profile from a soil science perspective and obtained six samples for OSL dating. The six dates range between 2.0 +/- 0.3 KaBP (Woodland Period) and 23.4 +/- 3.0 KaBP (120 cmbs-below the archaeology); the dates fit well with the archaeology, and are in proper chronostratigraphic order. The rather large errors at one standard deviation are of particular interest, indicating that the sediments were incompletely zeroed (i.e., not all grains were exposed .to sunlight) on deposition. Because complete zeroing would have occurred with eolian deposition, this finding tends to support the conclusion of Brooks et al. (2005), who interpreted slopewash as being the most likely mode of artifact burial at 38RD628. Based on site context, only colluvial/slopewash and eolian deposition were likely possibilities. Thus, with both the site and stream-head basin lithologic sequences being well dated, and their depositional processes reasonably well understood, it is now possible to correlate the two sequences, and the processes that formed them, with a fair degree of accuracy. A paper is being prepared for submission to Southeastern Geology.
In February 2006, Brooks and Taylor served as consultants to Steve Smith (South Carolina Institute of Archaeology and Anthropology, University of South Carolina) on research stemming from a Battlefield Protection Grant from the National Park Service. The research involved the collection of data necessary to formulate an Action Plan for the shoreline preservation of Civil War Archaeological Site 38CH1213, located on the north end of Folly Island, near Charleston, South Carolina. The site is undergoing rapid beach erosion. The geoarchaeological portion of the investigation during fieldwork was to assist in the interpretation of site limits and ongoing erosion processes. Field data used in interpretation included surface survey, mapping, and coring. Archival data to document changes in shoreline position included old maps, aerial and other photographs, and military documents. Reports of previous archaeological and geological (i.e., beach profiles) investigations were also consulted. Collectively, the data provide a clear picture of problems and prospects for site preservation; these will be presented in a forthcoming report by Smith.
At Mathis Pond in April, 2006, Brooks, Taylor, and Ivester served as consultants to Wendy Weaver, Brockington and Associates, for delineating site formation and depositional processes at 38AK862, a predominantly Early to Middle Archaic occupation on the eastern sand rim of the Carolina bay. Archaeological test units, block excavations, and backhoe trenches were inspected for stratigraphy. A continuous sediment column for grain-size analysis was extracted from the wall of one of the units with good archaeostratigraphy. Previous studies (e.g. Brooks and Sassaman 1990; Brooks et al. 1996) have demonstrated that examining up-column variation in grain-size is a sound method for locating buried surfaces in sandy sites with no visibly observable depositional or cultural stratigraphy. It becomes a particularly effective method when used in conjunction with archaeostratigraphy; in sandy sites, buried occupation surfaces tend to be represented by artifact density modes in vertical distributions. The modes occur at the tops (surfaces) of depositional events, usually defined by sediment fining-upward or coarsening-upward sequences, depending upon the depositional environment that produced them.
To document eolian deposition as the primary mode of artifact burial on the eastern sand rim of Mathis Pond, sediment samples were taken at a consistent depth of 40 cmbs (i.e., taken below the plow zone) from excavation units across the site. Assuming that the source for eolian sands were former, high energy, eastern shorelines (beach faces) during low water phases of the bay, and given that the dominant paleo-wind direction was from the west-southwest, then the sediments on the sand rim should fine to the east-northeast across the site. Sediment analyses by Ivester are in progress.
In May 2006, Brooks and Taylor consulted with Randy Daniel and Chris Moore (East Carolina University), and Keith Seramur (geologist with Appalachian State University) on four archaeological sites in the Tar River Valley near Greenville, North Carolina. The sites date to at least the Early Holocene, and are located on terrace margins above prominent escarpments with braidplain terraces below. The geomorphology, orientation, and sediments of the site landforms suggest formation, and artifact burial, by eolian deposition. Their location on the northeast side of the valley is consistent with the dominant paleo-wind direction being from the west-southwest (e.g... Carver and Brook 1989; Brooks et al. 2001). The wind would have plucked sand from the sparsely vegetated, wind-swept braidplains for eolian deposition onto the adjacent, higher terrace margins (e.g., Markewich and Markewich 1994). Deposition into standing vegetation (i.e., sediment baffling) is inferred from the geomorphology of the landforms; rather than exhibiting distinctive dunal morphologies, the highly localized, source-bordering sand bodies are amorphous and hummocky with downdrift lobes that thin rapidly and terminate abruptly.
The only other reasonable possibility for dominant mode of artifact burial at the Tar River sites is bioturbation after landform formation through eolian deposition, but this is ruled out by: 1) in situ artifact clusters (features); 2) polymodal, vertical artifact distributions indicative of a series of buried surfaces (e.g., Brooks and Sassaman 1990; Brooks et al. 1996); and, 3) at the Barber Creek Site, by a series of C-14 dates in proper chronostratigraphic order; the C-14 dates and archaeostratigraphy are in agreement. However, there is also an apparent problem with artifact burial by eolian deposition. According to Leigh (2006), the transition from braidplain to meanderbelt floodplain occurred -17,000-15,000 CalYBP. With the meanderbelt floodplain presumably becoming heavily vegetated during the transition, the braidplain, as a sediment source for eolian deposition, should have been effectively shutdown well before the formation of the archaeological site landforms. There is, however, a possible explanation that would accommodate burial by eolian deposition. Interestingly, all four sites have infilled paleochannels at their toe below the escarpment. In all four instances, it is apparent from the abandoned channel patterns that the long-term trend has been for the channel to "hug" these particular terrace margin locations. Briefly, it is hypothesized that an active channel existed at the toe of these terrace margin sites during occupation(s). This would account for both the attractiveness of these site locations to prehistoric populations (i.e., high ground overlooking an active channel), and for a mechanism promoting eolian deposition, whereby channel-related sands are exposed during low-water phases for eolian transport onto the adjacent landform. To test this hypothesis, the channels will be cored and dated to determine whether or not they were active during site occupations. The channel-related sands will also be compared to those deposited at the sites to see if they were likely sources. In addition, assuming that the site landforms formed in conjunction with eolian deposition, it is expected that grain-size analysis will indicate a series of discrete, buried surfaces whose locations will coincide with artifact density modes in vertical distributions (e.g... Brooks and Sassaman 1990). Downdrift, it is also expected that, in addition to thinning, site sediments will fine, and archaeological deposits will become more recent and shallower.
Finally, Keith Seramur conducted a microscopic analysis of quartz sand grains from one of the Tar River sites (Barber Creek site) and compared these with quartz sand grains from the adjacent braidplain. As expected, the water laid sand grains from the braidplain exhibited v-shaped pitting under magnification. In contrast, under low magnification, the site sand grains in aggregate looked like crushed glass; under high magnification, individual grains looked like miniature, multidirectional flake cores that had been reduced by grain-on-grain impact. The characteristics of the site sand grains were surprising in that conventional geological wisdom says that the surface texture of eolian sand grains should exhibit frosting and upturned plates. However, as Seramur notes (2006 personal comm.), that model is based on the Saharan desert, which represents very different climatic and depositional regimes. If verified by additional work, this could be a major breakthrough for those of us looking for a definitive sediment signature of eolian deposition; a signature whose attributes do not overlap with those of other depositional environments.
In November 2002, David Maki and Geoffrey Jones of Archaeo-Physics, LLC - Shallow Subsurface Geophysical Survey of Minneapolis, MN conducted a remote sensing survey of 0.98 hectare at the Galphin site (Maki 2004). The types of ground-based sensors employed during this survey included electrical resistance, magnetic field gradient, and ground-penetrating radar. In 2003 and 2004, multi-spectral data regarding the site was collected from aerial and space-based sensors. Information from each of the individual sensors was combined into a composite map. The SERDP team then identified a number of anomalies and attempted to assign each to an archaeological feature category. The task of archaeologists with the SRARP was to test the validity of these predictions-would the anomaly be present and would it correspond to the predicted archaeological feature category?
Fieldwork consisted of the excavation of 51 backhoe trenches (1 x 4 m), 20 shovel test pits (0.35 x 0.35 m), and 2 test units (1 x 0.5 m) yielding 2,002 artifacts. The majority of the artifacts date to the colonial period occupation of the site; however, a small number of prehistoric artifacts were also recovered. As a result of this fieldwork, the overall effectiveness of the remote sensing methods seems to be influenced by a variety of factors. In many cases, the excavations produced no visible differences between areas of positive and negative resistance. Alternatively, the ground-penetrating radar (GPR) and the magnetic field gradient surveys produced somewhat clearer results. Many of the, GPR anomalies were either identified as modem cultural features, such as plowscars, or natural undulations across the landscape rather than remnants of the defensive fortification that we were seeking to identify. Although each of the anomalies targeted as a result of the magnetic field gradient survey produced metal artifacts, here again a number of these were not colonial in origin but rather modem historic or wire pin flag fragments resulting from previous archaeological surveys conducted at the site.
Following analysis of the artifacts recovered during this survey and associated research, Robert Moon and I co-authored a technical report entitled Ground Truthing of a Multi-Sensor Remote Sensing Survey at the George Galphin Site (Herron and Moon 2006). We are in the process of analyzing artifacts from previous excavations at the site and attempting to sort through a mountain of associated documents collected by a Galphin descendant. This research will result in a comprehensive site report of the archaeological work conducted at the George Galphin site to date.
Today, Salem lies nestled within the heart of the modern-day city of Winston-Salem, where each year thousands of tourists and local schoolchildren visit the restored buildings of Old Salem and learn about its past as retold by costumed interpreters. Chartered in the 1950s, Old Salem's public presentation focused on the town's white, Moravian founders. However, by the 1990s, Old Salem, Inc. began to sponsor research that examined the contributions made by African Americans to the town's history (see Dowe 1994; Ferguson 1991, 1993; Ferguson and Hartley 1997; Hartley 1995; Hughes 2005; Palmer Gillies 1999; Sensbach 1998; Taylor 1993; and Ziegenbein 2002). The recent reconstruction of an 1823 log church for African Americans and the restoration of its descendant-an 1861 brick church-exemplify a new effort aimed at capitalizing on the growing market in African-American heritage tourism. Next to these structures lie Salem's Parish graveyard, a colonial graveyard for the interment of non-Moravians, white and black, and a later, segregated graveyard for African Americans.
Like daily life for Salem's residents, interment in Gottes Acker revolved around the choir system. According to the Moravian archivist C. Daniel Crews, choirs were: ...group[s] within a congregation consisting of individuals of like life circumstances [e.g. Children, Little Boys, Little Girls. Older Boys, Older Girls, Single Brethren, Single Sisters, Married People, Widowers, and Widows]...For many years the Single Brethren, the Single Sisters, and the Widows lived, labored, and worshipped together in close fellowship in their respective Choir houses (Crews 1996:5)
As a result, burial of the deceased in choir plots reinforced a gendered ideal of the proper Christian family. However, the role that this gender ideology played in creating both Gottes Acker and its subaltern, Salem's Parish Graveyard, remains largely unexplored. While Gottes Acker and Salem's Parish Graveyard were created for the interment of two different populations, lying as they do at the opposite end of Salem and looking radically different from one another, they trace their origin to the same period in time.
This thesis contends that far from being discrete and unconnected entities on the landscape, Gottes Acker and Salem's Parish Graveyard were and continue to be intimately connected spaces. Moreover, it argues that their mutual development provided for the existence of a segregated graveyard for African Americans after 1815. In effect, the historical connection between Gottes Acker and Salem's Parish Graveyard takes the form of a discourse in landscape.
The analysis of a material culture, archival and spatial data supports this assertion. These include an analysis of gravestones to create a linguistic typology that looks for patterns of similarity and difference between choir cohorts through time. Additionally, a number of primary documents concerning the church's policies regarding graveyards and mortuary practice are examined. A series of historical maps augments these archival sources. Finally, the recreated spatial layout and order of interment in Gottes Acker
Salem's mortuary discourse is not, however, simply a discourse confined to issues of location or proximity. Rather, a broader socio-political context: a context that revolves around issues of race, class, gender, and sexuality has and continues to construct both graveyards discursively. It includes the deployment of material culture and language to articulate a constantly negotiated ideology that emphasizes the familial bonds of a Christian community. The goal therefore, is to take two entities on the landscape, commonly treated as if they are disconnected, and reestablish their mutual dependence through the writing of a history that emphasizes a series of geo-social relations during their formative period (1766-1815). That said, chapters focus on the analysis of the following four discursive arenas: 1) the organization and placement of both graveyards on the landscape; 2) their aesthetic qualities; 3) the deployment of language on the landscape through the placement of gravestones; and 4) the control of access and ritual behavior within and between mortuary spaces.
With the help of descendent community members and aerial photos, the approximate locations of 24 buildings have been identified. However, the exact footprints of most of these structures are still unknown. To date, consultation has focused on conducting archival research and preliminary site mapping, as well as aerial photo rectification in preparation for future archaeological investigation resource assessment.
It is hoped that as the project continues to grow, it will provide an opportunity for comparative analysis with other contemporary African-American schools such, as those located in Dunbarton and Meyers Mill, historic towns of the SRS.
As Hally (1993, 1996) and Blitz (1999) have discussed, Lawton and Red Lake do not fit Hally's expectations. While they are located only a few kilometers apart, neither is larger or has indications of serving more important political and religious functions. Therefore, there is no clear primary and secondary center in this relationship. Jared Wood's (2005) discovery of another small mound town within a few kilometers of both Red Lake and Lawton, still further complicates the political picture. The site, known as Spring Lake, also does not appear to have served a larger number of important functions than the other nearby mound towns.
Blitz (1999) has suggested that archaeological examples like this one may represent a social form that is not hierarchically organized, as Hally's model assumes. Instead, using ethnohistoric data on Creek communities. Blitz argues that closely spaced mound towns like Red Lake, Lawton, and Spring Lake may be centers in a confederacy where there are no hierarchical political relationships among the centers.
When we reconstruct the structure of both Lawton and Red Lake from the work that has been conducted so far, the two towns appear to be completely different kinds of centers. Lawton has two mounds and a well-defined residential zone enclosed within a ditch and palisade wall. Red Lake has three small mounds, spread over a much larger area, and has no evidence for a ditch encircling the immediate vicinity of the mounds. However, a closer comparison shows that these mound centers are very similar both in terms of scale and structure. Each town is comprised of a core area of two or more mounds arranged around an open space that we have interpreted to be plazas. Ringing the core is an area of midden that we interpret as the domestic occupation zone. When you superimpose Red Lake over Lawton, it becomes clear that these core areas and associated dense occupation zones are about the same size.
Two things appear different, however. First, the occupied area of Red Lake seems larger than that of Lawton. This is an artifact of the work done at each site. In working at Lawton, we chose to focus our energy on the area inside the ditch. Mississippian occupation continues outside of that ditch along the slough edge in both directions. We have not tested enough to find the actual extent of that occupation, but it is clearly Hollywood phase.
The only really clear difference between these two towns is that the core area at Lawton is set off from the rest of the site by the ditch and presumed palisade wall. In a sense, we have what we think is a sacred precinct surrounded by and separated from the rest of the community. Although there is no ditch at Red Lake, we have not yet conducted the work to determine if its sacred precinct is separated from the rest of the community by a wall.
Jared Wood's (2006, pers. comm.) work at the nearby Spring Lake site seems to indicate that the mound complex at that site was on about the same scale as Lawton and Red Lake and laid out in a similar fashion. The site has a single mound adjacent to a small plaza, all surrounded by dense occupation that continues up and down the adjacent waterway. Pottery recovered by Wood indicates that Spring Lake dates to the Hollywood phase, as well.
These similarities in scale and structure bring us back to questions about the political relationships among these sites. Considering only scale and structure, these sites do not appear significantly different. It is true that Red Lake has three mounds, Lawton two, and Spring Lake one, suggesting a potential hierarchy. However, it is not clear how each mound functioned, so it is not possible to understand if the number of mounds translated in any way into religious or administrative differences that would be indicative of an administrative hierarchy. We have found evidence at Lawton that each of the mounds served different functions: one was residential for elites and the other ceremonial (Stephenson and King 2004). Unfortunately, we do not have the information to compare Lawton to Red Lake mound function at this time.
At Lawton, the civic-ceremonial core of the site is set apart by a wall and ditch. It is possible that the labor invested in these features suggests that Lawton's civic-ceremonial space had a greater importance than civic-ceremonial space at other sites. Again, it is possible that the mound and plaza cores of Red Lake and Spring Lake also are separated from the rest of the occupation. The work simply has not been done to evaluate that possibility.
There are obviously many questions to answer before it is possible to clearly understand the political relationship among these sites. However, the layout and distribution of the mound towns in the middle Savannah River valley may provide some insights into those relationships. The settlement system in this part of the Savannah River valley, as far as we understand it, is reminiscent of the talwa or town systems of the Creek Indians, particularly of the 17th and 18th centuries. Creek communities had a sacred core that included a winter council house, a square ground where summer councils and important ceremonies took place, a ball pole and field, and sometimes one or more small mounds (Howard 1968; Knight 1994). This civic-ceremonial space was surrounded by households that often were strung up and down the major drainages on which the towns were located (Etheridge 2004; Knight 1994; Worth 2000). This is a classic dispersed settlement system. Following Bruce Smith's (1978) arguments for Mississippian settlement in the Mississippi River valley, the middle Savannah River valley makes ecological and social sense in the ridge and swale floodplain setting of the Coastal Plain.
In many instances, Creek towns were scattered up and down major drainages, sometimes close enough to one another that the scatter of households for one town butted up against the households associated with a nearby town. Town affiliation was important among the Creeks, and although some towns were more important than others, each had a great deal of autonomy (Knight 1994; Saunt 1999).
It is possible that what we are looking at in the middle Savannah River valley is a social system not unlike the Creeks of the 18th century. There was social ranking, but it was embedded within the clan system and expressed in terms like "older brother" or as part of the Red and White symbolism of war and peace (King 2003). In these systems, there were some communities that everyone recognized as more important, and there was a certain deference shown to those communities and their leaders. However, individual towns were largely politically independent (Knight 1994; Saunt 1999) y/e think it is fair to call these entities chiefdoms (King 1999). There were certainly social and political relationships among them, but there was no overarching hierarchy linking them all into one social system.