Colonial Williamsburg Digital Library

Physical /Structural Evidence

There has been some concern that these pits do not represent root cellars, since no trace of structural remains survives in association with these features. This absence, however, should not be taken as proof that there were no structures over these features, based on other evidence from the site itself, and from other Tidewater sites. First, let me deal with the lack of structural remains. Given the amount of disturbance caused by plowing on the site, the fact that only the root cellars remained intact is not surprising. For example, there was no evidence of a hearth at the Pettus quarter at Kingsmill, since, as Bill Kelso has pointed out, all traces of a stick and mud chimney could easily be destroyed by plowing (1984:104). Additionally, if the structures were of wooden sill construction or built with a shallow brick foundation (either continuous or pier), all traces could have also been destroyed by plowing. To use another example from Kingsmill, the North Quarter site, containing two root cellars, displayed only a small portion of a robbed foundation wall trench (Kelso 1984:124). Thus, it is not inconceivable that no tangible structural evidence would have remained after substantial plowing in the area.

Additionally, the Carter's Grove pit features share similar sizes and configurations to those of other area root cellars. The cellars at Carter's Grove ranged in size from 2' 8" square to 9' x 6', while the root cellars at the Kingsmill Quarter (where 18 2 cellars were located in all) measured between 2'9" x 2'0" to 5' x 8' (Kelso 1984:120). Those at Monticello's Mulberry Row ranged between 2' x' 3' to 4' x 6'. The Kingsmill Quarter root cellars were from 2' to 3'6" deep below the plowzone, while the Carter's Grove pits ranged from 4" to 3'6" in depth. Also, some root cellars at the Kingsmill plantations, Monticello, and other sites have displayed internal structural/architectural components, such as wooden linings, floors, and partitions (Kelso 1984:105, 107), similar to those seen in Pits A and B at Carter's Grove.

The presence of a trash filled ravine to the east of the root cellar features, and the location of what appears to be a colonial well suggest that the site was serving as a domestic area during the late 18th century. The well was located at too great a distance to be of use to the main house, and although the ravine trash could conceivably be associated with the main house, it is more likely that such garbage would be discarded in ravines closer to the kitchen, which was located on the east side of the mansion.

Stratigraphic Evidence

What clues can the stratigraphic evidence in the features, as well as the mends and crossmends, both within and between the pits provide? How does the Carter's Grove evidence compare with that of other root cellars?

There were crossmends among five of the Carter's Grove root cellars and many of the pits contained mends between their various layers. Those pits which display crossmends were open and being filled simultaneously. Kelso states that many of the Kingsmill cellars were filled with trash at the time of or just before the destruction of their associated building, since rubble and mortar layers of the uppermost fill contained artifacts which crossmended with lower layers (Kelso 1984:120). Similar fill (containing brick and mortar debris) and crossmends between the upper and lower layers of the some of Carter's Grove features suggests that the features were filled quickly and intentionally, possibly at the time of a building's destruction.

From the large number of pits found under single structures (18 at Kingsmill Quarter), it has also been suggested that some pits were created, used, filled, and capped in sequence (Kelso 1984:120), possibly during the lifetime of the building. New pits were then excavated to take the place of those recently filled. Later pits are actually sees cutting earlier ones, similar to Pit E at Carter's Grove, (which appears to be three separate root cellars) and Pits J and K, which adjoin. Additional proof that root cellars were excavated, used and filled during the lifetime of a building comes from the Brush-Everard Kitchen root cellar. The cellar was constructed after the kitchen was built, and since the kitchen is as original building, was filled while the building was still standing.

Artifacts from the root cellars appear to consist primarily of secondary refuse, defined here as object discarded at a place different from their location of use (Schiffer 1972:161). The ceramics and glass from the root cellars were of diverse types, generally fragmentary in nature, and except in a few instances, each vessel was represented by only one or two sherds. These characteristics are consistent with those of secondary fill (Wise 1976). The root cellar assemblages seem to typify what Stanley South has labeled "adjacent secondary refuse" (1977:48), which accumulates in yards and is dispersed through a variety of human, animal and environmental activities.

A small test trench through the ravine (CGER 717) yielded a crossmend with one root cellar (CGER 643 - Pit L), indicating that the ravine was probably the primary area of trash disposal for the quarters. Here, debris from the quarters would be discarded, and this refuse would contain the majority of each broken ceramic or glass vessel. (The assemblage from the ravine contains large pieces of glass and ceramic with numerous mends, typical of a primary trash disposal area.) Smaller fragments, not swept up or discarded initially with the original vessel, would eventually find their way into the yard, under the quarter, or perhaps into an open root cellar by falling through the trap door or cracks in the floorboards. Intentional filling of the root cellars, either during the life of the building, or directly after its destruction, could easily make use of the debris swept up from the quarter yard and under the buildings. Such clean-up behavior was seen at Tazewell Hall in a filled drainage ditch and a backfilled cellar. Much of the material in the root cellars is broken into small pieces and the bone shows signs of rodent gnawing and weathering. These would be typical characteristics of material which had been yard scatter. Since there is at least one ceramic mend with ravine debris, it, seems likely that the root cellar material was generated through activity on the site.

It is also likely that some of the artifacts represent objects deliberately placed in the cellars by the slaves, or objects discarded directly into the cellars. Of the thirteen Carter's Grove pits, profile drawings of only three of these have been recovered to date. Pits A and B (CGER 715 and 716) showed evidence of thin layers of sand or silt at their bottoms, with the sand suggesting their possible use as food storage pits. These features had also remained open long enough to become partially filled with silt prior to their deliberate filling in the late 18th century. The silted layers in Pit A (CGER 715C and D) contained numerous large artifacts, including a saddle tree, an iron padlock and key, a scythe fragment and a possible gridiron handle, in addition to refined and coarse earthenware ceramics. Objects such as these, which were largely unbroken, may have been stored in the root cellars.

Artifact Evidence

After arguing that the Carter's Grove features do appear, at least physically, to conform to other excavated root cellars, then the question remains of analyzing the artifacts for their representativeness of the occupants' household possessions. Are the artifacts from the Carter's Grove features similar in content to those of other root cellars?

To attempt a full-scale analysis and comparison of all the artifact assemblages from excavated root cellars in Virginia and the surrounding area is an immense task, the possibility of which is currently being looked into by the Department of Archaeological Research for future study. For the purposes of the Carter's Grove Slave Quarters Furnishings Committee's needs, a few generalized conclusions can be drawn from the information gathered so far.

To date, the Carter's Grove root cellar assemblages have been catalogued using the Department of Archaeological Research artifact inventory in dBase III +. Raw counts of all items (see the DAR inventory and the attached memo listing problems in resolving the Noel Hume inventory with the actual assemblages) are available, but no formal minimum vessel count of the ceramics and glass has been accomplished.

Therefore, analysis for this report will be based on the sherd or fragment (as opposed to the object) level.

As far as other sources of information on root cellars and slave sites in Virginia and elsewhere, Appendix A shows the material which I have been able to gather, and that which has been requested but not yet received. Generalized conclusions concerning root cellars as an artifact indicative of Afro-American behavior have been drawn from a less-than-thorough perusal of these documents.

The artifact assemblages recovered from the Carter's Grove root cellars are domestic in origin. Treating the artifacts recovered from all of the pits as one unit shows the following functional breakdown:

The composition of the assemblage does not reflect the quality of household possessions which the Burwell family would have owned at the end of the 18th century. The predominant ceramic is undecorated creamware, with earlier forms of tableware, such as white salt glazed stoneware and delftware also common. Porcelain makes up 11% of the refined tablewares and this presence would appear to go against the theory that these items were slave possessions. It should be noted here that porcelain has been found within root cellars at Monticello (Kelso 1986), Kingsmill (Kelso 1984:205), Mount Vernon (Outlaw 1985), and on slave sites in South Carolina (Wheaton et. al.). William Pittman examined the porcelain from the features, however, and in his opinion the vessels, most dating to the mid-18th century, were of mediocre quality. There was one fragment (saucer rim) whose manufacture date fell around 1720.

Combining the ceramics, bottle glass, table glass, and cutlery groups gives a total of 79% kitchen or food and beverage related items. This high percentage is consistent with that shown by Wheaton et. al. as indicative of slave sites in South Carolina (1983:285). When compared with Stanley South's Revised Carolina Artifact Pattern for British colonial sites (1977), where kitchen related items generally comprise 59% of all 5 artifacts, and architectural items, 27%, it can be seen that there are significant differences between the figures associated with slaves and those with British colonial sites. Patterning within the Carter's Grove artifacts is more consistent with the slave pattern of Wheaton et. al., which shows 77% kitchen and 18% architectural category medians. The low architectural percentage of the Carter's Grove site (and the resultant higher kitchen ratio) is easily explained by the fact that all structural evidence of the quarters had been destroyed by plowing, and the subsequent removal of the plowzone prior to archaeological investigation.

Lumping the ceramics into one group for all of the Carter's Grove features shows the following breakdowns:

The presence of English-made ceramics within a supposed slave assemblage is not surprising. Root cellars throughout the area have largely contained ceramics of English and European manufacture. Even on slave quarters sites in the deep south, where slaves were more isolated from the Anglo-American culture, English ceramics are commonly found. For example, at the Yaughan and Curriboo slave house sites in South Carolina, excavations showed that non-local ceramics comprised between 7.2 and 15.3% of the total artifact assemblages (Wheaton et. al. 1983:333). English made ceramics were also found on slaves sites at Cannon's Point Plantation (Otto 1984), on St. Simon's Island, Georgia (Moore 1981), and Butler Island, Georgia (Singleton 1980). Colonowares, supposedly of slave origin, were not well represented is the Carter's Grove assemblage. The only two decorated specimens of colonoware recovered in Williamsburg, however, were from the root cellars.

When the tableware ceramics are broken down into types, the following percentages are shown: 6

Most of the ceramics within the root cellar assemblage show signs of unusually heavy usage. The creamware plates in particular are heavily stained and scratched, and in some places the glaze has been almost completely worn away. The same heavy wear is evident on many of the coarse earthenware vessels, and even on English white salt-glazed stoneware, which is a more durable ceramic type than any of the coarse or refined earthenwares.

Although there was some whiteware (with a terminus post quern of 1820) contained within the root cellar assemblages, these sherds were treated in the Carter's Grove excavation report as contamination. Thus, the general dating of the root cellar assemblages based on the ceramic type of most recent manufacture (pearlware of varying forms of decoration with beginning dates of 1779-1790), indicating that the features were filled at the end of the 18th century. For this report, the whiteware (a total of 4 fragments, all of which were found in the uppermost layers of the cellars) has also been treated as contamination. It should be pointed out, however, that if these pits were actually filled after 1820, then the majority of the ceramics being discarded in these cellars had been manufactured over 50 years previously.

Several pipestems from the root cellars had been reworked using a knife to make a mouthpiece from a broken stem end. In one example, the stem had been whittled into a mouthpiece approximately 2" from the bowl of the pipe. Since English clay pipes were an inexpensive items in the late 18th century, such frugal behavior would seem indicative of a person without ready access to either cash or new pipes. This particular artifact also brings to mind the description by Ferdinand-Marie Bayard of the interior of a Virginia slave cabin, which includes the statement "An old pipe, very short and a knife blade, which were sticking in the wall were the only effects that I found in the dwelling" (as cited in Vlach 1987:11).

Another trait which the Carter's Grove root cellars seem to share with the Kingsmill, Mount Vernon and Monticello root cellars is the large number of buttons recovered from their fills. The Carter's Grove features contained 29 buttons, largely undecorated copper alloy examples. For contrast, a domestic assemblage from a trash pit at the Dr. Barraud House dating to the late 18th century, contained one button, and, with less than half of the ceramics vesselized to date, 63 ceramic vessels so far. 7 A rough ceramic vessel count for the root cellars was estimated by Kathleen Pepper as falling between 45 and 60 vessels. with plates and bowls forming the most well represented vessel forms. Vlach discusses clothing as "...essential to a slave's sense of identity since it was his most permanent and personal possession" (Vlach 1987:4). Bill Kelso believes that the large number of buttons recovered from root cellars may be related to slaves using old clothing in the construction of quilts (Kelso 1984:202).

Potential Problems

There are some possible problems in using the Carter's Grove material for comparison with artifact assemblages from other root cellars as a means of assessing the Carter's Grove assemblage typicality. These include:

• 1.The level of data recovery for the Carter's Grove assemblages is not known. Mr. Noel Hume did not normally screen the soil from his excavations, which will make it difficult to compare with some of the other root cellar assemblages, particularly in terms of small faunal and botanical remains.
• 2.The amounts of bone which were included with the Carter's Grove root cellar artifacts does not begin to match the quantities of bone recovered from other area root cellars. The original inventory mentions the presence of bone in assemblages which cannot at present be located.

Conclusion

When compared with other excavated examples of root cellars, the Carter's Grove material seems typical. Despite beliefs that slaves did not have refined ceramic tablewares, or what might be thought of as luxury items (mirrors or jewelry, for example), artifact assemblages associated with slaves show these ideas to be false. In a culture heavily influenced and controlled by Anglo-Americans, the use of imported items by slaves would be expected. How slaves were obtaining refined English ceramics for their tables may never be known, but an interesting question that does seem answerable is the age of ceramics used in slave households. To accomplish this, a site with documented beginning and end dates is needed. At least one such assemblage appears to exist.

Dennis Pogue, archaeologist for The Mount Vernon Ladies Association, has contacted our office about an assemblage excavated several year: ago on the documented site of the quarters of the household servants at Mt. Vernon. These were slaves Martha Custis Washington brought from Williamsburg upon her marriage, and the quarters themselves are documented as having been destroyed in 1793 (Outlaw 1985). A good beginning date on these quarters has been established at around 1760, and with the good terminal date, important question about whether slaves were using hand-me-down or out of fashion ceramics and other possessions can be addressed. Our office has received a copy of the inventory of the root cellar assemblage, and the detail in this inventory is very good. There are also over 25,000 bones and other food remains contained within the assemblage, which is a good sample size for making some statements about slave diet.

Although this is far from a complete and definitive study, I hope this will answer some of the committee's questions, or at least spark some lively debate. I will be glad to answer questions about any of my statements, and share any of the material which I have received on other slave sites. The recently completed inventory of the Carter's Grove artifacts is included with this memo.

BIBLIOGRAPHY

Kelso, William

1971

A Report on Exploratory Excavations at Carter's Grove Plantation James City County, Virginia (June 1970 - September 1971). Edited, and with additional data on continued excavations September 1972 - March 1972 by Neil Frank. Colonial Williamsburg Foundation Research Report No. 111.

1984

Kingsmill Plantations 1619 - 1800: Archaeology of Country Life in Colonial Virginia. Academic Press, New York.

1986

The Archaeology of Slave Life at Thomas Jefferson's Monticello: "A Wolf by the Ears". Journal of New World Archaeology. Volume VI, Number 4. pp. 5020.

Moore, Sue Mullins

1981

The Antebellum Barrier Island Plantation-In Search of an Archeological Pattern. University Microfilms, Ann Arbor.

Otto, John Solomon

1984

Cannon's Point Plantation 1794 - 1860: Living Conditions and Status Patterns in the Old South. New York, Academic Press.

Outlaw, Alain

1985

Mt. Vernon Archaeology Interim Report on the Root Cellar (44Fx762/40), July 1985. Unpublished report on file at the Virginia Division of Historic Landmarks, Richmond, VA.

Schiffer, Michael B.

1972

Archaeological Context and Systemic Context. American Antiquity. 37:156-65.

Singleton, Theresa

1980

The Archaeology of Afro-American Slavery in Coastal Georgia: A Regional Perception of Slave Household and Community Patterns. University Microfilms, Ann Arbor.

South, Stanley

1977

Method and Theory in Historical Archeology. Academic Press, New York.

Vlach, John Michael

1987

Afro-American Domestic Artifacts in Eighteenth-Century Virginia. Material Culture. Volume 19, No. 1. pp. 3-24.

Wheaton, Thomas R., Amy Friedlander, and Patrick Garrow

1983

Yaughan and Curriboo Plantation: Studies in Afro-American Archaeology. Report prepared by Soil Systems, Inc. Marietta, GA.

Wise, Cara 10

1976

Date and Status in Eighteenth Century Delaware: An Archaeologist's View. Transactions of the Delaware Academy of Science. 1974 and 1975. Delaware Academy of Science, Newark, Delaware.

INFORMATION ON SLAVE SITES AVAILABLE AT DAR

Kingsmill Plantations Root Cellars

• Excavation Notes - on file at DAR
• Finds Lists - on file at DAR

Governor's Land Root Cellars

• Excavation notes and finds lists have been requested but not yet received.

Mount Vernon House for Families Cellar

• Finds List - on file at DAR
• Interim Report - on file at DAR

Monticello Mulberry Row

• Finds list - on file at DAR