Colonial Williamsburg Digital Library

Rich Neck in the Seventeenth and Eighteenth Centuries

The site is located on a piece of property that was known as "Rich Neck" in the early seventeenth century. Rich Neck, and plantations like it, made up the community of Middle Plantation. This loose cluster of settlements would ultimately become the city of Williamsburg, and for a time, Virginia's capital (McFaden and Muraca 1994:9). But at the time of Rich Neck's initial European settlement, it was still the frontier, approximately eight kilometers from the first capital, Jamestown.

George Menefie, Council member and merchant, held the first patent on Rich Neck, which he conveyed to Richard Kemp, Secretary of the Virginia colony, in 1636. Kemp's residence at Rich Neck was excavated in the mid-1990s (site state number 44WB52). On Kemp's death in 1650, his widow, Elizabeth Wormely Kemp, married Thomas Lunsford. Records seem to indicate that by 1653 Elizabeth Wormely Kemp Lunsford was again widowed (McFaden et al. 1994:12). In 1665, the property was sold to Thomas Ludwell.

Like Richard Kemp, Thomas Ludwell was Secretary of State for the colony. And although he owned extensive tracts of land in several counties, he maintained Rich Neck as his residence. Archaeological excavations show that he enlarged and added onto the country seat built by Kemp (McFaden et al. 1994:14). Among the people transported to the colony by Thomas Ludwell was his younger brother Philip, who also rose to prominence, becoming Deputy Secretary of State by 1676/7. No doubt their cousin, Governor Sir William Berkeley, had considerable affect on their fortunes in Virginia—not the least of which was Philip Ludwell's subsequent marriage to his cousin's widow, the Lady Frances Culpeper Berkeley, and the acquisition of the Berkeley plantation at Green Spring.

The Ludwells divided themselves between the plantations, with Philip Ludwell's son by a previous marriage, Philip Ludwell II, making his home at Rich Neck for a number of years. Archaeological investigations show that his new plantation house was built with materials salvaged from the previous structures (McFaden and Muraca 1994:15). Philip II was certainly living there by 1698, as his first child was born at Rich Neck in that year. His remaining children were born at Rich Neck in 1701 and 1704 and at Green Spring in 1706 and 1716. Therefore, it is thought that Philip Ludwell II moved away from Rich Neck between 1704 and 1706 (Smith 1993:2). When Philip I died in England ca. 1717, Philip II inherited the property, but Rich Neck was never again the family's primary residence.

Philip Ludwell II continued the family tradition of political prominence, holding positions of leadership in the government as well as in the parish, at the College of William & Mary, and in the local militia. His son, Philip Ludwell III, inherited Rich Neck in 1726/7, while still a minor. Philip III's children were all born at Green Spring between 1737 and 1751, reinforcing the idea that Rich Neck was not the family's home. It seems that plantation 4 managers were in charge of operations on the property from 1706 forward, although there is no direct mention of such persons until much later in the eighteenth century.

Following in his forefathers' footsteps, Philip Ludwell III participated in Virginia's political life. He was a member of the House of Burgesses and sat on the Governor's Council. By 1760, however, Philip III had left the colony a widower, taking his children with him. He died in England in 1767 (Smith 1993:5). Rich Neck was ultimately divided between his two surviving children, Hannah Philipa Ludwell Lee, and Lucy Ludwell Paradise (Virginia Magazine of History and Biography [VMHB] 1913,1924).

Before Philip III left for England, he hired plantation manager Cary Wilkinson to oversee his Virginia holdings. Lucy Ludwell Paradise and her husband John Paradise, owners of the core of the Rich Neck holdings, retained his services after Philip III's death. Twenty-one enslaved individuals—10 men, 5 women, 3 boys, and 3 girls—were living at Rich Neck when Philip III's estate was inventoried. They were apparently tending livestock, and perhaps growing tobacco and running a mill, based on the tools and supplies listed in the same inventory (VMHB 1913:401). These people were presumably under the indirect supervision of Wilkinson, who was based at Green Spring (Franklin 1997:67).

William Wilkinson took over for his uncle in 1783 and stayed on until 1800. Lucy Ludwell Paradise, the last Ludwell to reside at Rich Neck, finally returned to Virginia in 1805, dividing her time between that estate and a house in Williamsburg.

Rich Neck Today

Site 68AP is located in the City of Williamsburg, in the eastern part of the state of Virginia (Figure 2). The site is near the intersection of Jamestown Road and State Route 199 (Figure 3), not far from College Creek, a tributary of the James River, which in turn drains into the Chesapeake Bay. The area is characterized by steep ridges interspersed with broader upland terraces (McFaden et al. 1993:3). The site is located on one such terrace.

The species of plants and animals found near the site today reflect many years of human occupation and use. As part of a plantation, the area was heavily farmed from the seventeenth through the nineteenth centuries. This is corroborated by the presence of a plowzone encountered archaeologically. The early to mid-twentieth century saw some logging on the property, as may be observed from the growth pattern of the trees found today (McFaden et al.1994:3).

Currently, the site is surrounded by a suburban housing development. Wild flora include hardwoods such as oaks and holly, as well as evergreens like loblolly pine. A variety of wild and commensal animal species inhabit the periphery of the development including deer, raccoon, squirrel, and various waterfowl. College Creek is still home to a variety of fish and several species of turtle. The seventeenth-and eighteenth- century residents of Rich Neck may have exploited any or all of these species.

5

Figure 2. General project location.

Figure 3. Project location within the City of Williamsburg.

6

Excavation Strategy

As stated in the introduction, site 68AP was first identified while testing the perimeter of the Rich Neck Slave Quarter. The 1994 shovel testing campaign consisted of 75 centimeter units placed every five meters, yielding a plowzone sample of slightly more than 2%. A shovel test at the Rich Neck grid point 565 N/465 E came down on dark brown feature fill which covered the entire unit (Figure 4). Interestingly, this feature was not predicted by analyses that examined concentrations of artifacts in the plowzone. Likewise, areas with high concentrations of plowzone artifacts did not always overlie significant features (Agbe-Davies 1994:5). When archaeologists expanded the shovel test to a 2 x 2 meter unit, one edge of the feature was exposed. An auger test showed fill extending at least one meter below the interface between the plowzone and natural subsoil.

Figure 4. Rich Neck Slave Quarter and associated shovel tests.

8

Figure 5. Overall site photograph.

In March 1995, a small area around the feature was machine stripped, indicating that it was approximately 4 x 4 meters in size. Close by, excavators noticed a second, smaller feature, which looked like a small root cellar or trash pit, hereafter designated master context 3 (M3). At this time excavators mapped the two features and sank a 1 x 1 meter test unit in to the southeast corner of the larger feature (Feature 1). The soil was saturated, so excavators had difficulty distinguishing stratigraphic levels. To compensate, they assigned arbitrary levels every 10 centimeters, stopping after encountering clay approximately 70 centimeters from the top of the surrounding subsoil.

Later that year, excavation began with the intention of recovering all significant archaeological material in anticipation of the lot's eventual development. A machine was used to clear an even larger area of plowzone, ultimately exposing an area approximately 15 by 13 meters. The intent was to find other features associated with the two primary features (Feature 1 and M3), presumed to be either cellars or trash pits. After machine stripping, the last few centimeters of plowzone were shovel-shaved and trowelled down (Figure 5).

Except for six shovel tests and the 2 x 2 meter area over the cellar, the plowzone was not screened. However, all artifacts found in the course of removing the plowzone were saved as context 68AP00000, or in the case of the plowzone immediately above Feature 1, context 68AP00030.

The features were all excavated with trowels and screened through ¼-inch mesh. Chemical and flotation samples were also collected from every feature or layer of fill. The exceptions to this sampling rule were tree holes and several episodes of subsoil erosion, which were detected in Feature 1. From these, only chemical samples were taken. Furthermore, in the case of the erosion layers in Feature 1, the fill was trowelled, but not screened.

9

A number of other samples were taken as well. Oyster shell was particularly abundant—over 300 specimens came from master context M4 alone (M4 being the first layer in Feature 1). Samples of brick were collected in anticipation of acid extraction and xeroradiography studies (for example, Hoff 1994; Galucci et al. 1994). Finally, a 25 x 25 centimeter column sample collected with very tight stratigraphic and elevation controls, was taken out of Feature 1 for flotation and the subsequent analysis of botanical remains. These samples remain at the DAR. They were not analyzed at the time of writing.

Excavation began with the various small features found scattered across the site (Figure 6). It was anticipated that some of them might be postholes indicating the presence of a building in the project area. The next focus was M3, the possible root cellar or trash pit. It was sectioned and taken out in two halves, 68AP00003 and 68AP00032. Feature 1 was the final feature to be excavated. This strategy was followed to allow inexperienced field school students the maximum exposure to archaeological fieldwork before having to cope with what was believed, quite correctly, to be a stratigraphically complicated feature.

Feature 1 was divided into quadrants for excavation. The southeast and northwest quadrants were removed first to allow archaeologists to observe complete sections across the length and breadth of the feature. Next, the northeast quadrant was removed. The southwest quadrant was not excavated. Each quadrant was excavated following the layers (contexts) observed within each quadrant. These layers were then correlated into master contexts which represented all contexts from the same deposit, regardless of which Figure 6. Plan of all features. 10 quadrant contained them. Like all the compass orientations in this report, the directions refer to the site grid, grid-north being 44 degrees east of true north.

Features and layers were distinguished from one another by differences in soil color, texture and inclusions. Each distinct deposit was treated as a separate entity and given a unique context number. Information about its location, description and contents were recorded on context forms which are currently on file with the DAR. Stratigraphic relationships between these contexts were recorded using the Harris Matrix system (Harris 1989).

All features were drawn in plan view and larger features like Feature 1 and M3 were also drawn in section. These drawings were then incorporated into AutoCAD computer generated drawings of the entire Rich Neck complex. Photographs of the entire site, as well as detailed photographs of individual features, were taken and remain on file at CW's DAR.

Artifacts recovered were bagged according to the context from which they came. Laboratory technicians washed, labeled and identified the artifacts, and determined which artifacts were fragile and in need of conservation. The artifact inventory was entered into a computer using the program Re:discovery, which facilitates the integration of excavation data with laboratory data. Artifacts from the flotation samples were not catalogued at the time of writing. Ceramics were pulled to be cross-mended and create minimum vessel counts (see "Artifacts" below). A sample of the identifiable bones was also pulled for analysis.

Description of Strata and Features

Testing prior to excavation demonstrated that plowing had obliterated any occupation layers. As expected, there was no intact stratification outside of feature fill. Under the leaf litter and topsoil, a plowzone overlay the entire site with an average thickness of about 18 centimeters. The brown (Munsell color 10 YR 4/3) sandy loam contained artifacts primarily from the eighteenth century. The features cut into a pale brown (10 YR 5/8) sandy clay, which comprises the subsoil in this area.

None of the numerous small features exposed by the plowzone removal proved to be regularly shaped or arranged (Figure 6). Most of them were quite shallow, but ranged from 3 to 35 centimeters. Almost all of them terminated in irregular shapes, suggesting that they were tree holes rather than postholes. Artifacts were very sparse in these features, mostly fragments of wine bottle glass and nails, and a few ceramics including Yorktown-type coarse earthenware and white salt-glazed stoneware.

M3 was a sub-rectangular feature filled with an olive brown (2.5 Y 4/3) sandy clay loam to a depth of 21 centimeters (Figure 7). At its greatest extent, it measured 1.64 by 1.23 meters (Figure 8). Observed from ground level, it does not appear to be perfectly aligned with Feature 1; however, this does not preclude them being contemporaries. M3 contained artifacts not unlike those found in Feature 1. The two contexts that make up M3—68AP00003 and 68AP00032—are the north and south halves of the same feature fill. One chemical and one flotation sample were taken from this feature, which has a terminus post quem (tpq) of 1755.

11

Figure 7. Section of M3.

Figure 8. Plan of M3.

This feature is clearly not a borrow pit dug to extract dirt or clay, nor a pit dug merely to serve as a repository for garbage. The sides of the pit are very regular and represent careful digging of the hole by its eighteenth-century excavator. Furthermore, it must have been covered during its use-life, or filled immediately after being dug, as there was no silt or other evidence of erosion at the bottom of the pit. At the base of M3, there was a burned area in the subsoil. However, it was quite small and did not extend very deep into the clay, nor did the fill contain any great quantity of charcoal or ash. It does not seem, then, to have been the site of a major burning episode.

Approximately 3 meters to the northwest of M3 is another, much larger, feature. Feature 1 has been designated as a cellar. It measures 5.22 by 4.1 meters at its maximum extent. Its sides are nearly vertical, with the exception of what might be interpreted as a bulkhead entrance, which comprises approximately 2 meters of the cellar's maximum horizontal length (Figure 9). Four major fill episodes were observed in Feature 1, as well as numerous thin deposits related to erosion or exposure to the elements. Its total depth was 1.3 meters and consisted of 12 discrete layers, some of which were small lenses which do not appear in the profile drawings (Figures 9 and 10).

12

Figure 9. Section of Feature 1 running east-west.

Figure 10. Section of Feature 1 running north-south.

The largest and most artifact rich of the master contexts was M4, the uppermost layer in Feature 1. This fill was a dark yellowish brown (10 YR 4/4) clay loam with scattered brick, mortar and charcoal inclusions. Large amounts of shell and bone were found throughout. Thickness ranged from 36 to 82 centimeters. This layer of fill extended across almost the entire cellar, including part of the possible bulkhead entrance (see "Architecture," below). The base of the layer sloped down away from the bulkhead at the western edge of Feature 1. The tpq of this master context is 1740, based on the presence of molded white salt-glazed stoneware. Two creamware fragments were recovered from this master context, 13 but since they were close to the top of the feature, near a large tree root and another intrusion observed only in profile (Figure 9), it is thought to be intrusive.

The next master context, M12, represents a very brief erosion episode. M12 was a sloping layer of subsoil-like clay. The layer was a brownish yellow (10YR 6/6) clay, with a thickness ranging from 2 to 11 centimeters. No artifacts were recovered from this layer.

Master Context 6 was a dark brown (10YR 3/3) loam which ranged from 15 to 21 centimeters in thickness. Prominent inclusions were charcoal, bone, brick, and shell, although not in the quantities noted in M4. Both M4 and M6 slope sharply approximately 1 meter from the west edge of Feature 1, as may be noted from Figure 9. This "shelf" may reflect the presence of an obstacle, now gone, which blocked the gradual slope of the cellar fill contents. The tpq for this layer is 1755.

The next stratigraphic level is an erosion layer, M7, a dark yellowish brown (10YR 4/5) clay, 13-35 centimeters thick. A few artifacts were found in this layer, mostly bone, glass and nails. One fragment of Buckley suggests a tpq of 1720, but as the excavator noted, most of the artifacts in that context came from an area just next to the layer's interface with M6. Other than the Buckley, wine bottle glass provides a tpq of 1650. This master context was trowelled, but not screened.

Master Context 8 was a yellowish brown (10 YR5/4) sandy clay. This thick fill is very sandy and in the sidewall shows traces of different washing-in events. Most of the artifacts in this layer were very large, many of them were tools of some kind: a pot hook, an iron ring, a hoe and a sword part were all found. Again, a lone fragment of Buckley points to a tpq of 1720, otherwise a single sherd of Nottingham suggests 1700. This is not sheet refuse, as indicated by the size of the two slipware platters recovered.

The next layer, M10, was somewhat obscured by the fact that the cellar was repeatedly filled with rainwater and groundwater in the course of archaeological excavation. Furthermore, the fill was very similar in color to surrounding layers, being an olive gray (5Y 4/2). Thus, it could only be distinguished in terms of its clayey texture. The one place where this layer was effectively isolated was in the northeast quadrant. Here, M10 was 13 centimeters thick and included bone, charcoal, brick, mortar and shell—though not in appreciable quantities. This layer seems to have been fairly barren except for those inclusions. One significant artifact is a fragment of a dip-molded empontilled wine bottle, which gives this master context a tpq of 1730.

The next several master contexts represent erosion events or other weathering processes. M9 was a brownish yellow (10YR6/8) clay, which at its maximum depth was 23 centimeters. It occurred primarily adjacent to the walls of the feature and lensed out toward the center of the cellar. This layer contained very few artifacts or inclusions, but did provide a tpq of 1720.

The next layer was another erosion layer, M14, a yellowish brown (10YR6/4) sandy clay with brick, mortar and bone inclusions. Its maximum depth was 8 centimeters, with a minimum of 1 centimeter. Wine bottle glass was found in this context, giving a tpq of c. 1650.

14

M15 consisted of a yellowish brown (10YR5/6) clay with no inclusions. The tpq of c. 1650 came from one of the few artifacts recovered, a fragment of wine bottle glass.

M16 was a very thin lens (barely 1 centimeter) of grayish brown (10YR5/2) sandy clay. No tpq could be determined for this master context.

M17 and M18, described below, had no direct stratigraphic relationship to layers M14-16, and we only know that M17 and M18 are later than M10, as are M14-16. Master context M17 was a yellowish brown (10YR5/6) sticky clay only 7 centimeters in depth, for which there was no available tpq. M18 was a brown (10YR5/3) sandy clay 2 centimeters deep, which was nearly mistaken for subsoil because of its similarity with the surrounding natural soil. However, inclusions of bone, charcoal, brick and shell indicated that this layer did contain cultural material. The tpq which was established is 1730.

The next layer was a dark yellowish brown sandy loam with animal bone, charcoal, and notable mortar inclusions. M19 was estimated by the excavators to have an average depth of 20 centimeters.

The final layer in Feature 1 was M13, a yellowish brown (10YR5/6) sandy clay. It was found in only a very limited area and contained very few artifacts or inclusions. It was quite thin (1-9 centimeters) and was underlain by natural subsoil.

Figure 11 is a Harris Matrix which shows the relationships between the various layers in Feature 1. The numbers in the matrix are the individual context numbers. Table 1 correlates these context numbers (also used in Appendix A, the artifact inventory) with the master contexts described above.

15

Figure 11. Harris matrix diagram of Feature 1.

Table 1.
List of Master Contexts

Description	Master Context	Context Number
Plowzone	M11	68AP-0, -30
Tree holes	M20	68AP-12, -13, -14, -15, -16, -17, -23, -27
Root cellar	M3	68AP-3, -32
Feature 1	M21	68AP-37
	M4	68AP-1, -2, -5, -6, -7, -8, -35, -36, -41, -42, -45, -50, -77, -78, -79, -80, -84
	M22	68AP-44
	M12	68AP-55, -57
	M6	68AP-9, -31, -43, -49, -52, -81, -83
	M7	68AP-54, -59, -60, -86, -89
	M8	68AP-48, -51, -61, -62, -65
	M10	68AP-63, -66, -75, -82, -85
	M9	68AP-67, -88
	M13	68AP-73, -74
	M14	68AP-68, -87
	M15	68AP-69, -86
	M16	68AP-70
	M17	68AP-64
	M18	68AP-47
	M19	68AP-58, -72, -76, -90

16

Architecture

The irregularity of shape and placement of the various scattered features shown in Figure 6 indicate that they are tree holes rather than post holes. This means that no architectural features, other than the cellar itself, were found at 68AP. Without any architectural features, such as postholes or foundations, or even a hearth, very little can be said about the physical characteristics of the building served by this cellar. One solution to this problem is to compare Feature 1 with similar features or sites.

A search through archaeological reports for the seventeenth- and eighteenth-century Virginia Tidewater has yielded a number of sites with similar characteristics from a range of time periods, and bearing a variety of interpretations. The earliest such feature (1620-1640) is found at Site "D" at Carter's Grove. Here, a 11.5 by 7.25 foot (3.5 by 2.2 meter) pit with a flat bottom 3.5 feet (1.07 meters) below subsoil was interpreted as a borrow pit mined for the manufacture of daub, which later served as a watering hole for cattle (Luccketti n.d.).

Several features excavated at Kingsmill also seem similar to Feature 1. A brick-lined half-basement at the Utopia Quarter (ca. 1641-1700) was approximately 15 by 12.5 feet (4.6 by 3.8 meters) (Kelso 1976:2). The North Quarter (ca. 1750-1775) also had an unlined cellar 3 feet (0.9 meter) deep which measured approximately 14 by 12.5 feet (4.27 by 3.8 meters) (Kelso 1984:103, 124-127).

Feature 2 at 44CC297 is described as a possible interior root cellar or roofed exterior cellar from the late seventeenth or early eighteenth century. It was ovoid in plan, roughly 8 by 7 feet (2.4 by 2.1 meters) in extent and 2.35 feet (0.7 meter) deep. It was irregularly shaped, with gradually sloping sides which terminated in a rectangular, square-sided bottom measuring 2.5 by 1.2 feet (0.76 by 0.36 meters) (Jones et al. 1991:36-37).

Corbin's Rest in Northumberland County (44NB180) included a large pit (Structure 6) interpreted as a wood-lined cellar constructed ca.1680-1710. This feature was much larger than Feature 1 at 46.5 by 37 feet (14.17 by 11.27 meters) maximum dimensions, but was very like it in shape and proportion. No other architectural features were found in the vicinity. The archaeologist concluded that "[This] example probably functioned similarly to a large dairy and cool storage house and it is possible that an upper floor a kitchen was installed with the servant cooks employing a half story loft above this as sleeping quarters" (Hodges 1990:94) . A ledge and remnants of a mold around the interior of the cellar were interpreted as the base of a sill that supported the building's superstructure (Hodges 1990:90).

Structure 1 at 44HT38 dates to the years 1680-1730. The cellar for this post building measured 14 by 12 feet (4.27 by 3.66 meters), not including the builders trench, which made the archaeologists decide that the cellar may have been lined (Higgins et al. 1993:64). 18 Structure 13 contained a cellar approximately 20 feet square (6.1 by 6.1 meters) and 2 feet (0.61 meter) deep that was filled sometime around 1720-1725. This cellar was lined with brick and wood (Higgins et al. 1993:171). This cellar has a bulkhead entrance to one side.

The several cellars at the Pope Site (44SN180) were interpreted as belonging to domestic structures. The largest measured 20 by 18.4 feet (6.1 by 5.6 meters) and had no associated architectural features other than a robbed out brick foundation. Two smaller contemporary cellars (6.4 by 4.5 feet and 7.69 by 6.36 feet) were clearly post buildings (Reinhart 1987:13-34).

Excavations at College Landing in Williamsburg identified two structures that may be compared to 68AP. First, archaeologists uncovered a small cellar dating to between 1750 and 1815, designated the "lot 37" cellar. There were no extant architectural remains other than a hearth located within the cellar hole itself and a shelf in the subsoil near the bottom of the feature: a sill for laying a robbed-out brick foundation (Hudgins 1977:32). In this case, the bulkhead was near a corner of the cellar, leaving room for the hearth and chimney to be centered on the same end. The second structure, labeled the "colonial" cellar, consisted of a basement measuring 16.25 by 9.5 by 3 feet (3.4 by 2.9 by 0.9 meters), including the bulkhead entrance. Remains also included a chimney base and a root cellar (Hudgins 1977:52).

The mid to late eighteenth century site at 44JC298 was slightly different in that the cellar hole was a basin-shaped feature approximately 8-10 feet (2.44-3.05 meters) in diameter. The structure over the cellar was supported by two main posts, the holes for which were discovered in the bottom of the cellar, and piers on the exterior of the cellar (White 1991:10). This feature dates to from the late eighteenth to the early nineteenth century.

Feature 53 at 44GL357 was a square, partially brick lined cellar measuring 16 by approximately 20 feet, and 3 feet deep (4.88 by 6.1 by 0.91 meters). It is interpreted as the remnants of a full basement for a one-story brick structure (Stuck et al. 1996:207). The fill from this feature dates from the second and third quarters of the eighteenth century. This cellar was also associated with a small root cellar.

Table 2 summarizes the foregoing information.

One common thread running through all these interpretations is that, regardless of whether or not other architectural remains are found, all of the rectilinear pits (as well as several of the others) were interpreted as cellars within or underneath structures, rather than outdoor manufacturing, farming or processing facilities. The sole non-structural example cited above (Site D) was basin-shaped rather than square-sided. Therefore, seems reasonable to conclude that 68AP, with its generous dimensions, square sides and straight walls was once a deliberately excavated cellar, with some form of superstructure.

A second question is whether or not the "structure" served by this cellar was anything as elaborate as a dwelling house or rather consisted of a few planks thrown over an exterior storage pit. As the excavator of Corbin's Rest wrote, "Cellar Houses, 'covered 19

TABLE 2. CELLARS IN TIDEWATER SITES

NAME	DIMENSIONS (m)	SUPPORT	LINING	BULKHEAD	TYPE OF STRUCTURE	ROOT CELLAR?
Utopia Quarter	4.6 x 3.8 x 1.83	post holes	brick	yes (step)	dwelling	———
North Quarter	4.27 x 3.8 x 0.9 (est)	brick foundation	———	yes (step)	dwelling	adjacent
44CC297 fea 2	2.4 x 2.1 x 0.7	———	———	yes (step)	roofed cellar or root cellar	———
Corbin's Rest str 6	14.17 x 11.27 x 0.52	wood sills	wood	yes (step)	dairy or kitchen	———
44HT38 str 13	6.1 x 6.1 x 0.61	brick foundation	brick/wood	yes (unknown)	dwelling or tavern	within
Pope Site fea 7	6.1 x 5.6 x 0.6	brick foundation	brick	———	dwelling	———
College Landing lot 37	6.1 X 4.9 x 0.61	brick foundation	brick	yes (step)	dwelling	within
College Landing "colonial"	3.43 x 2.9 x 0.91	———	———	yes (no step)	dwelling	adjacent
44GL357 fea 53	4.88 x 6.1 x 0.91	brick foundation	brick	yes (step)	dwelling	———
68AP	5.22 x 4.1 x 1.4	———	———	yes (no step)	———	adjacent

20 cellars', 'roofed cellars' and 'pit houses' are familiar finds throughout the 17th century" (Hodges 1990: 94). It seems that one way to discover the nature of the 68AP structure would be to examine the specific characteristics and interpretations of several of the rectilinear features described above. These include Utopia Quarter, North Quarter, 44CC297 Feature 2, Corbin's Rest Structure 6, 44HT38 Structure 13, the Pope Site, the two College Landing structures, and 44GL357 Feature 53.

As Table 1 shows, cellars come in a range of sizes. The average area of the cellars in square meters, excluding the much larger Corbin's Rest is 22.35, very close to the size of 68AP Feature 1 at 21.4 square meters. Of the other features, all but one (44CC279) are definitely proper buildings rather than roofed or covered outdoor storage pits (Figure 12). This includes even the one other cellar without a foundation or post holes (College Landing "colonial"). The excavator of College Landing writes, "the absence of postholes, piers, or wall impressions, indicates that the structure was probably erected on either ground laid sills or shallow brick or wooden piers" (Hudgins 1977:54) . The same was probably true at 68AP.

The lack of architectural features such as posts or foundations was characteristic of the other eighteenth and nineteenth century structures at Rich Neck (Franklin 1997, McFaden et al. 1993:17). It is interesting to note that the structure at the Rich Neck Slave Quarter is also without visible means of support. This building is known only from the pattern of its root cellars; no foundation, piers or footings were recovered (Franklin 1997). It is possible that the later dwelling continued an architectural precedent set by Feature 1. The table also shows that it is not unusual for such cellars to be unlined as Feature 1 appears to have been.

Figure 12. Roofed cellar (from Doepkens 1991).

21

Figure 13. Other features on Rich Neck Plantation.

The extension of Feature 1 (along the western edge) appears to be a bulkhead. Although a few of the bulkhead entrances described above were build of wood or brick, there are several examples noted above in which the bulkhead entrance to a cellar was simply carved out of clay. One of these had no steps at all (College Landing "colonial"), as is the case with Feature 1.

At this time, there is no reason to definitively link Feature 1 with the root cellar (M3) in terms of use-life. One might conclude from their similar orientation that they were contemporary, even part of the same structure. This interpretation has been employed at the North Quarter and College Landing "colonial" sites, which have root cellars nearby. However, there are reasons to question a similar interpretation at 68AP. As stated above in the "Description of Strata and Features," M3 showed little evidence of erosion, suggesting that it was sheltered from the elements during its use-life, and then quickly filled when it was no longer needed. However, this scenario contrasts with the depositional history of Feature 1 which was exposed to considerable weathering as it was being filled.

Furthermore, demonstrating conclusively that these two features were a part of the same occupation is difficult. There are no crossmends between M3 and Feature 1 to indicate that they were filled using the same secondary refuse, and artifact ratios are not very comparable (see "Crossmends" and "Comparisons with 68AL" below). Finally, the tpq for M3 is 1755, ten years after the tpq of 1745 for M6 (and by extension, the strata above it.

A hypothetical structure over the cellar and the root cellar could be readily aligned with the seventeenth-century structures from the earlier Rich Neck component. Certainly, its orientation is not far off that of the eighteenth-century Slave Quarter, but it is more similar to that of the earlier component, down to the orientation of the hypothetical gable ends (Figure 13). So Feature 1's overall placement in the landscape seems to point to some relationship with these structures. Nevertheless, the artifacts and (lack of) architectural 22 features resemble the later period structures, indicating how well 68AP forms an interpretive bridge between the earlier and later occupations of the plantation.

It is difficult to say from the architectural remains who occupied or used the structure at 68AP. Yet a number of parallels exist with structures identified as slaves' or servants' quarters, ranging from the third quarter of the seventeenth-century (Utopia Quarter) to the late eighteenth-century (North Quarter) (Kelso 1984:72, 75, 125).

The documentary record for Rich Neck plantation enhances our understanding of who may have occupied or otherwise used 68AP Feature 1. Given the tpqs for the various fill layers (1650-1745, or 1755 if one includes M3), the structure was in use during the lifetime of Philip Ludwell III and possibly that of Philip Ludwell II. The tpqs of the majority of the master deposits (see "Description of Strata and Features") are from the decades following the removal of the Ludwell family to Green Spring, though the use of the structure may predate this shift. Therefore, it is most likely that the people who used the structure at 68AP were Ludwell slaves or perhaps an undocumented overseer.

Artifacts

Site Formation Processes

Because the fill from the features at 68AP is secondary rather than primary refuse, the analysis of this site downplays the activities which preceded disposal, and focuses instead on the ultimate acts which shaped the archaeological record as it is encountered. Rather than debating the primary use of this cellar, and focusing on understanding the intent and actions of its builders and users, attention is devoted to understanding the way in which the cellar was filled and the cultural activities which this filling reflects.

When a structure is leveled, rubbish is commonly used to fill in any remaining holes in the ground. When filling is deliberate and rapid, the fill at the bottom of the hole contains more recent artifacts than at the top, as landscapers initially use garbage at hand, but are forced to find fill further afield to complete the task. Such is the case in the ethnographically-observed demolition of early twentieth-century structures which are razed to the ground (White and Kardulias 1985:72). If a structure is filled incidentally, or slowly, one would expect older types of artifacts in the lower levels of fill and more recent types in the upper layers.

31

The tpqs from Feature 1 do not provide a clear picture. Here, the available tpqs from the uppermost stratum to the deepest are: M4 1740; M6 1745; M7 1720; M8 1725; M10 1730; M9 1720; M14 ca., 1650; M15 ca., 1650; M19 1725. This does not form a regular sequence from latest to earliest, but there is a general trend toward earlier dates from stratigraphically lower levels.

The mean ceramic dates are less ambiguous, although the presence of large amounts of English tin-glazed ceramics tended to slant the mean dates toward mid-century. Using the method as described by South (1977), the mean date for M4 was 1746, for M6 1745, for M8 1730.

But beyond the generic description of filling events at the site, what do these fill sequences mean? There are a number of processes which contribute to archaeological deposits: loss, discard at the point of use, ritual deposits, deposition during abandonment, activities of later occupants (e.g.: squatters), disturbance by human landscaping activities, the collapse of structures and erosion, secondary refuse discarded by later occupants of a different portion of the site (Schiffer 1985:24), and destruction or razing of structures.

Several of these processes can be eliminated immediately. Despite jokes about the possible "ritual" significance of the cow skull found at the very bottom of the cellar, Feature 1 was not a ritual deposit. And clearly the fill layers from the cellar are too extensive represent items lost into it during its tenure as part of a structure. Nor do these layers represent the detritus of activities taking place in the cellar itself. The fractured quality of the ceramics attest to the fact that these deposits are "secondary refuse" (Schiffer 1985), consisting of small fragments. Furthermore, Feature 1 appears to have been deliberately abandoned, because of the lack of "de facto refuse"—in other words, usable artifacts recovered from use contexts or activity areas (Schiffer 1977:23). So there remain only a few processes to discuss: secondary uses of the structure; collapse and demolition; and refuse disposal by later residents of adjacent structures.

The activities visible in the archaeological record more closely parallel those of refuse disposal than demolition (White and Kardulias 1985). We did not find later artifacts in the lower layers and earlier ones in the upper layers, as has been suggested as typical of razing activities. None of the layers appeared to be the remains left following the burning and reduction of architectural debris. Any evidence of razing should have been in the lower layers of fill, prior to the addition of "foreign" rubbish, and prior to any evidence for exposure to erosional processes. But in this case, the debris was scattered among the household rubbish, separated into discrete layers, which seems to argue against a single major razing project.

After its life as a late seventeenth-early eighteenth-century outbuilding, the structure over the Feature 1 may have served any number of purposes. It may have been used to house and maintain the dozens of slaves who continued to live at Rich Neck after the Ludwells had left for Green Spring between 1704 and 1706. Alternatively, it could have been a plantation work or storage space. The possibilities are practically endless, though.

Ethnoarchaeological research has indicated that structures often have a protracted use life which includes a variety of functions, as appropriate to the condition of the structure. 32 This is especially true in the case of vernacular buildings. "Where more than a minimum of capital and labour is invested in construction, but where buildings are subject to irreversible decay over a limited time span, they will be liable to fulfill a series of functions over their useful life. This has been called a 'devolutionary cycle' " (David 1971:119). For example, a building might start out as a dwelling, then be converted to an outbuilding such as a kitchen, later it could become a storehouse, or an animal enclosure (David 1971:123).

Clearly, the initial function of a structure is not its only important one within a cultural system. The cellar and the building above it were probably many things to many people. And in the end, it served one last function—as a refuse pit. The decision to finally fill in the cellar hole may have been triggered by the departure of the Ludwell family for England in the 1750s (Smith 1993:5), an event which may have resulted in the shuffling of personnel and redistribution of activities at their plantations. However, the evidence from the artifact analysis is far from conclusive.

The collapse and demolition of the structure that sat over the Rich Neck cellar is difficult to identify archaeologically. Items that might be construed as architectural debris were recovered from all layers. The remains of several wooden boards were recovered from master contexts M4 and M7. But there was little notable architectural hardware other than nails in any layer.

It is highly probable that the materials of which the structure was built had been removed for use elsewhere, as building materials, or even as firewood (Lange and Rydberg 1972:422). The lack of architectural features such as postholes, foundations, or chimneys, and the lack of significant quantities of architectural materials such as boards, bricks, and mortar seem to indicate that this structure was picked clean at the end of its primary use-life, and that the materials the archaeologists encountered probably bore little relation to the building which stood over the cellar.

The fill layers seem to consist of several separate mass disposals of household trash from the slave quarter at 68AL. The fill was not dumped in all at once, nor does it seem to have been collected from earlier deposits, as is often the case in demolition projects. So we may conclude that the activities which took place here were more focused on discard from the quarter than filling the cellar. We see that the fill is not composed of debris from the destruction of the building which stood over it, but rather from occupants of a later structure in the same area, who periodically disposed of large quantities of fairly fragmented household trash, using the most convenient receptacle available.

Although numerous erosion layers (M7, M8, M9, M13, M14, M15, M16, M18) indicate that the cellar was exposed long enough to allow the accumulation of many centimeters of silt and eroded subsoil, this is not necessarily evidence of prolonged exposure. Our experiences while excavating the cellar showed that several centimeters of silt could build up after an evening's rainstorm. The layer recorded as M10 could have taken as little as a single rainy summer week to accumulate. In addition to the layers attributable to erosional forces, several master contexts seem to be the byproduct of exposure to weathering. M12, M10 and M17 are all very sticky clay layers which could be the result of weathering breaking down the structure of the sediments in the layers below them.