The Clifts Plantation Site:
Summary of Documentary Evidence and Intra-site Chronology
The Clifts Plantation (44WM33) is located on the south shore of the Potomac River in Westmoreland County, Virginia. The site lies on a tract of land now owned by the Robert E. Lee Memorial Association, Inc., a group devoted to the preservation of Stratford Hall, the 18th-century mansion that was the birthplace of Robert E. Lee. The site was excavated over a three-year period, from 1976 to 1978 (Neiman 1978, Neiman 1980a, 1980b, 1990). Artifacts and records from the excavation are curated by The Center for Historic Preservation, Mary Washington College, and by The Robert E. Lee Memorial Association at Stratford.
Permanent English settlement of Westmoreland County began in the late 1640's. Many of the earliest settlers in the region were immigrants from Maryland, among them Nathanial Pope, who first patented the land on which The Clifts was located in 1651 (Virginia Land Patents 4:32). Pope, who lived up river from the Clifts tract at the confluence of Mattox (Appomattox) Creek and the Potomac, was among the county's wealthiest residents at his death in 1660 when he left The Clifts tract to his son Thomas (Westmoreland 10:115). Thomas Pope turns up in the Westmoreland County court records sporadically through the ensuing 25 years as a "planter of Westmoreland" and "merchant of Bristol" (Westmoreland 4:89,266, V:404). During his stays in Virginia, he occupied a plantation located at the mouth of Pope's Creek at the western edge of 2450 contiguous acres that included The Clifts. Archaeological evidence indicates that The Clifts tract, comprising the eastern half of Thomas Pope's 2450 acres, was first occupied during this period, presumably as a tenant farm.
At Thomas' death in 1685, the Clifts portion of his property passed to his sons, Richard and John, with a dower interest to his wife, Joanna, who acquired another third at John's death in 1700 (Westmoreland 15:324). Joanna Pope appears to have remained a Bristol resident during her entire life. Thomas' youngest son, Nathaniel, appears in the county's records for the first time in 1704 when he married the daughter of a Westmoreland Justice of the Peace (Westmoreland 8:260). Like his father, he appears to have pursued a career on both sides of the Atlantic, being styled both "merchant" and "mariner" (Westmoreland 7:59,327). In 1708 his mother gave him power of attorney to manage The Clifts. The letter of attorney refers to her son as "Nathanial Pope of Pope's Creek" (Richmond 5:116-117). The phrase does little to resolve the ambiguity of whether or not Nathaniel and his wife actually resided at The Clifts. His wife's deceased first husband owned land on Pope's Creek and The Clifts is situated three miles down river from it (Westmoreland 17:246).
In 1716 Joanna and Richard Pope sold the Clifts, including what was referred to in the deed as "the manner house erected on the second clift," to Thomas Lee (Westmoreland 15:324-345). While the Popes had circulated the among members of Westmoreland County's political elite, Lee's family had been part of the colony-wide elite for two generations. Lee himself would become a member of the Governor's Council and eventually acting governor of the colony. Up until 1729, Lee lived at his father's plantation on Lower Machodoc Creek, when the dwelling burned to the ground. Sometime thereafter he erected the brick mansion, today known as Stratford Hall, on the Clifts tract about a quarter mile from site of The Clifts (Wyrick 1971). The "manner house" and the buildings surrounding it were demolished to make way for a road running between Lee's new mansion and a landing on the shore of the Potomac. Recent dendrochronological indicate the construction of Stratford began in 1738. They raise the question of the location of Lee's residence in the decade following 1729. The Clifts is a possibility, but there is no documentary evidence to make it more than that. In fact, a court order from 1729, two months after the Machodoc fire, reveals that Lee was taking steps to rebuild at Machodoc in that year (Westmoreland IX:255). Archaeological evidence indicates occupation of The Clifts continued at least until 1730.
The site lay in a field that was actively plowed up until the excavation began in 1976. The plowzone over the central portion of the site was excavated by hand in 124 quadrats, nearly all 10-feet square, and screened through 1/4-inch mesh. The outlying portion of the site was stripped mechanically. All sediments from all features were screened through quarter-inch mesh.
The architectural layout of the site throughout the 60-to-70 year occupation was conditioned by the immediately surrounding topography. The principal dwelling -- the "manner house" of the 1716 deed -- and the surrounding cluster of structures are situated on a finger of dissected upland. About 1300 feet to the north lies a cliff, one of a series of bluffs that stretch along the Potomac shore, plunging roughly 130 feet to the river. About 160 feet to the west of the dwelling lies a steep-sided ravine in which a spring surfaces. In the absence of a well, this was the water source for the site during its occupation. To the east of the dwelling the land slopes off more gently. The area to the west of the dwelling, and closest to water, was the site of most of the ancillary structures on the site, while the land to the east was the site of large fenced enclosure, presumably a garden.
The basic architectural elements of the site can be enumerated quickly. All buildings were framed around hole-set posts. At the center of the site lay the principal dwelling (Structure 1). This building had a complex history of modification. It began life with a 3-unit, center-chimney, cross-passage plan. About 1700, the cross passage disappeared and the service room on the end of the building was incorporated into the hall (Neiman 1990, cf. 1980a,b.). Early in the occupation the house was surrounded by a palisade fence with bastions on opposite corners. Just to the southwest of the dwelling lay two superimposed, and hence successively built, multiple-bay structures (Structures 2 and 3). Analysis of plowzone artifact distributions suggests that the earlier of these buildings was a work house, devoted to bulk processing and storage, while the later building was a dwelling that housed slaves who dominated the plantation's labor force after 1700.
Arrayed in an arc to the west of the dwelling and the quarters lay a series of 9 small, single-bay outbuildings (Structures 4-12). Their construction dates spanned the occupation and there is a trend to larger numbers of larger buildings. Burnt daub and wood in their post molds indicate 6 of these structures were smokehouses. Plowzone artifact distributions indicate the one or more of the later structures were dairies. Among the outbuildings were interspersed numerous, irregularly shaped pits, dug into subsoil and filled with artifact-laden sediments. To the east, on the opposite side of the dwelling, lay the remains of two successive fence systems, one comprised of ditch-set uprights, the other of hole-set posts. On the eastern edge of the garden, were 18 grave shafts, 16 of which contained human skeletal remains (Aufderheide et al. 1981). Finally, a pair of larger, two-bay outbuildings, both dating to the second half of the occupation on the southern end of the garden (Structures 13 and 14).
The chronology of the site was worked out using the occurrence seriation method. In the 15 cases where stratigraphic relationships existed among the deposits whose assemblages were seriated, the chronological order inferred from stratigraphy agreed with the seriated order. The seriation allows the occupation of The Clifts to be resolved into 15 periods, or more precisely, the assignment of a limited number of depositional units containing relatively large samples of artifacts to 15 periods. However, because the sizes of the assemblages they contain are small, many depositional units are not included in the seriation. Other means must therefore be employed to fit such units into the chronology. With a given number of types, the fineness with which time can be resolved in a set of assemblages scales with their sample sizes. This means that if deposits containing small artifact samples are to be seriated, the seriation will have to be one based on fewer types and therefore result in less fine temporal resolution. The seriated assemblages at The Clifts were lumped into four major phases, each having 15-to-20 year durations. Each phase groups a set of assemblages and, more importantly, a set of types that were first introduced to the site during it. Each of the four type-groups then can serve as a "type" for what in effect is a second, coarse-grained, occurrence seriation that includes many more assemblages. These smaller assemblages, and by inference the deposits containing them, were assigned to one of the four phases on the basis of the latest group of types they contained. Deposits containing only one or two artifacts could only be slotted into the chronology in more general terms, based on a terminus post quem derived from the order of ceramic-type introductions.
Two additional lines of evidence were used in constructing the chronology. An extensive program of refitting ceramic sherds recovered from different deposits allowed the attribution of sherds to individual vessels. The initial seriation made it possible to determine the phase during which pieces of each reconstructed vessel entered the archaeological record for the first time. This yielded a terminus post quem for the deposition of other sherds belonging to that vessel. Otherwise undated deposits that contained sherds from phased vessels were assigned termini post quem on that basis. Stratigraphic relationships between seriated and undated deposits supply similar information. In some cases, they also yield a complementary terminus ante quem.
The foregoing represents a departure from standard approaches to chronological inference in historical archaeology that are based on termini post quem from historically documented beginning manufacturing dates of ceramic types and on weighted means of manufacture dates. A seriation-based approach offers finer-grained chronological control than the first of these methods. Unlike the second, it offers immediate insight into the empirical correctness of the results, based on an evaluation of the extent to which the pattern of occurrences in the permuted matrix fits the abstract model on which the method is based. This is not to say that evidence concerning historically documented dates of manufacture has no place in chronological inference, merely that there are better means for building relative chronologies within sites. Termini post quem, based on documented manufacture dates, come into their own in anchoring relative chronologies in absolute time.
Absolute temporal boundaries for the four-phase chronology at The Clifts were derived in precisely this fashion. Occupations of the site appears to have begun c. 1670. This date is derived from the fact that the Phase-1 assemblages are dominated by lead-glazed coarse earthenwares manufactured by Morgan Jones, a potter known to have operated a kiln in Westmoreland County as early as 1669 and who left the county in 1681. Additional corroboration for the date comes not from ceramics, but from the palisade erected around the principle dwelling during Phase 1. Documents indicate that wealthier planters in Westmoreland and adjacent Northern-Neck counties erected such fortifications around their households during 1675 Indian scare that helped precipitate Bacon's Rebellion in the Virginia Colony.
Although Jones' wares continue to show up in assemblages until the end of the occupation, the majority of sherds appearing in Phase-2 and later contexts show signs of depositional recycling. Staffordshire Yellow Slip Earthenware was introduced to the site at the beginning of Phase 2, in the form of cups exhibiting finely combed decoration that, on the basis of documents and kiln excavations, appeared on Staffordshire products in the period from c.1680 to c.1700. Together these observations suggest a Phase 1-Phase 2 boundary at c.1685.
A beginning date of c. 1705 for Phase 3 is suggested by the appearance in its earliest deposits of English stonewares with lustrous brown finishes associated with Burslem and Nottingham. These wares went into manufacture c. 1700. That Phase 3 spanned the first decades of the 18th century is corroborated by the appearance of Rhenish Blue-Gray Stoneware mug sherds in Phase 3 deposits bearing the "AR" cipher of Queen Anne (1702-1714) impressed at the rim.
The arrival of Slipped White Saltglaze Stoneware at the site signals the beginning of Phase 4. The initial date of manufacture for this ware is c. 1715. Its first appearance in the Westmoreland County probate inventories dates to 1726 (Westmoreland Inventories 1:44). Sherds of Buckley Earthenware, with its distinctive red and yellow banded body, first appear early in Phase 4, although a similar black-glazed, red-bodied ware is present in Phase 3 deposits. Buckley does not appear on Chesapeake sites until c. 1720. Finally, the earliest Phase-4 deposits contain sherds of Rhenish Blue-Gray Stoneware mugs bearing the "GR" cipher of King George I or II (1714-1727, 1727-1760) on a central medallion. A c.1720 beginning date for Phase 4 is therefore suggested.
Dating the end of the occupation is a more difficult matter. Among the last ceramic types introduced to the site was Unslipped White Saltglaze Stoneware. The earliest documented date of production for this ware is 1720. The earliest evidence of its introduction to the Chesapeake is recorded in the recovery of several examples from Corotoman, documented to have burned to the ground in 1729. Hence the presence of the ware at The Clifts suggests that the occupation continued at least to c. 1730. Just how much longer it continued is problematic. The absence of molded White Saltglaze Stoneware plates from the assemblage indicates abandonment had occurred by c. 1740. This is also the dendrochronologically documented date for the construction of Stratford Hall, an event that likely occasioned abandonment of the site, if it was still occupied.
Fraser D. Neiman
15 November 1998
Aufderheide, A.C., F.D. Neiman, L. Witmers and G. Rapp
1981 Lead in bone II: Skeletal-lead content as an indicator of lifetime lead ingestion and the social correlates in an archaeological population. American Journal of Physical Anthropology 55:285-291.
1978 Domestic architecture of the Clifts Plantation Site: the social context of early Virginia building. Northern Neck Historical Magazine, 28:3096-3128.
1980a The "Manner House" before Stratford (Discovering The Clifts Plantation). Robert E. Lee memorial Association, Stratford VA.
1980b Field Archaeology of The Clifts Plantation Site, Westmoreland County, VA. Ms. on file Robert E. Lee Memorial Association, Inc. Stratford VA.
1990 An Evolutionary Approach to Archaeological Inference: Aspects of Architectural Variation in the 17th-century Chesapeake. Ph.D. Dissertation, Yale University.
Virginia Land Patents
4 Virginia Land Patents, Book 4, Ms. on file Virginia State Library.
2 Deeds, Wills, etc. 1661-1662. Ms. on file, Westmoreland County Courthouse, Montross, VA.
3 Orders, 1662-69. Ms. on file, Westmoreland County Courthouse, Montross, VA.
4 Deeds, Patents, Wills, 1665-1677. Ms. on file, Westmoreland County Courthouse, Montross, VA.
5 Orders, 1678-1689. Ms. on file, Westmoreland County Courthouse, Montross, VA.
7 Orders, 1698-1705. Ms. on file, Westmoreland County Courthouse, Montross, VA.
8 Orders, 1705-1721. Ms. on file, Westmoreland County Courthouse, Montross, VA
9 Orders, 1731-30. Ms. on file, Westmoreland County Courthouse, Montross, VA
10 Deeds and Wills, Book 1. Ms. on file, Westmoreland County Courthouse, Montross, VA.
15 Deeds and Wills, Book 6. Ms. on file, Westmoreland County Courthouse, Montross, VA.
17 Deeds and Wills, Book 8. Ms. on file, Westmoreland County Courthouse, Montross, VA.
1 Inventories, Book 1. Ms. on file, Westmoreland County Courthouse, Montross, VA.
1971 Stratford and the Lees. Journal of the Society of Architectural Historians 30(1).