Colonial Williamsburg Digital Library

LOT HISTORY

The use of the area presently occupied by the Peyton Randolph site during the Middle Plantation era (1634-1699) is generally unclear. The only archaeological feature representing this time period was a boundary ditch crossing the site in a northwest-southeast direction. This ditch may be related to the palisade which was built in 1634, traversing the entire Middle Peninsula between what are now called College and Queen's Creeks. No other conclusively seventeenth-century or aboriginal features were encountered during the course of excavation.

The first historical documentation involving the area under investigation was on November 11, 1714, when the Trustees of the City of Williamsburg conveyed eight lots to William Robertson. The lots sold to Robertson were numbered 232, 233, 234, 235, 236, 237, 207, and 208 on the City Plat. Unfortunately, the first plat of the city no longer exists, and the exact location of the numbered lots is not known. However, careful examination of the surviving historical records has placed the lots mentioned above in the block bounded by the eighteenth-century locations of Scotland Street on the north, N. England Street on the west, Nicholson Street on 4 the south, and Queen Street on the east. Their configuration was probably as follows:

Written in the conveyance of the eight lots to Mr. Robertson was the stipulation that he build upon each of the lots within two years, or they would revert to the Trustees:

… that if the said Wm. Robertson his heirs or Assigns shall not within the space of Twenty four Months next ensueing the date of these Presents begin to build & finish upon Each Lott of the said granted Premisses One good Dwelling house or houses of such dimensions & to be placed in such manner as by One Act of Assembly made at the Capitol the 23d day of October 1705 … then it shall & may be lawful to & for the said Feoffees or Trustees & their succesors Feoffees or Trustees for the Land appropriated for the building & Erecting … to have hold & Enjoy in a like manner as they might otherwise have done if these Presents had never been made (Stephenson 1952, from York County Deeds and Bonds, III, 28-29).

Apparently Robertson found himself unable to meet these requirements, because on November 19, 1715 he sold lots 233 and 234 to Philip Ludwell. On April 20, 1717 the Trustees gave title of lot 232 to John Tyler, who was apparently already living on it. Again, on January 17, 1718 the Trustees conveyed title to lot 235 to Samuel Cobbs, also already in possession. Robertson was able to retain four of the lots, however, until 1723 when he sold them to John Holloway. At that time the lots contained several houses, orchards, and a windmill (Stephenson 1952).

In July of 1724, only six months after he had purchased the four lots, Holloway sold one lot, probably 237, to John Randolph. The record of sale describes the lot as "… adjoining to the Lot whereon the said John Randolph now lives …" The record also mentions "messuage" on the lot, which probably describes a building built by Robertson soon after the initial purchase in 1714. How Randolph acquired the adjacent lot on which he lived at the time of this sale, presumed to be 207, is unclear. There is also a real possibility that the lot Randolph bought from Holloway was actually 207 and he was already residing in Structure S on lot 237. Since there is no record of the prior sale in the York County records, the assumption is that the transaction was recorded in the Hustings Court or in General Court, the records for which have not survived. Randolph sold the only other lot he owned (174) to Archibald Blair the same day in 1724 that he bought 237. The exact location of lot 174 is not known, but consensus has placed it across N. England Street on the southwestern corner of England and Scotland Streets.

John Randolph, new owner of lots 207 and 237, was born at his family home at Turkey Island in 1693. He attended the College of William and Mary from 1709 5 to 1711, becoming Deputy Attorney General for the Counties of Charles City, Henrico, and Prince George in 1712. Between the years 1715 and 1717, he traveled to England to study law. Upon his return to Williamsburg, Randolph was appointed clerk of the House of Burgesses. After serving as agent for the Colony for many years, in 1732 he became the only colonial Virginian to be knighted. He was elected Speaker of the House of Burgesses in 1734 and remained "Mr. Speaker" until his death in 1737.

John Randolph married Susannah Beverley of Gloucester County around 1718. At the time of his death, four children were surviving: Beverley, born c. 1720; Peyton, born c. 1721; John, born c. 1727; and Mary, whose birth and death dates are not known. Susannah Randolph survived her husband and remained owner of the house until her death sometime after 1754. It is at this date that the last historical reference to Lady Randolph appears. The exact date of her death and place of burial, however, are unknown.

The Randolph's eldest son Beverley resided in Gloucester, becoming a Burgess for the College and later Sheriff of Gloucester County. Younger son John Randolph (often known as John Randolph the Tory) became a prominent lawyer, Burgess, Attorney General and Speaker of the House. He was also quite interested in horticulture, probably writing A Treatise on Gardening by A Citizen of Virginia around 1765. Perhaps wanting a home commensurate with his brother's, he built Tazewell Hall, the largest private residence in the city, on South England Street. When the break with Great Britain came, John fled to England. His body was returned to Williamsburg after his death in 1784 and was buried alongside his father and brother in the vault at the College of William and Mary.

Peyton Randolph, just as his father and brother, attended the College of William and Mary and studied law in England. He was admitted to the bar on February 10, 1742/3 and began his career as Attorney General for the Colony on May 7, 1744. Soon afterward, on March 8, 1745/6, he married Betty Harrison of Berkeley. They resided at his mother's house where Peyton had grown up. Peyton's public life continued in various assignments as Attorney General, with increasing involvement in inter-colonial affairs. He was elected Speaker of the House in 1766. As the problems leading to the eventual revolution were becoming more apparent, Peyton Randolph was increasing his contact with prominent personalities from the other colonies. In 1774 the First Continental Congress met in Philadelphia, and Randolph was unanimously elected president. He was also president of the Second Continental Congress, held on May 10, 1775. Declining health prevented Peyton Randolph from continuing to preside over the Congress, although he was in attendance to see John Hancock elected to his former position. Peyton Randolph died in Philadelphia of what seems to have been a stroke at about eight in the evening on October 22, 1775. He was buried in the chapel at the College of William and Mary.

Peyton Randolph was survived by his wife, Betty, who inherited his entire estate since the couple had no children. Betty Randolph also survived the Revolution and the occupation of the city of Williamsburg by Lord Cornwallis' troops. She died in 1783 and was probably buried in the vault at the College with her husband.

6

The following advertisement of public auction appeared in the Virginia Gazette on February 15, 1783:

TO BE SOLD, By public auction, in Williamsburg, on Wednesday the nineteenth of February next, The houses and lots of the late Mrs Betty Randolph, deceased, together with a quantity of mahogany furniture, consisting of chairs, tables, mahogany and gilt framed looking glasses, and desks, a hansome carpet, a quantity of glass ware and table china, and a variety of other articles; also kitchen furniture complete. The above mentioned house is two story high, with four rooms on a floor, pleasantly situated on the great square, with every necessary outhouse convenient for a large family, garden and yard well paled in, stables to hold twelve horses, and room for two carriages, with several acres of pasture ground. Twelve months credit will be allowed for all sums above five pounds, on giving bond with approved security, to carry interest from the date if not punctually paid. The EXECUTORS. (Stephenson 1952).

The house was sold to Joseph Hornsby on February 21, 1783. Hornsby sold the property to Thomas Peachy sometime around 1800, but the record of the transaction has been lost. The house remained in the Peachy family until the late 1860s when it was sold to Richard Hansford. The succession of ownership for the remainder of the nineteenth century to the present is as follows:

1884-1893	Moses R. Harrell
1893-1897	John Dahn
1897-1920	E.W. Warburton
1920-1921	Williamsburg Inc. (real estate company)
1921-1927	Mary Proctor Wilson
1927-1938	Merrill Proctor Ball
1938-present	Colonial Williamsburg Foundation.

Merrill Proctor Ball's daughter, Gertrude Deversa, continues to occupy the east wing of the house as her residence. Her recollections of living in "Mr. Speaker's" house are related in the oral history section of this report.

ENVIRONMENTAL SETTING

The Historic Area of the City of Williamsburg is situated along a drainage divide running east-west along the peninsula. The ravines cutting into the dividing ridge from the south eventually channel run-off water into the James River, whereas those on the north side of Duke of Gloucester Street direct the flow into the York River. The area around the Peyton Randolph House has an elevation of about 80-85 feet above mean sea level. A fairly steep ravine which existed just to the east of the back lot became in the early 1940s the location for the Colonial Parkway tunnel.

Since the setting is now semi-urban, much of the flora around the Peyton Randolph House is cultivated rather than wild. Walnut, pecan, weeping willow, and crepe myrtle are but a few of the trees surrounding the yard. In earlier periods, however, Williamsburg had the characteristic flora of the Southeastern Coastal Plain—typified by temperate deciduous forests with areas of black and loblolly pines. 7 All of the Williamsburg area has been logged many times over the past 350 years, undoubtedly altering the landscape and ecology considerably.

Various faunal species common in the area in the eighteenth century are still found in Williamsburg and its environs. White-tailed deer, which are rather prolific on the outskirts of town, especially in protected areas along the Colonial Parkway, are not found in the Historic Area. However, grey squirrel, raccoon, opossum, skunk, rabbit, mice, and rats live peacefully within the town. Even occasional grey and red foxes are seen within the city limits. Birds common in the area include many types of water fowl, song birds, eagles, hawks, and vultures. Snakes, land and riverine turtles, fish, shellfish, and crabs are found in abundance in stream-side and marshland environments.

The climate of the Williamsburg area is fairly mild, mitigated by its proximity to the Chesapeake Bay and the Atlantic Ocean. The average annual temperature is 59° Fahrenheit, with the average winter temperature 38.9° F and the average summer temperature 76.8° F (Lewis 1975:141). Extremes in temperature do occur, however, both in the winter and the summer. Fifteen degree temperatures in January and February are not uncommon, and readings in excess of 100° F in July are predictable. The average annual rainfall is 42.5 inches, with occasional snow (Bick and Coch 1969).

8

THE FIRST EXCAVATIONS

The first archaeological excavations in the Peyton Randolph back yard were begun by the Colonial Williamsburg Foundation's Architecture Department in November 1938. The work was done under the direction of Francis Duke and was completed in the early fall of 1940 (The Archaeological Report is included in Appendix 3.) The primary objectives of the early excavations were to supply details for the reconstruction of the then no-longer standing east wing of the house, as well as for the most accurate possible restoration of the main house. One must assume the diggings in the back yard were aimed towards the eventual reconstruction of the various outbuildings, but only one small building was actually reconstructed. This structure, a shed, was constructed on original eighteenth-century foundations in 1940 (Photo 2).

The so-called "east wing" of the house was thought to have been a separate building, later connected to the original main structure by the central section. This building (Structure S) had been razed before modern memory, and does not appear in any of the early photographs of the Randolph House. The only evidence of the structure prior to the 1939 archaeology was its apparent presence on the Frenchman's Map and a "scar" left on the exterior east wall of the connecting central section (see Photo 3)—the only above-floor-level brick wall in the entire Peyton Randolph House.

Photo 2. - Reconstructed Outbuilding - Structure N

Excavations in 1939, prior to the construction of the house to its present appearance, revealed a 20 by 35 foot foundation some 15 inches thick forming a cellar under the east wing. The cellar had two levels of paving with a fireplace and chimney base built along its north wall.

Many photographs were taken during the 1930s reconstruction/restoration and archaeology. One of the more revealing was taken of the excavated cellar in the east wing (Photo 4), showing a dump truck in the background. This confirms the suspicion that much of the fill removed from the site during the excavations was transported elsewhere. Given this indication, it appears, with little leap of logic, 10 Photo 3. - East Elevation of House, Before Restoration Photo 4. - Excavation of the Foundations of Structure S in 1938
that the backfill dumped after completion of the 1930s archaeology may have been brought in from another site. This supposition significantly affected the treatment of the backfill during the 1982-1985 excavations.

In addition to the excavation of the east wing foundations, the work in the late 1930s attempted to clarify changes in the various entrances to the house. Most significant among these were the entrances to the basement under the older west wing of the house. The excavators located a bulkhead entrance in approximately the same location as the reconstructed one is today, and another just to the east. The first entrance had been altered several times. The oldest entrance was from the north, but was changed to the west with the addition of the covered passageway from the kitchen. Later, the access to the basement was moved to the east and a porch placed over the former bulkhead. It is unlikely that the present reconstructed porch-tower existed at the same time as the covered passageway from the kitchen.

The outbuilding area, or back yard of the Peyton Randolph House, was also investigated in the 1930s. East of the marl walkway, which currently exists extending northward from the entrance to the central section of the house, the foundations of four outbuildings were located. Three were superimposed upon each other and probably represent storage buildings. The fourth was located just to the south of these and was reconstructed as a shed. More information regarding these structures is included in the main body of this report.

Documentation of the early excavations was reasonably complete considering the standards of the time period. Somewhat detailed maps were made of the excavation areas, and several photographs were taken which proved invaluable to the later excavators. Unfortunately, a significant amount of information was lost due to the destruction of several builder's trenches and the interruption of 11 stratigraphic sequences. Also unfortunate was the apparent apathy towards artifacts which is reflected in this passage from the 1939 Archaeological Report:

Few fragments, and none of importance, were found among the outbuildings. The east wing excavation yielded some china fragments in unusually good preservation, some of them being almost whole pieces. But nothing was found of particular architectural interest. Fragments were consigned to the Educational Department (Archaeology) (Duke 1939:15).

THE SECOND EXCAVATIONS

Another section of the current area of investigation was excavated in 1955, by a crew directed by James Knight of the Architecture Department. The method of excavation had by this time changed from the relatively haphazard digging carried out in the 1930s to systematic cross-trenching. The area cross-trenched extended from the edge of the earlier excavations all the way to Scotland Street, a plot of several acres. Three twentieth-century houses were still standing when the cross-trenching took place. Although the presence of these houses impeded the trenching operation, oral history informants maintain that all three structures contained basements, and thus the archaeological record in the area of the houses was already obliterated.

The newly-instituted method of cross-trenching consisted of digging parallel trenches to subsoil, measuring a shovel blade in width and a shovel handle apart, at forty-five degree angles to the north-south layout of the city. Such a technique would easily locate brick foundations. When a brick foundation was encountered, the walls were followed, thereby exposing the building. Scale drawings were made of the foundations, but no record was kept regarding the location of the trenches. Seven structures and a well were discovered during the 1955 excavations. Some regard for recovered artifacts was expressed at this time, although the provenience is less than exact and many fragments were given to the people who lived in the houses on the lot. Interviews with the surviving members of the excavation crew, and the results of the later archaeology, indicate that some soil around the newly-discovered foundations was screened (Derry and Brown 1987). The cross-trenches were refilled with virtually the same earth that was removed (except where screening occurred). This fact, given the great amount of area that was cross-trenched in the Historic Area during that period, greatly influenced the treatment of the trench backfill in the 1982-85 excavations.

THE THIRD EXCAVATIONS

The third excavations were a preliminary investigation of a portion of one of the outbuildings located during the 1955 cross-trenching expedition. These were conducted in the winter of 1977-78 by the Department of Archaeology under the direction of Ivor Noël Hume. An interim report on the findings, written by Eric Klingelhofer, is included in Appendix 4. This brief excavation was the first carried out by a trained archaeologist and the first to give credence to a refined 12 stratigraphic sequence. In addition, Mr. Klingelhofer's notes on the excavations were organized using the Harris Matrix System, a concept that was also employed in the later work. These preliminary excavations were limited to the western half of Structure A, which was then covered with a temporary structure rather than backfilled.

EXCAVATION

The back lot of the Peyton Randolph House was the first major excavation carried out by the newly-established Office of Excavation and Conservation. The specific goals of the project were primarily to provide information leading to the reconstruction of the appropriate dependencies, walkways, fence lines, garden features, and work areas. In addition to these rather straight-forward "where", "what", and "when" questions, there was the opportunity to address questions regarding, for example, the value of data gathered from disturbed contexts, the significance of close spatial control (even piece-plotting) on a prolific urban site, and the uses of in-depth soil analysis. Because of these and other more open-ended research goals at Peyton Randolph, the system used was an adaptation of a system that British archaeologists Biddle and Kjøbye-Biddle (1969), Barker (1982), and Harris (1979) have been using for many years known as "open area excavation." Open area, or "block," excavation does not use standing balks between squares for stratigraphic reference. Although the method was new to Colonial Williamsburg archaeology, and most historical archaeology done in the Tidewater Virginia area, it is hardly new to America. This approach was taken in the 1950s by Binford at Hatchery West (1970), and was presented as a method for carrying out settlement/subsistence studies in American archaeology by Struever in the 1960s (Struever 1968).

The decision to eliminate the traditional balks was prompted by a change in some research goals. Hester, Heizer, and Graham, in their text on field methods (1975), suggest that the square and balk method has primarily been used to explore "vertical" problems such as chronological sequences of artifacts. The best available chronological description of colonial artifacts, Noël Hume's Guide to Artifacts of Colonial America (1969a), as well as a number of pamphlets and journal articles on the same subject, was a product of the square and balk method. However, Hester, Heizer, and Graham caution that "if one is seeking 'horizontal' information, then one should work for broad exposure of buried cultural remains, using large-area excavation" (1975:76). Because, as stated earlier, the purpose of the Peyton Randolph excavations is to offer a reconstruction of the back yard by detailing the nature of the spatial divisions and the use of particular areas in the past, broad horizontal exposure was a logical choice.

Vertical information need not be forfeited using this excavation approach. Sections formerly drawn from standing balks can be created from detailed elevation records and plan maps. Elevations are taken at least every 2.5 feet, more often in areas of complex stratigraphy. Many more cumulative section drawings can be generated from these records because one is not limited to those sections left 14 standing. Additionally, many archaeologists have found single standing sections "grossly misleading" (Barker 1982:44).

Open area excavation best suits large sites with complex stratigraphy, a description that aptly describes the Peyton Randolph site. Project goals required establishing the function of individual outbuildings or activity areas during various periods of use. This involved the isolation of living surfaces, features, and other deposits contemporary with the use of each building and the recognition of meaningful groups of features (macro-features). Of particular importance is the ability to accurately separate intact layers, then recognize and correlate contemporary stratigraphic units over a broad area. This has proven to be a complicated task for the Peyton Randolph back lot, where intensive occupation over two centuries has been combined with previous "archaeological" trenching to produce a very complex and disturbed stratigraphic sequence. Leaving many large unexcavated balks of soil would only obscure much of the stratification on the site, further complicating matters. To a field archaeologist trying to follow a consistent layer across a site, there is little difference between the destroyed stratigraphy in a 1955 cross-trench and the inaccessible strata in a standing balk.

In order to ensure that meaningful spatial patterning of the recovered artifacts at Peyton Randolph was recognized, a basic coordinate system was used to establish a grid. The north-south baseline was laid along North England Street, with "point zero," an arbitrarily selected point, near the corner of North England and Nicholson Streets. East-west grid coordinates were projected at ten-foot intervals from the north-south baseline. The site lies entirely within the northeast quadrant of the system. The ten-by-ten-foot units were laid out, and each was designated by its northwest corner stake. Each of these standard ten-foot squares was further subdivided into sixteen two-and-one-half foot sub-units. The exact horizontal and vertical coordinates of all diagnostic artifact groups, features and structural remains were also established using this grid system and a fixed datum point. Tightly controlled locational data was gathered so that, when appropriate, they could be subjected to computerized spatial mapping programs ranging from absolute three-dimensional graphic plotting to simulated two-dimensional distributional plotting (e.g., SASGRAPH). Such computerized mapping could supplement field identification of artifact concentrations and recognition of their broader spatial patterns through careful exposure in situ. Previous work on many historic sites has demonstrated the usefulness of these mapping programs in identifying specific patterns of refuse deposits and areas in which specific activities occurred. Many of these patterns and activity areas could not have been determined by visual inspection of excavation units alone.

Abandoning the extensive use of balks and increasing the control over provenience required some alteration in the recording system. Context numbers were given to both units of stratification and certain types of interfaces following the guidelines suggested in the Harris Matrix System, (Harris 1979:59). In order to reduce the size of the number labeled on individual artifacts, record spatial control, and facilitate computer mapping, each context number represents not only a particular unit of stratification, but a specific ten-foot section of that unit. For example, plowzone in the 250N 100E unit was given the context number 271, but the same layer in 260N 100E was given the number 272. Finer divisions of layers were obtained by dividing the unit into sixteen equal 2.5-foot-square sub-units, each designated by an alphabetic tag following the context number.

15

Once a layer or a feature was uncovered and defined it was given a context number. These numbers were assigned sequentially as the context was defined. Each context was described on the CONTEXT RECORD SHEET (Fig. 3). A description of the form and its contents are as follows:

01.	CONTEXT # -	self-explanatory
02.	PROJECT -	project name, e.g., PEYTON RANDOLPH
03.	DATE -	date context number assigned
04.	SITE -	block and area, e.g., 28G, 28H
05.	UNIT -	coordinates of the northwestern corner of the unit containing the layer or feature being described.
06.	TYPE -	what the context represented, e.g., plowzone, disturbance, layer, feature, post hole, etc.
07.	GENERAL DESCRIPTION -	brief description of the location and salient characteristics of the context
08.	STRATIGRAPHIC CONTEXT	relationship of the context with others
09.	SOIL CONDITION	relative amount of moisture in the soil
10.	SOIL COLOR	Munsell number
11.	LT. MTR.	f-stop and shutter speed of light when the soil color determined (using standard setting of 100 ASA)
12.	SOIL TEXTURE	relative amounts of clay, silt and sand
13.	SAND TEXTURE	sand grain size
14.	SOIL COHERENCE	degree of cohesiveness of soil particles
15.	MOTTLINGS	description of the five most prevalent mottlings in order of frequency
16.	ORGANIC REMAINS	choice of inclusions, if any
17.	INORGANIC REMAINS	choice of inclusions, if any
18.	BOUNDARY/INTERFACE DELINEATION	description of the area of contact between two contexts
19.	SKETCH #	number of all field drawings in which the context appears
20.	PHOTO #	number of all photographs in which the context appears
21.	SOIL SAMPLE #	site plus context number
22.	PEEL #	site plus context number
23.	WET SCREEN #	site plus context number
24.	RECORDED BY	three initials of the person filling out the form
25.	EXCAVATED BY	three initials of the person(s) excavating the context
26.	TPQ	terminus post quem date added to the record after artifacts have been examined by the lab
27.	METHOD	whether the context was excavated by unit, sub-unit, or individually, and screen size
28.	APPROVED BY	initials of the field technician who approved the information on the record
29.	ENTERED BY	initials of the person entering the data into the computer

16

Fig. 3 - Context Record Form

17

In order to enter this data into a computer for easy management, two masks were created using INFOSTAR+ (version 1.6, 1984). The CONTEXT RECORD SHEET was divided into two forms for data entry purposes because of the large amount of disk space required for the more than 3000 contexts. The first mask contained all data on the sheet except the STRATIGRAPHIC CONTEXT information. The second mask containing this data was used to construct a Harris Matrix (Harris 1979). Ideally, the data gathered daily from the field would be entered as the excavations proceeded; however, due to a variety of circumstances, neither the hardware or the program necessary to enter the information was available until the excavations were virtually complete.

Since the most important aspect of site interpretation is based primarily on stratigraphic relationships, a brief explanation of the Harris Matrix System is in order:

The primary goal of the study of the stratification of a site is to make a stratigraphic sequence. The stratigraphic sequence may be defined as the sequence of the deposition of strata or the creation of feature interfaces on a site through the course of time (Harris 1979:86).

Simplistically a matrix diagram of the major layers in 28G would appear as follows:

TOPSOIL
CW TRENCHES
PLOWZONE
BSL 1
BSL 2
PLANTING BED 1
SUBSOIL

As the sequences were determined by the excavators, field technicians, and the site supervisor, the layers and features associated with those layers were excavated. The STRATIGRAPHIC CONTEXT section of the CONTEXT RECORD SHEET was filled out with the appropriate information. Those context numbers which denote the same layer or large feature were given a MACRO-FEATURE designation. This designation applies to a group of context numbers which compose a unit of study, such as plowzone, CW cross-trenches, or a structure. Combining these units of study according to how each relates to the other is done schematically using the matrix diagram. The diagram can then be separated into "periods," showing the changes chronologically through the occupation of the area. Periodization is "the process by which the stratigraphic material from a site is arranged into periods and phases based upon stratigraphic, structural and artifactual data" (Harris 1979:125). It is from this periodization process that the sequence of major events in the history of the Peyton Randolph back lot can be reconstructed.

18

SOIL ANALYSIS

ARTIFACT ANALYSIS

Artifacts recovered in the field were brought into the O.E.C. Laboratory on a daily basis. As quickly as possible, the bags were emptied and washed according to established laboratory procedure. After washing, artifacts were dried, sorted into categories, and labelled with the site and provenience number. Oyster shell, brick fragments, and some iron pieces were counted and placed in temporary storage. The remainder of the artifacts were coded according to type. Computer cards were key-punched with the appropriate codes and fed into a mainframe computer at the College of William and Mary. The coded and entered artifact lots were left in open storage in Study Collection Rooms 1 and 2 until the end of the report-writing process. All faunal material was remanded to the Staff Zooarchaeologist, and any artifact in need of conservation was removed to the Conservator. Ceramic and some glass artifacts were mended and cross-mended under the direction of the Collections Supervisor, in consultation with the Site Supervisor. Until completion of all reporting activities, the artifacts remained in open access for analysis and study.

SPECIAL STUDIES

25

IV.A. STRUCTURE A - EARLY EIGHTEENTH-CENTURY DWELLING

PHOTO 5. - Structure A

Structure A was first excavated during the 1955 exploration of the Peyton Randolph back lot (Photo 5, above). It was probably a wood frame building measuring approximately 20 by 16 feet, containing a hearth in the northeastern corner. The building's sill was supported by nine brick piers. The presence of the piers was hidden from exterior view because the area between them was filled by a one-brick-wide string course. The bricks in the string course were of the same firing and size as those of the piers, so it may be assumed they were contemporary. The purpose of the "façade" of string brick was either one of aesthetics, making the building appear it was more substantial than it actually was, or one of practicality, perhaps helping keep unwanted animals from under the house. The filler bricks supported little weight since many of them were simply laid one on top of the other without benefit of mortar or the interlock needed to support a sill. The depth of the facade and piers varied from three courses in the east to five in the west. This compensated for the natural slope of the land towards North England Street, and for the continuing compaction of a seventeenth-century drainage/boundary ditch upon which Structure A had been constructed. The builders of Structure A were apparently well aware of the earlier ditch because the southwestern pier of the structure rested on a line of rowlock brick placed across the feature (Photo 6). The support brick were also the same as the building brick, evidence that they were laid before the house was built in anticipation of compaction problems.

Even though the foundations of the small structure had been previously excavated, much archaeological evidence pertaining to the construction of the building remained. A large support beam probably ran north to south under the 26 Figure 5 - Placement of Joist Nails, Interior Structure A 27 floor through the center of the structure since three central bricks lined up in this direction. No comparable arrangement ran east to west. Impressions of the sill had been left in the wet mortar on the exterior of the chimney base. The size of the sill was surprisingly small, little more than 4.5", based on the slight elevation difference between the wall foundation and the firebox or floor level. Several rows of upright nails found in situ in the southeastern corner of Structure A suggest wooden joists approximately two feet apart with floor boards running north-south (Fig. opposite).

Photo 6. - Rowlock Wall Under Foundations of Structure A

The small corner fireplace (Fig. 6, below right) was contained by the north and east exterior walls of the structure. Five courses of brick remained, and while the top three were well-mortared, the bottom two were held in place by a fine sand mixed with oyster shell bits. The bottom course rested on subsoil, and extra brick were placed around it to fill the gap between the base and the walls. The fourth course from the bottom contained one brick which was mortared to the top of the northeastern pier. Sufficient area remained on top of this pier, however, for the sill to have been in place when the chimney was constructed. The possibility that the fireplace was an addition to a standing building rather than original to its construction cannot be ruled out.

Figure 6. - Detail of Corner Fireplace, Structure A

The firebox consisted of one course of brick mortared to redeposited fill at the level of the fourth course from the bottom. The fill contained ash, plaster, mortar, and rubble, but no diagnostic artifacts. A circular area of the firebox was heavily burned, and the bricks were black, worn, and 28 crumbled. The hearth bricks in front of the firebox were mortared onto fill, and the outer or southwestern edge had fallen and slumped downward. The fill was similar to that under the firebox, containing ash, mortar, plaster and brick (only in lesser amounts). As few artifacts were found in the fireplace fill, no date was obtained for this part of the chimney feature either. However, one artifact of note was found in this fill: a manganese decorated delft tile fragment. This may indicate a repair or rebuilding of the hearth and fireplace, but more likely it was a residual artifact already in the fill and has no connection with Structure A's fireplace.

Photo 7. - Detail of Fireplace "Wing"

Also of note were two brick pads or "wings" on either side of the chimney base (Photo 7, above). Neither of these "wings" was mortared to the chimney base, although the two remaining courses in the northwest "wing" were mortared to each other. This mortar layer and the upper course of bricks extended northward over the former sill. Impressions of what is believed to be that sill remained on the underside of the mortar. Elevations taken on these bricks indicate a sill of approximately 4.5". The other bricks of the "wing" were mortared onto redeposited fill similar in color to that found under the firebox. Again, unfortunately, no diagnostic artifacts were found. The southeast "wing" consisted of only one course, and was badly damaged by a later posthole which destroyed part of the chimney base in the same area. Neither of these "wings" was designed to bear substantial weight, for unlike the chimney base they rested on redeposited fill rather than stable subsoil. One must assume these corners to the fireplace were flooring or low shelves filling the gap in the interior between the chimney and the walls.

The demise of Structure A is illustrated in soil differences evident in a balk left standing within the interior of the structure (Fig. 7). Under the topsoil and the remnants of a 1975-76 bus turnaround, the interior of Structure A contained four major layers. The uppermost layer (28G/81) was a mixture of brick rubble, mortar, and plaster with very little soil or artifacts. The western half of this layer was excavated in 1978. The second layer (28G/88) contained a concentration of plaster as well as numerous mid-eighteenth-century artifacts. Under the plaster was a layer of dark brown and yellow loam (28G/108) [Munsell: 10YR5/6], containing numerous artifacts and some charcoal. The fourth and last stratigraphic layer, before subsoil, was a medium brown loam [Munsell: 10YR5/3], which was probably the original topsoil present when Structure A was built. Unfortunately, it contained few artifacts. Eight cross-mends between the brownish yellow loam and the plaster layer indicate they were part of a similar or the same series of events. No cross-mends were associated with the other two layers, but this was not 29 Figure 7. - Profile of Standing Balk in Structure A unexpected since only 26 of the 434 sherds found within the structure came from the rubble and buried topsoil layers.

None of these layers can be physically associated with layers outside the structure since the 1955 cross-trenching followed the brick foundation walls all the way around the structure, leaving an island of undisturbed strata in the interior. The association was partially reconstructed through cross-mends. Several layers excavated outside the west wall of Structure A in 1978 contained ceramic fragments which cross-mended with the two artifact bearing layers on the interior. Figure 8 illustrates these layers, the thirty cross-mends, and reconstructed elevations. These exterior layers were described by Eric Klingelhofer in his "Interim Notes" (see Appendix 4) and include a thick mortar layer (ER2175F) sealing a half-inch deposit of brown loam (ER2175J) that contained numerous shell fragments, then another half-inch of ashy loam (ER2175K). These layers sealed a one-inch layer of loam (ER2175L) containing mortar, plaster, brickbats, and numerous fragments of delft tile. Under this tile layer was another ashy loam layer (ER2175M, N), sealing a large scatter of pipe stems (ER2179A). These pipe stems created a fairly flat surface over a two-inch layer of brown loam (ER2175T) and two deposits interpreted as the infilling of depressions in the brown loam (ER2171J and ER2175P). Below the brown loam and above natural was reportedly a three-inch layer of buff sandy loam (ER2175V).

Klingelhofer thought these layers represented the daily sweepings from Sir John Randolph's law office (pipe stem layer ER2179A), a renovation (mortar, plaster, and tile layers), a dower house for Lady Randolph (loam layers ER2175J and K), and then demolition (mortar spread ER2175G). However, cross-mends illustrated in Figure 8 indicate that everything from the ashy loam (ER2175M) beneath the delft 30 Figure 8 - Cross-Mends to 1955 Trench Fill, Structure A 31 tiles up to the mortar spread (ER2175G) was destruction debris. Though this was described as five layers, only two inches existed between the mortar and the ashy loam. In fact, all of these layers also cross-mended to each other (Fig. 8). This certainly supports the argument that all these layers represent one event of relatively short duration rather than renovation, use, and destruction.

The association becomes even clearer when horizontal distributions are considered along with the vertical ones. Figure 9 combines a distribution map from the 1982 excavations with a sketch map from the 1978 excavations. The 1982 section shows the distribution of ceramics inside Structure A provenienced to 2.5-foot sub-units. The 1978 section shows the placement of two recognized artifact scatters—mortar and tiles. Taken together, they illustrate a localization of trash disposal. It would appear that after the wooden structure was removed the brick foundation stood open while dumping occurred near the south half of the west wall, both inside and outside the foundation. Perhaps this was an attempt to fill the gaping hole left by the removal of Structure A and to level the downslope west of the foundation. All this material was functioning as landfill and must have been trash from another structure, probably the main house, given the quality of the items represented. The mortar/plaster layers may represent debris generated from the construction of the mid-section of the main house. Since the point of dispersal for the debris appears to have been N. England Street near the southwestern corner of Structure A, this could indicate that a cart was used to haul in the trash from some distance away. It is thus likely that these deposits were secondary refuse and not directly relevant to the interpretation of Structure A.

The same may be true of the plaster layer inside Structure A (28G/88). Although the cross-mends seem to indicate the ceramics in the plaster layer and the layer below (28G/108) were part of the same deposit, the artifact distributions are different (Fig. 10). Artifacts were concentrated in the southwestern corner in both layers but are more densely distributed at the lower level and more spread out in the plaster. The most plausible explanation is that the lower layer was the product of actual dumping; when the plaster was added, it was raked smooth to cover and seal the derelict foundation/landfill. Spreading of the plaster dislodged and dispersed some of the trash beneath it, indicating that the plaster and the rubble may have been secondary refuse also.

With many of the layers on the exterior of Structure A excavated in 1978 now reassigned to the post-occupation period, the functional interpretation of the structure as a law office and dower house as proposed by Klingelhofer has to be rethought. The only layers excavated in 1978 that were relevant to the occupation of Structure A are the pipe stem layer (ER2179A), the two-inch layer of brown loam (ER2175T), and two deposits cutting this loam (ER2171J and 2175P). Three inches of buff sandy loam under the brown loam were excavated but few artifacts were found.

The building was occupied some time after 1720 and destroyed before 1770, probably between 1755 and 1765. The occupation layers produced a terminus post quem of 1720 based on the presence of white salt-glazed stoneware and Yorktown pottery. Three fragments of clouded ware and one of refined agate-ware were found in the debris layers. These sherds provide a TPQ of 1750, and the absence of cream-ware, popular after 1770, suggest destruction occurred before that date. This destruction date could be reduced to 1755-1765 based on mill-edged clouded ware 32 Figure 9. - Structure A, Artifact Scatters 33 Figure 10. - Distribution of Selected Artifacts in Plaster Layer and Below Plaster Layer 34 plate fragments. According to Colonial Williamsburg's Resident Archaeologist: "Archaeological studies have so far failed to establish the presence of … the clouded glaze wares, in American homes before the mid-1750s … [Also] the milled edge seems to have been popular in the period 1755-1765" (I. Noël Hume 1970:409).

The uppermost occupation layer (ER2179A) was described by Klingelhofer as "a 1-inch thick layer of soft loam containing in some places more pipe stems than soil … small pieces of bottle glass were found among the pipe stems." It covered a "fairly flat surface presented by infilling of … depressions." It is suggested that this represents repair work on the ground in preparation for a pipe stem paving similar to the one reported by Audrey Noël Hume at the Hugh Orr dwelling, also in Williamsburg. This paving, too, was composed of fragments of clay tobacco pipes mixed with small pieces of window glass and a number of glass and ceramic items (A. Noël Hume 1963:22). Almost none of the pipe stems near Structure A show any sign of having been smoked, indicating they were probably wasters or part of a damaged shipment transported to the site for use as paving.

Figure 11. - Oswald's Simplified Typology

Since these items were part of a walking surface, the pipe bowls from this layer were fragmentary, but enough could be pieced together for comparison to Adrian Oswald's simplified general typology (Oswald 1975:35-38). The overall form of these bowls is most similar to Oswald's type 11 (Fig. 11), but the bases were somewhat smaller than the one illustrated. Oswald reports that bowl type 11 has a date range of c. 1730-1760 and has a base that declines in size over time. Therefore a date range for the smaller-sized bases could be narrowed to about 1740-1760.

In 1978 the Binford mean date formula was applied to the diameter of the holes in the stems. Laboratory personnel derived a mean hole diameter of 4.18 for 35 1,506 fragments, producing an extrapolated date of 1772.69. The stems were remeasured in 1983, expanding the sample to 1,789; a date of 1774.06 was determined. However, neither estimate likely represented the actual date of deposition of a layer characterized by 1740-1760 pipe bowls and sealed by several layers dating to the 1750s or 1760s. Rather than declaring Binford's formula invalid, the nature of the sample was evaluated. The sample was indeed suspect since it was probably a shipment of wasters, and was at any rate probably from the same shipment rather than the assorted domestic garbage from which the formula was derived.

The sample was also compared to those above and below it, and to the pattern established by J.C. Harrington and presented by Ivor Noël Hume in A Guide to Artifacts of Colonial America (1969a). This set of data is presented in Figure 12. Figure 12. - Pipestem DATING Comparisons to Harrington's Charts and Binford's Regression Formula Dates
All the other samples date to the 1750s or 1760s and conform to the patterns presented by Noël Hume. In fact, even with relatively small samples, the other layers sort out in an acceptable pattern. Those predating destruction predict 1750 dates while those believed to have been deposited after demolition predict 1760 dates. Consequently, the pipe stems in the paving must represent an abnormal collection. Either the pipes were wasters and the abnormally small bore diameter relates to their status as rejects, or if they represent one shipment, perhaps 36 damaged in handling, the averaging effect of several shipments reflected in domestic garbage may be necessary for Binford's regression line formula to be applicable.

While the dating estimate based on bore diameters could not be used in this unusual case, the estimate of 1740-1760 based on bowl type seems acceptable. Destruction layers believed to date to 1755-1765 sealed the deposit, and the pipe stems overlay a portion of another deposit that may date to the 1740s. This latter deposit was described in the 1978 field notes and is believed to be the westernmost extension of the grey ash layer found in 1982 under the clay walkway bed north of Structure A. This ash contained Astbury ware, a ceramic type most popular in the 1740s (I. Noël Hume 1969b:35). The occupation layers, also beneath the pipe stem layer, contained Yorktown pottery, a type out of production by 1745.

If this stratigraphic relationship is accurate, then the pipe stem scatter may have been part of an up-grading or formalization of pathways in the 1740s or soon after. Judging from the presence of an occupation layer (ER2175T) beneath the pipe stems, Structure A must have been occupied before the addition of that paving. There is no evidence that the addition of the paving accompanied any major change in the use of the structure. Both the brown loam (ER2175T) below the pipe stems and the loam (ER2179A) surrounding the pipe stems contained few artifacts compared to the destruction layers above. Ceramic fragments totaled 141 and 130, respectively, while the secondary trash deposits above contained 2,279 fragments. These two layers (ER2175T and ER2179A) were also similar in ceramic content. Refined- to-coarseware ratios were almost identical for these two deposits, approximately 85 to 15.

Figure 13. - Ware Types - Specific Categories

In an attempt to uncover the function of Structure A, the occupation layers were analyzed according to several criteria: specific ware type, context of use, and vessel form.

Figure 13 describes ware types according to specific categories. The counts from the destruction layer inside Structure A were also considered here for, comparative purposes. These graphs show that the occupation layers consisted mostly of delft and white salt-glazed stoneware. The destruction layer had more Chinese porcelain, and also a wide variety of types. This would probably be true of trash hauled in from the street. This secondary deposit contains a greater diversity of material since it was composed of objects discarded, perhaps from a 37 Figure 14. - Ware Types - General Categories
number of original sources. By contrast, the sparse collection from the occupation layers is the product of behavior specific to Structure A.

A summary of the various artifacts from the occupation layers is provided in Figure 14. These items are grouped according to functional categories and broader contexts of use. This perspective, modeled after Lyle Stone's Fort Michilimackinac report (1974), was taken to provide information on the specific use of Structure A.

The most plentiful artifacts in the uppermost occupation layer were pipe stems (59%) and window glass (16%). However, these were the major constituents of a paving surface and were less plentiful in other occupation layers (pipe stems 24%, glass 6%). The pipe stems showed no sign of having been smoked. Another eighteenth-century walkway uncovered in Williamsburg reportedly also contained "fragments of clay tobacco pipes mixed with small pieces of window glass" (A. Noël Hume 1963:22). Apparently in both cases, these two items did not serve their usual functions in personal and structural contexts. This fact is even more apparent if one compares percentages of artifacts by these functional categories with and without these paving materials.

PAVING LAYER — ARTIFACT TYPES

	% of Artifacts:
Context of Use	Paving Layer/Remaining Occupation Layers
	with Paving Materials	without Paving Materials
Personal	32%/70%	10%/14%
Household	63%/10%	89%/84%
Structural	6%/20%	1%/ 1%
Crafts	0%/0%	0%/ 1%

39

As this breakdown illustrates, eliminating the paving materials from the data set brings the percentages closer together, making the two occupation layers similar in all categories.

The percentages shown in the chart represent totals for all the occupation layers minus the paving materials. This occupation material, most likely discarded or lost near its place of use, reflects an unexpected function for the building. This Peyton Randolph-period outbuilding, where one may have expected a craft or service activity to be carried out, appears, instead, to have been the site of domestic activity. In fact, less than one percent of the artifacts could be attributed to a craft, while an overwhelming 86% reflect a household context. This percentage, taken with the 13% representing personal items, implies this building was either a residence or at least a place where food and beverage consumption occurred.

Ideally, these counts should be in numbers of vessels rather than fragments. However, the material was quite fragmentary (as is usually the case with occupation layers constantly walked upon), thus accurate vessel counts were difficult to obtain. The wine bottle necks and bases sorted into a minimum of only nine bottles, and the many small ceramic sherds were mended into only seventeen vessels:

CERAMIC VESSELS FOUND IN PAVING LAYERS

Ware	Vessel Form	No. of Fragments
Colono-ware	Bowl	1
Delft	Punch bowl	16
	Punch bowl	22
	Plate	1
	Plate	4
	Tankard	1
	Ointment pot	5
Chinese porcelain	Shallow bowl	1
	Plate	1
Coarseware	Milk pan	14
	Milk pan	1
Brown-glazed earthenware	Pitcher	2
	Pitcher	2
Fulham-type stoneware	Tankard	1
White salt-glazed stoneware	Tankard	29
	Tankard	15
	Teapot	41
	Teapot	3

40

Only two of these vessels are milk pans; one is an ointment pot, the remainder are teawares or tablewares. Evidently much more food and beverage consumption took place in Structure A than food preparation, which is not surprising considering the huge kitchen complex on the lot and the small hearth within Structure A.

The identity of those participating in these activities is not at all clear. The person(s) residing in, or making use of, Structure A prior to the 1755-1765 destruction date, was apparently taking food from delftware plates and serving tea from white salt-glazed teapots. Tea drinking during the first half of the eighteenth century was relatively restricted, as the commodity itself was highly priced and teawares were not plentiful or readily available (Roth 1961:66; Shammas 1980:14). This fact, coupled with the presence of mirror glass, another expensive item during the period, suggests that those residing in or using Structure A enjoyed a social position above that of the great majority of domestic slaves, but far below that of the Randolphs themselves, whose teawares and dinnerwares at the time include a number of export porcelains, such as blue-and-white and Imari styles. If Structure A was actually lived in, the occupant could have been a tutor, craftsman, overseer, or favored slave.

In summary, Structure A remains enigmatic. The occupation layer contained no artifacts that would typify any specialized function; however, chronological questions have been answered. It was certainly demolished during Peyton Randolph's lifetime, possibly after his mother's death, and most likely between 1755 and 1765.

The artifacts and stratigraphy in and around Structure A were analyzed more intensely than those associated with any of the other structures in the Randolph lot. The reasons for this approach to the analysis of Structure A material was simply because it was the only structure excavated that contained or was associated with a significant number of artifacts that could be identified as relating to the structure and reasonably intact stratigraphy. A similar analysis of the other structures is not possible because of the relative lack of these important components. Further analysis of Structure A material, and that from other parts of the site, has been accomplished and is continuing. Such analysis is, however, beyond the scope and purpose of this descriptive report. Another report, dealing with detailed and interpretive analysis of several aspects of the Peyton Randolph archaeology, is now in progress and is scheduled for completion in 1990.

41

IV.B. STRUCTURE B - OUTBUILDING

PHOTO 8. - Structure B

Located only 0.35 feet south of Structure A, though at a higher stratigraphic level, were the scant remains of a small building designated Structure B (Photo 8). The foundations were situated within units 230-240N 10-20E. The north wall was the most intact of the three remaining; it was ten feet in length, one and a half bricks wide, and constructed in English bond. The east wall was also in fairly good condition, 8.2 feet in length, also one and a half bricks wide, and constructed in English bond. The south wall was completely robbed, probably in the mid-nineteenth century. However, a row of seven bricks, laid flat and oriented north-south, was found along the exterior of the south wall's original location. These bricks may represent the remains of steps leading to a door on the south side of the structure. Only what seems now to be a pier remained of the southwestern corner of the building. This pier was three courses high, of indeterminate bond, and measured 1.6 feet north-south and 1.3 feet east-west. No traces of the west wall, save the southwestern corner, remained. Exterior dimensions of the small structure were eight feet north-south by ten feet east-west. Interior measurements were 5.4 feet by 7.8 feet.

The building was first uncovered during the 1955 excavations, with the exception of the southwestern corner, which was missed by the cross-trenches. At that time a 3.6 foot square brick appendage was centered along the north wall of the structure (Fig. 15). The appendage rested on fill above the foundation wall of nearby Structure A. Although the appendage appears on the Knight Map and seems to be a small chimney base, no traces of it survived the construction of the bus 42 Figure 15. - Brick Feature over South Wall of Structure A (from J.M.K. map)
turnaround in 1975. Knight coded the structure as "second period colonial" on his overall map, but coded the appendage as "third period." The appendage was necessarily constructed after the demise of Structure A (c. 1760). Structure B itself, however, may have been built while "A" was yet standing, although this is unclear since no builder's trenches which could closely date Structure B's construction were left intact after the 1955 excavations.

Structure B was constructed on the Brown Sandy Loam 2 (BSL 2) layer of sheet refuse which built up over the 1720 to 1800 period. The southwestern corner of the structure, previously undisturbed by the 1955 excavations, was sealed by Brown Sandy Loam 1, the sheet refuse layer which followed BSL 2 and which was deposited primarily in the first half of the nineteenth century. A layer of rubble and plaster (28G/ 1752) within the interior of the structure could represent the destruction of the building. This rubble has a TPQ of 1820. Given this evidence, it appears that the building was probably constructed in the third quarter of the eighteenth century and survived until the second quarter of the nineteenth.

43

The function of Structure B is not easily determined. The predominant artifact type within the structure was bone. Two fully articulated bird skeletons (a duck and a rooster) were excavated from the interior (Photos 9 and 10). However, these animals may have been yard fowl that hid in the building and subsequently died, revealing nothing about the actual function of the structure. The structure and associated layers were so highly disturbed in this area that little other than a description of the building as it existed when found can be reported upon with any certainty.

Photo 9 - Articulated Remains of Rooster

Photo 10 - Articulated Remains of Duck

44

IV.C. STRUCTURE C - OUTBUILDING

PHOTO 11. - Structures C and D

In the northeastern section of the 28G area, in units 250-70N 70-80E, the foundations of two structures were located (Photo 11, above). The larger and later of these two was Structure C. Structure C's foundations measured twenty feet north-south by sixteen feet east-west. As with most other Williamsburg buildings, the foundations were oriented on the north-south, east-west axis. The strata associated with Structure C were highly disturbed by various activities in and around this area, especially on the interior of the building.

The interpretation of the surrounding strata was complicated by strata associated with the earlier Structure D underlying the area, and by overlying layers and features associated with three major occurrences. The first of these was the plowing of the 28G area during the mid-nineteenth century. This resulted not only in the disruption of the area's stratigraphy by the act of plowing itself, but in a systematic redistribution of bricks from various other locations. In this instance, the bricks were moved from an area being plowed to an area which was not being plowed, i.e., the area around the then-newly-planted pecan trees. The trees themselves were disrupting agents, as the roots from the easternmost tree caused a substantial amount of damage to the western walls of both Structures C and D, and also damaged the northern walls. The roots and the trees made excavation of the area impractical and destroyed a large amount of the stratigraphic record relating to the structures. Lastly, cross-trenching undertaken throughout 28G in 1955, and the subsequent uncovering of the brick foundations by the excavations, interrupted the stratigraphic connections between interior and exterior as well as destroying almost all of the builder's trenches.

In combination, these disruptions caused a serious clouding of the stratigraphic relationships in the area, making determinations of function and period of use much more difficult than they otherwise would have been. Fortunately, enough of the stratigraphic record remained to make some strong correlations.

The foundation remnants of Structure C's four walls present an intriguing picture because none of them match. The northern wall was composed, for the 45 most part, of stretchers made from brick bats. The western wall had headers along most of the wall along with nine rowlocks at the north end. The east wall was made up of headers. Only the south wall appears to have been constructed with workmanship in mind. This southern foundation which was constructed in English bond and was the wall facing the main house.

A small amount of hard-packed marl in 260-70N 80E, measuring about 0.2-0.3 feet in depth, was probably all that remained of the floor to Structure C. The marl was found under the destruction layer of Structure C, and was above the destruction layers of Structure D and Brown Sandy Loam 2. As there are no other reasons for a layer of this type in this location it would seem this marl was once a part of the floor to Structure C. Among the artifacts found in the marl were white salt-glazed stoneware, slipware, and blue underglazed Chinese export porcelain.

The major occupation layer with which Structure C was associated is Brown Sandy Loam 1, a sheet refuse layer that dates from about 1775 to the mid-nineteenth century. (Please refer to section IV.J. of this report for a detailed discussion of this layer.)

A combination builder's/robber's trench remained along the northern wall of Structure C. This trench was probably used to rob bricks from the north and east walls of Structure D as Structure C was being constructed. This trench was sealed by a thin layer of marl and filled with brick rubble. After removal of Structure C's walls, excavation uncovered a fragment of pearlware, thus placing the construction date of Structure C to post-1779. This suggests that Structure C was most probably constructed under the ownership of Joseph Hornsby (1783-1800).

There is no hard evidence available to aid in the determination of the structure's specific function. It would seem, however, that Structure C was a storage building of some sort. This conclusion is due in part to the relationship of Structure C to Walkway # 2. Walkway # 2 undoubtedly serviced Structure C, evidenced by the brick paving left unplowed between the two pecan trees. The substantial nature of Walkway # 2, which is similar in this respect to Walkway # 1 (both discussed in detail in another chapter), shows that it could accommodate carts and wagons as well as foot traffic. It is probably safe to conclude that Structure C's function may have been that of a storage building which may have needed servicing by just such a passage. The artifactual and structural remains give no indication of Structure C's use as a smokehouse, dairy or other service building.

Structure C fell out of use at or just before the mid-nineteenth century, when the area was brought under the plow. A large number of nails were found about the perimeter of the building. This, along with the lack of consistency and stability of the foundations, makes it virtually certain that Structure C was a wood frame building.

46

IV.D. STRUCTURE D - OUTBUILDING FOR STRUCTURE A

PHOTO 12. - Structure D

Structure D was located inside and below the remains of Structure C, principally in units 260-270N 80E. The foundation remains of the structure give the impression of a small, humble building, measuring 12.25 feet north-south by 11 feet east-west on the customary Williamsburg orientation axis (Photo 12, above). The four walls were all of a rather disjointed nature. The north wall, lying under the north wall of Structure C, was apparently partially robbed or destroyed during the construction of the latter building. The western wall was not excavated during the 1950s (or during the current excavations) because of its proximity to the easternmost pecan tree on the back lot. The eastern wall was a virtual farrago of bricks, brick bats, and mortar from both Structures C and D. The south wall was the only one of the four that was not highly disturbed, even though the western portion of this wall had been removed at some point, probably also during the construction of Structure C. All four walls of Structure D were constructed almost exclusively of brick bats, with the few whole brick used mostly as stretchers. There was nothing remotely resembling any type of bonding style in any of the walls, as the reused brick in the foundation had simply been mortared together with shell mortar.

Structure D was first excavated in 1955 during the cross-trenching carried out in the area. This cross-trenching resulted in removal of all the builder's trenches and all but one of the robber's trenches. The plowing of the area affected Structure D just as it did Structure C, disrupting the strata and causing a redistribution of a large quantity of brick into the area.

No remains of an above-ground floor for Structure D were found. However, what appears to have been a "living surface" was isolated in the interior of the 47 structure. This living surface was composed of three lenses, found only inside the foundations, and part of a major layer which was encountered in both the interior and exterior of the structure. The uppermost of these three components was a charcoal and brick layer which was located immediately under the destruction layers of Structure D and was, in fact, mingled with them. The brick in this layer was probably a result of the mixing. The soil matrix was a dark brown sandy loam [Munsell: 10YR4/3] with heavy charcoal scatter. Some of the artifacts removed from this layer date as late as the first half of the nineteenth century, but their presence was probably a result of destruction activity intruding into this charcoal layer. Other, more representative ceramics recovered from the layer included slipware, white salt-glazed stoneware and North Devon gravel coarseware.

The next level of the living surface was composed of a brown layer [Munsell: 10YR5/3] and a dark yellowish brown layer [Munsell: 10YR4/6], both found only within the interior of Structure D. The brown layer contained several different artifact types, all of which dated to the early part of the eighteenth century. These included Westerwald stoneware, Colono-ware, white salt-glazed stoneware, various coarsewares, slipwares, and black-glazed redware. There were no artifacts found in the darker layer.

The remaining layers (those lying below the so-called living surface) were all stratigraphically below Structure D with the exception of Brown Sandy Loam # 2. The BSL 2 layer occurred both in the interior and the exterior of Structure D. It was, however, much thinner on the interior of the building than on the exterior. This thin lens of BSL 2 seems to represent deposition prior to the construction of the building. The remaining, thicker exterior deposition of BSL 2 corresponds to the occupation of Structure D.

A great deal of the interpretation of Structure D depends upon the stratigraphic relationships of the various soils. From a comparison of the layers on the interior and exterior it is fairly easy to determine the way in which the layers associate with the various phases of the building (construction, use, and destruction). The construction of Structure D took place after BSL 2 began to be deposited but before the deposition was complete, evidenced by the thin layer of BSL 2 in the interior and the thicker build-up on the exterior. BSL 2, a sheet refuse layer, was first deposited about the time of the major development of the property in the second decade of the eighteenth century. Structure D was probably constructed shortly afterward. The layer best related to the occupation of Structure D was this same Brown Sandy Loam 2. BSL 2 extended over most of 28G and was the lowest stratigraphic layer found in the area. The layer of sheet refuse was the predominant cultural stratum on the whole site from about 1720 to the end of the century (see Chapter IV.J.)

The destruction layers relating to Structure D were of various types, the common denominator being brick rubble. A sizable portion of this brick rubble probably resulted from the numerous disruptions in the area. However, as some of this rubble was directly sealed under the marl floor to Structure C, it is safe to assume that this rubble came from the razing or falling of Structure D. The layers most closely associated with this destruction activity also contained a vast quantity of "destruction-type" debris such as mortar, nails, window glass, etc. These layers also contained ceramics which were helpful in dating the destruction event. Included among these ceramics were agateware and creamware. The presence of 48 creamware indicates that Structure D could not have been torn down prior to 1762, when creamware was introduced.

With the exception of one possibly intrusive sherd of Bennington ware, there were no artifacts of a later date. The lack of later artifacts leads to the assumption that Structure D was probably torn down in the 1760s or 1770s, and certainly before the building of Structure C.

The ash content of the soils on the interior of Structure D seems to indicate that the building may have functioned as a smokehouse. The dimensions and the meagerness foundations of the structure indicate that it was wood framed; a large quantity of nails found around the perimeter tends to substantiate this interpretation.

In summary, Structure D was probably a wood-framed service building, possibly a smokehouse, which stood for approximately fifty years between the 1720s and the 1760s or 1770s when a major re-orientation of the back yard occurred. This re-orientation was based on the construction of the mid-section of the main house and the subsequent construction of Structures C, E, and F, as well as the re-building of walkways to service them.

49

IV.E. STRUCTURE E - GRANARY (?)

PHOTO 13. - Structure E

Structure E is located 21 feet south of the south wall of Structure A (Photo 13, above). The building was almost totally excavated in the 1955 archaeological investigation, with only a five-foot segment of the east wall being inaccessible at the time. Most of the structure was again excavated in 1984, with the exception of the southern quarter, which was left untouched because of its proximity to the access driveway for the Randolph House life-tenant and the presence of a large walnut tree.

In its entirety, the building measures approximately 28.5 feet and 18.5 feet. Only the east, north, and west walls were uncovered; the south wall (if still intact) remains located under the access drive. None of the three walls uncovered bore any resemblance to one another. The east wall appeared to be the earliest of the three; nineteen feet of it was excavated in 1984. The wall was a brick and a half in width, set in English bond, and fairly well-constructed out of principally whole, previously-unused brick. The northeastern corner had been destroyed, making the relationship between it and the north wall somewhat ambiguous. The north wall was only one brick wide, one course deep and truncated in several places by subsequent disturbances in the area. No particular bonding style was recognizable in the remains.

The west wall was obviously built much later than either the east or north walls. It was only one brick wide, consisted of mostly used brick bats, poorly 50 constructed (not even straight), and casually bonded with sand mortar. The west wall was constructed over the north wall at the northwestern corner.

In addition to the walls, the building contained a series of piers running down the, center the length of the structure. Three piers were located, and another virtually destroyed, by cross-trenches during the 1955 excavations. Three others were found during the 1984 excavation. The piers were constructed on 2.5 foot centers and composed of new brick with shell mortar. Fill around one of the piers had a TPQ of 1779. This TPQ ties in closely with the underpinning of the "Granary" by Humphrey Harwood in August, 1785 (see Appendix 2). If these piers do represent the work of Mr. Harwood on that date, the function of the building during that period is established historically.

Photo 14. - Detail of Collapsed Pier, Structure E

In addition to having been underpinned, the building was later partitioned in the center from east to west. Two short appendages were constructed at right angles abutting the interior of the east and west walls at the center of the building. An east-west 3.5 by 0.7 foot pier was also constructed in the center of the interior, either to partition the building into two equal halves, or to brace the center of the floor. Whatever the case, the center pier obviously bore considerable weight as it was actually pushed into the ground several inches (Photo 14, above). The weight was concentrated in the center of the pier as if from a support beam which ran north-south down the middle of the structure, rather than east-west across it. The two short corresponding appendages attached to the east and west walls show no evidence of excessive weight from the floor above.

51

With the function of Structure E possibly established as a granary, the temporal span of the building's use can be addressed. The older east wall and the center piers for Structure E were built on Brown Sandy Loam 2, a sheet refuse-bearing layer which was deposited over a long period of time during the eighteenth century. Planting Bed 4 was also located under the foundation walls and brick piers. This planting bed was probably in use up until the fifth decade of the eighteenth century. No compaction into the planting bed fill was exhibited in the foundation wall or piers to the structure, indicating the bed had some time to stabilize before Structure E was erected. If it is assumed that Structure E is the granary referenced in Harwood's ledger, then the building had been standing long enough in 1785 to require repairs. Allowing ten years for compaction and ten years for deterioration, it may be conjectured that the structure was built between 1750 and 1775. The clay base to what appears to be a work area around the structure has a TPQ date of 1762, conforming to the proposed construction date.

All of the north, east, and west walls, as well as most of the piers, had to be removed in order to excavate Planting Bed 4. Removal of the west wall revealed several post holes directly under the wall which appear to be a continuation of a long-standing north-south fence line at the western edge of the property. The fence was probably maintained even after the west wall for Structure E was constructed. This is evidenced by the observation that other post holes associated with the many repairs to the fence line are more numerous and often have later TPQ dates than the post holes under the west wall. The original west wall of Structure E many have been of post construction, as no evidence of a brick wall previous to the one which remained was found. In any case, the west wall of the structure seems to have always been used as part of the western fence line. (For a more detailed discussion of this fence line, refer to Chapter V.N. of this report.) Soil excavated from under the west wall of the structure contained one fragment of whiteware, producing a TPQ of 1820, thus suggesting that the west wall must have been constructed after this date. Unfortunately, none of the artifacts removed from under the east wall could reveal a more precise TPQ than "eighteenth century." Soil under the north wall, however, did contain pearlware (1779) as the latest datable artifact.

The demolition date of Structure E is based on several events occurring in the immediate area around the building. A post hole (28G/ 1290) which was part of a nineteenth-century east-to-west fence line separating the plowed area to the north from the yard area to the south intruded the east wall of the structure. This post hole had a TPQ date of 1880. In addition, plowing in the 28G area is thought to have been started around 1860, possibly in response to shortages caused by the Civil War. Most of the northern half of the area formerly occupied by Structure E was plowed. In all likelihood, the often repaired, ill-constructed building was demolished or fell down some years earlier.

52

Figure 16. - Structure F (from J.M.K. map)

53

IV.F. STRUCTURE F - UNIDENTIFIED BUILDING

The brick walls designated "Structure F" were not investigated in the 1982-1985 excavations. The structure (or structures) were located under the access driveway, prohibiting the excavation of the area without the expense of relocating the life-tenant's access drive (Fig. 16). All that is known of the brickwork is derived from the Knight map of the 28G area, cross-trenched in 1955. This map depicts brickwork from two different periods: "colonial" (first period of construction) and "modern" (fourth period of construction).

The depicted colonial brickwork suggests a building at least 24 feet north-south and 12 feet east-west.¹ Though fragmented into five pieces, the east wall was more intact than the west, of which only three short sections in the northwestern quadrant of the structure still existed. No intact corners were located in the cross-trenching. The colonial period walls were in English bond with shell mortar, approximately 0.8 feet in width.

The so-called "modern" brickwork indicates a structure approximately 26.5 feet north-south. Only parts of the west wall existed, so the width of the building is unknown. Knight coded the west wall of Structure E "modern" as well, and upon re-excavation this wall was deemed to date to the mid-nineteenth century. It is likely, therefore, that the west walls of both E and F were contemporary.

Unlike Structure E, Structure F was probably actually two separate buildings. Structure F-1, the older of the two, was located about six feet further south than the more recent Structure F-2. It is possible that the second structure shared the east wall with the first, but this can only be substantiated through an archaeological excavation of the area.

Though totally conjectural, Structure F could be the lumber house mentioned in Peyton Randolph's probate inventory (see Appendix 1).

54

IV.G. TOPSOIL

A grey-brown sandy humic topsoil [Munsell: 10YR5/2] lay below the sod in the 28G area, sealing the entire site. No structures or features were visible from the ground surface, and though the area grades slightly from east to west, little or no erosion appears to have taken place in recent years. Thickness of the topsoil varied from 0.15 feet in the east to 0.75 feet in the west. Artifacts recovered from the topsoil layer range from modern plastics and foil wrappers to residual finds from the eighteenth and nineteenth centuries.

Figure 17. - South Wall - 1982 Excavation Area

Evidence suggests this topsoil layer was brought from another area and deposited over the site, probably after 1972. The section (Fig. 17) along the southern edge of the 1982 excavation area indicates three layers of fill occurring after the James Knight cross-trenching in 1955. One is a topsoil layer approximately 0.1 foot thick covering the trench fill; a second is a clayey orange layer which is believed to be the driveway of the Miller House, occupied in 1958 (see Photo 15). This drive was probably used until the house was removed in 1972, apparently after which nearly half a foot of topsoil was dumped and graded over the site. The depth of deposition of this secondary fill was less in the northern section of the site, away from the Miller foundations, suggesting Colonial Williamsburg's Landscape Department felt less fill was needed there to level out the property. Neil Black, miller at the nearby Windmill, remembers fill dirt being trucked in for this purpose (see Oral History, Appendix 7, by A.K. Foster). 55 Apparently the process was repeated after the 1975-76 bus turnaround was discontinued as a great deal of topsoil was found over this area. All of these modern sub-strata were given the topsoil designation and context numbers.

Two important conclusions can be reached after the examination of the topsoil. The first is that most of this soil was brought in during the past fifteen years, and is thus totally unrelated to the use of this lot in the eighteenth century. The second is the obvious alteration of the landscape in the area by the successive infilling, which must be a consideration when reconstruction takes place.

Photo 15. - Miller, Skillman, and Donegan Houses

56

MODERN DISTURBANCES

The integrity of the topsoil layer in the 28G area had been interrupted several times in the recent past by utility lines, archaeological cross-trenching, rodent burrows, tree removal, construction and destruction of twentieth-century buildings, and the construction of a gravel bus turnaround for Colonial Williamsburg tour buses. The archaeological cross-trenching is discussed in detail by Derry and Brown (1987) in a paper entitled "Excavation at Colonial Williamsburg Thirty Years Ago: An Archeological Analysis of Cross-Trenching Behind the Peyton Randolph Site."

The bus turnaround was constructed in 1975 to accommodate the increase in visitation to Colonial Williamsburg during the country's Bicentennial celebration. It consisted of two layers, a top layer of dark orange clayey sand which was extremely hard packed, presumably from having been continually compacted by the Colonial Williamsburg buses. The second layer was a lighter orange layer of sand, pebbles, and marl, equally well packed. Varying little in depth, the turnaround layers were 0.8' deep at their deepest point. At all points under the turnaround was a layer of very dark brown, disturbed humic soil which apparently represents the mechanical grading for the paved surface. This layer contained artifacts from the last three centuries, as well as brick rubble, mortar fragments, and oyster shell. It varied in thickness from 0.1 to 0.4 feet.

The construction of the turnaround caused significant damage to at least one area of the site. Photographs and a drawing made during the 1955 cross-trenching exercise indicate the presence of a hearth or some sort of brick feature overlapping the south wall of Structure A (see Fig. 15). The small feature was probably associated with adjacent and later Structure B, but the function of the feature will remain a mystery.

Two telephone cables traversed the site from the northwestern corner near Structure A to the southern extent of excavation. The northern cable predated the 1955 cross-trenching, while the southern cable post-dated this work. Both lines apparently serviced the Miller House, which stood on the site until 1972. The cable trenches intruded the plowzone layer below topsoil, planted at a depth of about one foot. Both cables lay across the hearth of Structure A, but left the feature unscathed. The trenches for the telephone lines were not visible in the topsoil layer at the southern end of the site, indicating a very recent deposition of this layer.

Two lightning rod cables also intruded the topsoil layer on the northern end of the site. One of the two large pecan trees on the site had been - damaged by lightning in the 1960s (Neil Black, personal communication), so one may assume the grounded cables were attached some time after that occurrence. According to Klingelhofer (1978), one of the cables was moved north during the 1978 excavations in the area. Neither of the cables were evident in the photographs from the 1955 cross-trenching. Fortunately, the trenches for the cables were only dug to a depth of several inches and were no threat to any features below topsoil. The trench 57 features were excavated separately from the surrounding topsoil, but were given the provenience of the topsoil layer through which they cut.

Another, deeper trench skirted the western edge of the excavation area. This trench is much deeper (c. 2.0') and contained an in-service electric cable. Though the trench was excavated, the cable was covered and avoided throughout the archaeological process.

Along the northern edge of the excavated area, a trench-like disturbance running east to west was uncovered. The boundary delineation of the feature was distinct, but wavy. The fill was a grey sandy humus approximately 0.2 feet in depth. Photographs from the 1950s indicate a hedge in this area between the Miller and Donnegan houses. Artifacts recovered from the trench were of twentieth-century origin. The only other significant modern intrusion in the 28G excavation area was an electric service line which supplied power to the Miller House. The line was no longer "live" and was therefore removed when excavated. The line was of the old wrapped-lead variety about one inch in diameter. The trench was apparently hand-dug, approximately 1.5' deep and about a foot wide. It ran from the western edge of excavation between Structures B and E eastward through Planting Bed 3, into the southeastern extent of excavation just south of Planting Bed 2.

58

IV.H. PLOWZONE

PHOTO 16. - Post Hole Through East Wall, Structure E

This very dark greyish brown mixture of sand and silt [Munsell: 10YR3/2] was the first major layer uncovered in the 28G area that was formed in situ. The layer contained brick crumbs, mortar bits, and small artifact fragments. The artifacts were not found resting on a flat surface, but were oriented randomly throughout the stratum. Plowzone consistently covered Structure A, Structure B, parts of Structures C and E, the two walkways and occupation layers with approximately 0.7 feet of soil, except in those areas disturbed during or after the 1955 trenching. Near the 90 feet east grid line, this very dark greyish brown soil thinned to 0.5 feet, and even less further to the east. The northern extent of the plowzone was not encountered during the 1982-85 excavations and it is likely that the several acres of land lying to the north of the excavation area were also under cultivation in the nineteenth century. The western extent of the plowzone was apparently North England Street. The eastern extent was not found, though it probably did not extend into the ravine where the Colonial Parkway now exists. The southern boundary of plowzone was located during the excavation of Structure E. An east-west fence line, discussed in more detail elsewhere in this report, was uncovered along the 208 North grid line. One of the postholes for the fence was dug though the east wall of Structure E (Photo 16, above). Plowzone was found north of this fence line but not south of it.

Within the plowed area, an "island" of unplowed ground was found around and between the two large pecan trees on the back lot. The base of the easternmost pecan tree was encircled by a pile of brickbats. This may indicate that the 59 gardener sensibly did not plow near the trees and piled the brick fragments struck by his plow around the base of the tree to avoid hitting them again in subsequent plowing. The brick paving at the northern extent of Walkway # 2 lies between the two trees and was not disturbed by the plowing. Further south, the walkway showed signs of scars left by the plowshares.

Plowzone was excavated in 2.5 foot square sub-units for greater detail in artifact distributions. Distributions of several ceramic types revealed the following general patterns. English delft and white salt-glazed stoneware were most prevalent in the area over Planting Beds 1, 2, and 3, and in the Structure A area. Distributions of creamware and pearlware showed a concentration around the Structure C and D area not exhibited by the delft and white salt-glazed distributions. This may be illustrative of the shift in the use of the back lot after the re-orientation associated with the building of the mid-section onto the main house.

An extremely high concentration of beverage bottle glass was revealed over Planting Beds 1 and 2, indicating that at least some of the paving or fill in the beds was affected by the plow.

Plowzone did not extend into the 28H area.

60

IV.J. BROWN SANDY LOAM 1 AND 2

The term "sheet refuse" has been used in recent years (Moir 1982) to describe the layers of soil and accompanying artifacts that become deposited and dispersed over a large area during the course of site occupation. This layer of refuse is of two types. The broader definition, used by Deetz (1977) and Moir (1982), describes the conscious or "quick time" disposal of trash from windows and doorways into the surrounding yard. The edible portion of this refuse would be consumed by domestic animals while the non-edible artifacts would be subsequently dispersed by human and animal activity. "Sheet refuse" in reference to the Brown Sandy Loam layers) which developed in the 28G area at Peyton Randolph refers to more unintentional accumulation of soil and domestic debris building up over more than a century. This type of "sheet refuse," described by Moir (1982), is decidedly different from that discussed by Deetz. This unintentional dispersal of trash would probably extend over a broader area and contain a wider temporal range of artifacts. The two types of "sheet refuse," however, will have similar distributional characteristics near dwellings and service buildings and may be similar in sherd size, both having been subjected to trampling.

Deetz (1977) found that the intentional discarding of debris out of doors and windows was discontinued, at least in the New England region, by the mid-eighteenth century. This "pitch-it-out-the-window" approach to trash disposal gave way to a more systematic, sanitary, Georgian method of ridding the home of unwanted debris—the trash pit. There was, of course, a concurrent and long-standing alternative method, the unintentional disposal of trash represented at Peyton Randolph. No trash pits or privies were found during the excavations, save the infilling of Structure A's foundations after the removal of the framing which did occur around mid-century. There appear, however, to be some spatial dimensions to refuse disposal. The Brown Sandy Loam "sheet refuse" (BSL 1 and 2) was found only in the 28G area, nearly 150 feet north of the main structure. The build-up began in the early eighteenth century and continued until the area was plowed in the mid-nineteenth. The 28H area, on the other hand, does not contain a "sheet refuse" layer; domestic debris was confined to features and externally-originating fill. The area was apparently kept clean.

Brown Sandy Loam 1 (BSL 1) was that part of the "sheet refuse" layer that built up during the late eighteenth and early nineteenth centuries. The plowzone layer, directly above BSL 1, was probably comprised of the top several inches of the layer, along with debris which continued to be deposited while the area was plowed. BSL 1 covered the entire 28G area west of the eastern north-south fence line dividing lots 237 and 207. It was not found on lot 237. Plowzone was directly over subsoil in this area.

The BSL 1 layer between the eastern fence line and Walkway # 2 was the subject of a spatial recording experiment carried out in the spring of 1983. The object of the experiment was to determine the most efficient size excavation unit, one which would render acceptable and accurate artifact distributions, and yet could be excavated quickly and easily. The area involved was piece-plotted,² the most 61 exact recovery technique possible although extremely time-consuming. The piece-plotted data was then entered into a mainframe computer at the College of William and Mary where, using SASPLOT, several graphic representations of the distribution of various artifacts were produced (Figs. 18-21). Figure 18 shows the piece-plotted data as collected, with each artifact receiving a mark. Figure 19 shows the same data indicating distribution by 2.5 foot square sub-units, while Figures 20 and 21 show the same data distributed by five-foot squares and ten-foot squares, respectively. As expected, the data became less informative as the excavation unit size increased. The 2.5 foot square sub-unit seems to be the most efficient excavation unit which would render acceptable distributional results. Some information was lost, however. For example, the distribution of nails along the eastern edge of Structure D is slightly misrepresented when combined into sub-units rather than piece-plotted (see Figs. 18 and 19). A 2.5 by 2.5 foot sub-unit can be, however, excavated and screened quickly by one excavator. Notes taken at the time of the experiment indicate that while it required one person approximately a day and a half to piece-plot a sub-unit, the same excavator could shovel and screen two or three similar sub-units in one day. The loss of detailed information was not acute enough to adopt the piece-plotting approach in other areas.

Brown Sandy Loam 2 was that part of the sheet refuse layer which began deposition with the earliest occupation of the site and continued until the third or fourth quarter of the eighteenth century with the construction of Walkway # 2. Similar to BSL 1, BSL 2 existed only west of the north-south fence line separating lots 207 and 237. The layer was also found buried beneath the redeposited subsoil bases to both walkways in the 28G area. Since the experiment described above had been conducted earlier, BSL 2 was excavated only in 2.5 foot square sub-units.

It is not known how far either of the sheet refuse layers extended north of the excavated area. The only archaeology done beyond the 1980s excavation boundary was the 1950s cross-trenching, which was primarily concerned with locating brick foundations and the site of William Robertson's windmill. Although stratigraphy was recognized at the time, it was often not recorded, reported, or analyzed. The exact southern edge of the "sheet refuse" is not known either. In the thirty feet of unexcavated area between the edges of excavation in 28G and 28H, the Brown Sandy Loams cease to exist. It appears that the part of the yard between Structures F and G may have been the mental, and possibly the physical, demarcation of that which was perceived as the "back lot" as opposed to the "back yard." The "back yard" appears to have been that area directly behind the main house that was kept clean. It was also the area in which most, if not all, of the food preparation took place. Contained within this "back yard" was the kitchen, smokehouse, dairy, and well in addition to living quarters for the servants. The "back lot" buildings, however, appear to be principally storage facilities with the exception of Structure A, a dwelling with its own occupation layers.

Distributions of artifacts within BSL 1 and 2 are illustrated in Figures 22 through 44. The distributional maps were derived from data collected in 2.5 foot square sub-units using Statistical Analysis System Graphics (SASGRAPHS) installed on the mainframe computer at the College of William and Mary Computer Center.

62

GRAPH - Figure 18.
Artifact Distribution Based on Piece-Plotting

GRAPH - Figure 19.
Artifact Distribution Based on 2.5 x 2.5 Foot Sub-units

GRAPH - Figure 20.
Artifact Distribution Based on 5 x 5 Foot Units

GRAPH - Figure 21.
Artifact Distribution Based on 10 x 10 Foot Units

89

IV.K. WALKWAY # 1

PHOTO 17. - Walkway #1, 28G Area

Walkway # 1 (Photo 17, above) ran about sixty-five feet north-south from the southern limit of excavation in 28G along the east sides of Structures E and A. At the northeastern corner of Structure A the walkway changed to an east-west course. This section of the walkway ran from the edge of the 1978 excavations along the north side of Structure A to the west side of Structure C, about thirty-seven feet. Walkway # 1 was composed of two main layers, an upper layer consisting of marl and crushed oyster shell, and a lower redeposited subsoil base. The extent of the walkway to the west is unclear, though remnants could be seen in the northern portion of excavation in the 1978 excavation area. These remnants were lost as the land grades towards North England Street. The southern side of the east-west section of the walkway was obliterated by the cross-trenching of the 1950s. Walkway # 1 was probably constructed as a service "road" connecting Structures C, E, and F to North England Street. Associated with Walkway # 1 was a possible earlier walk connecting Structures A and D.

90

Only a small amount of the marl layer to Walkway # 1 remained intact (roughly a four by four and one-half foot patch), the rest having been destroyed by the plowing of the area in the nineteenth century (see Figure 46). This patch of marl may indicate there was an entire marl paving over at least the north-south section of the walkway. The surface of the redeposited subsoil base contained marl which had evidently been pushed down into the clay base during the use of the walkway. The remnants of the marl and shell surface measured from 0.075 to 0.1 feet in depth. Stratigraphically this marl lens lay underneath the plowzone of the mid-nineteenth century and above both the redeposited subsoil base to the walkway and the cultural layer designated "Brown Sandy Loam 1" (BSL 1). This relationship to BSL 1 is tenuous. The marl only sealed BSL 1 on the very edge of the walkway. It appears that the marl was deposited at about the same time as that BSL 1 began building up. The marl was placed on the subsoil base to keep carts and wagons from "bogging down" when the clay became wet and muddy from rain.

Figure 46 - Walkway #1, 28G Area

The second layer of Walkway # 1, the subsoil base, ranged in depth from approximately 0.5 feet in the north to about 1.2 feet in the south. Due to a disturbed and highly irregular eastern edge, the width of the walkway varied from seven to fifteen feet. This subsoil base also ties into the marl work area around Structure E. The width and depth of the subsoil base indicates plans for heavier traffic than simply a foot path. Stratigraphically, the subsoil base lay just under BSL 1, marl and plowzone. The earliest sheet refuse cultural layer, BSL 2, was cut by, and lay under, the subsoil base. On the east side of Structure E, where the base is deepest, it also intruded Planting Bed 3.

91

Because of its stratigraphic location and its association with Structures E and F, the walkway appears to have been used in some form from the late 1750s to the mid-nineteenth century. The late 1750s are also the probable construction period for Structures E and F. The area appears to have been devoid of outbuildings and was subsequently plowed by the third quarter of the last century. The southern extent of the walkway appears to have been the area around Structure F, though that area was not re-excavated in the 1980s, as it was not encountered in the 28H area of the lot.

Walkway # 1 has two other important features. The first is its association with the marl work area around Structure E. It is in this area that the redeposited clay was deepest. The clay was underneath the several marl, shell, and mortar layers which together make up the work area, and the walkway merged with the work area on the eastern side of Structure E.

Secondly, Walkway # 1 adjoined Walkway # 2 on the western side of Structure C, thus servicing this building and offering access to the buildings serviced by Walkway # 2. Walkway # 1 had the same stratigraphic position as Walkway # 2 with one exception. Walkway # 2 had two layers which Walkway # 1 did not share: a brick paving near Structure C and a brick chip lens under the redeposited base.

92

IV.L. WALKWAY # 2

PHOTO 18. - Walkway #2, 28G Area

Of the two major walkways located in the 28G area, Walkway # 2 (Photo 18, above) was the easternmost. The walkway ran north-south through units 230-270N 50-70E and, like Walkway # 1, was composed of several levels. Stratigraphically from top to bottom these levels included a layer of brick, marl, redeposited subsoil, and finally brick chips. Walkway # 2 serviced Structure C and probably Structure D as well, in an earlier form. Walkways # 1 and # 2 connected near the northwestern corner of Structure C, forming a loop. Unfortunately, the area of this loop was highly disturbed by the growth of two large pecan trees, preventing complete excavation. The southern limits of the Walkway # 2 were truncated by the Miller House, making excavation in that area impossible.

Adjacent to the west wall of Structure C, between the two large pecan trees, an area of brick paving was uncovered (Fig. 47). The paving had been placed over the marl level of the walkway, but survived only between the two trees. It is possible the paving had once extended further south, but had been removed when the area became a plowed field in the mid-nineteenth century. The remaining section measured 9.5' east-west by 9.0' north-south. Since no evidence of brick paving remained south of the pecan trees, it is possible that the paving represented a work area adjacent to Structure C, not existing over the walkway to the south. Ceramics recovered from the brick paving include English delft, English stoneware, Buckley, Jackfield, white salt-glazed stoneware, and pearlware. Other artifacts found consisted of nails, window and beverage glass, and iron fragments. No artifacts leading to a functional interpretation of the building were located.

The next level of the walkway was only slightly earlier than the brick paving and was composed primarily of marl. This marl level, unlike the brick paving, was 93 present in most of the units containing the walkway. Stratigraphically, the marl layer was superior to, and later than, the Brown Sandy Loam 2 layer, and contemporary with the BSL 1 layer (early nineteenth-century). This early nineteenth-century date for the marl layer makes almost certain a close relationship between the walkway and Structure C. Virtually devoid of artifacts, only two ceramic fragments were recovered from the marl: pieces of white salt-glazed stoneware and pearlware. It appears that the marl layer had been placed on the redeposited subsoil base in order to even the grade for the placement of the brick paving. This assumption is reinforced by the lack of artifactual remains.

Figure 47 - Walkway # 2, 28G Area

A rather substantial redeposited subsoil base was located beneath the marl. The depth of this redeposited base was about 0.5 feet in the northern area, becoming slightly thinner towards the south. There were no artifacts found in this layer of the walkway. The layer was composed of a redeposited yellow sandy clay subsoil [Munsell: 10YR8/8] with light grey sandy clay and white silty clay mottlings [Munsell: 10YR6/1 and 8/2, respectively]. Though primarily sealed by the marl layer, the redeposited subsoil base was also sealed by BSL 1 and plowzone in several areas. The base was sealed by BSL 1 along the edges for a width of two to four 94 inches and was sealed by the plowzone in locations where the marl was not present, mostly towards the southern end. The marl layer here may have been plowed away, thus becoming incorporated in the plowzone.

The earliest layer in Walkway # 2 was composed of brick chips. It was neither as substantial nor as extensive as the other layers of the walkway. The layer was comprised of a yellowish brown and brown sandy loam [Munsell: IOYRS/3 and IOYR6/4], mixed with a heavy concentration of crushed brick. The layer was within Brown Sandy Loam 2 and was underlying the redeposited subsoil base. Unlike the three layers above it, the brick chip layer appears to have serviced Structure D rather than C, as it was associated with BSL 2 rather than the later BSL 1. The brick chip path was approximately three feet in width at its northern boundary, widening to nearly eight feet near the southern edge of excavation in unit 230N 60E. The length of the path was roughly twenty feet and the overall depth about one inch. The were no remains of the brick chip part of Walkway # 2 south of the 230N unit, so it is difficult to determine where the pathway led south of Structure D. The artifacts recovered from the brick chip layer were of little use in determining the temporal span of the path. English delft was the only ceramic type recovered, although undatable fragments of glass, bottle glass, shell and nails were also found.

In summary, the relationship between the earliest part of the walkway, the brick chip layer, and Brown Sandy Loam 2 indicates that the path was constructed during the middle phase of the BSL 2 build-up, in the second quarter of the eighteenth century. Falling from use after the destruction of Structure D in the mid-eighteenth century, it was replaced by a more substantial walkway at about the time Structure C was built. This new walkway was constructed using a redeposited subsoil base, forming the demarcation point between the two brown sandy loam sheet refuse layers. A small amount of the later BSL 1 (about 0.25 inches) had been deposited over the edges of the base, so it appears that the subsoil base was in use alone for a short while before the marl was laid down. The brick paving was laid over the marl and subsoil base near Structure C, possibly to provide a dry work space for whatever activities were taking place in the area. Walkway # 2 remained in use until the destruction of Structure C in the mid-nineteenth century when all outbuildings in the 28G area were razed for plowing.

95

IV.M. FENCE LINES

PHOTO 19. - Eastern North-South Fence Line

102

IV.P. PLANTING BEDS

The most dramatic and potentially the most informative finds made during the 1982-1985 excavation effort were the garden features uncovered in the 28G area. Four vegetable gardens or planting beds were found and excavated. Each bed was different from the other in several respects, which presented a number of excavation problems. The information to be gathered, the physical attributes of each bed, weather, and time allowed for excavation all affected the recovery technique. Upon the discovery of the first two planting beds it became evident that a unique opportunity to gather new information from previously undiscovered and relatively intact, discrete early eighteenth-century features at Peyton Randolph had presented itself. The standard archaeological questions, i.e., when, what, who, how, and why, were naturally the first to be addressed. Other questions arose regarding how the beds related to the owners of the main house, whether they were formal gardens or vegetable patches, the kind of fertilizer that was used, if any, and the non-dietary use of faunal material.

Photo 21 - Differential Drying Over Planting Beds 1 and 2

103

111

IV.Q. MIDDLE PLANTATION FEATURES

Figure 54 - Middle Plantation Era Boundary Ditch with Associated Post Holes

Several features located in the back lot of the Peyton Randolph house have been attributed to the Middle Plantation Period. Middle Plantation was established after the Indian Massacre of 1622, along a ridge between the York and James Rivers. The site was less swampy than the low-lying capital at Jamestown, and though accessible by navigable creeks to the two rivers, it was not located directly on either. In 1633/34 a palisade was erected in order to secure the area from Middle Plantation to Hampton Roads from Indian attacks. Also providing grazing and foraging land for cattle, the palisade passed very near the present Peyton Randolph site. After the second Indian Massacre of 1644, a new palisade was built just northeast of the original one but still within Middle Plantation.

The College of William and Mary was established at Middle Plantation in February of 1693. At that time the settlement consisted of Bruton Parish Church, Burwell's Ordinary, several stores, two mills, a smith, a grammar school and several residences. When Virginia's capital was moved from Jamestown to Middle Plantation in 1699, streets were laid out in a north-south, east-west pattern and the city was re-named Williamsburg, after the reigning monarch.

The only evidence of the Middle Plantation Era found at the Peyton Randolph Site was a boundary ditch (Fig. 54) and several post holes which appear to be in association with this ditch. The ditch was first discovered during the 1978 archaeological investigation of the west half of Structure A (see Appendix 4). According to Klingelhofer, author of the 1978 Interim Report, the ditch was approximately four feet wide and one foot, two inches deep, containing only a few nail fragments. The only section of the ditch excavated at that time was on the exterior of Structure A, though it extended under the southwestern corner of the 112 structure. The part of the ditch in the interior of Structure A was sealed by a silty dark brown loam believed to be buried topsoil from the time the building was constructed. Though the ditch had probably been filled in somewhat when Structure A was built, it had not fully compacted and was quite evident to the builders. A line of rowlock brick, of the same color and consistency of those used to construct the piers in Structure A, was discovered placed across the ditch supporting the southwestern pier (see Photo 6).

From the edge of the excavated area near Structure A, the ditch continued in a direction approximately 23° east of south, running under nearby Structure B, Planting Bed 4, and the east wall of Structure E (see Fig. 54, previous page). The ditch had fully compacted by the time Structures B and E were built as it had no effect upon their respective foundations. The ditch continued across the 28G area past the southern edge of excavation. It was located again in the 28B area near the southwestern corner of Structure R. The full extent of the boundary ditch is unknown, though it is probably many times more extensive than the 150 foot section which passed through the Peyton Randolph site.

The sections of the boundary ditch that were excavated in 1985 contained brown to dark brown sandy loam [Munsell: 10YR4/3] with fine-grained sand. They were sealed by Brown Sandy Loam 2 and Planting Bed 4, intruding sterile subsoil. The only artifacts recovered from any part of the feature were a fragment of bottle glass and several highly oxidized nails.

This boundary ditch has been ascribed to the seventeenth-century Middle Plantation Period for several reasons. It was located at the lowest stratigraphic level, intruding only subsoil. It was not oriented on the Williamsburg Period north-south, east-west axes as were all other features discovered at Peyton Randolph. Finally, it contained very few artifacts considering its extent, which indicates that little cultural activity had taken place in the area when the ditch was constructed or when it was being in-filled.

Three post holes, found near the southern edge of the excavation area in 28G, appear to be in association with the boundary ditch. All three were located at the same stratigraphic level as the ditch. Unfortunately, they do not seem to form any pattern which permits their interpretation, nor did they- contain any significant artifacts. They are described as follows:

V.A. MODERN INTRUSIONS AND TOPSOIL

The first modern intrusions encountered in the 28H area were in the section cross-trenched in 1955. This was in what was at the time the Skillman House front yard. Part of the disturbance was apparently related to the construction and destruction of the Skillman House. Fortunately this activity did little but mix the topsoil layers, leaving the rest of the stratigraphy basically intact. A modern clay and gravel walkway associated with the house was located just under topsoil at the northern end of the excavation. A lead-encased telephone (?) line, which traversed the northern three excavation units from west to east leading to the Skillman House, was also relocated. This line is shown in a 1955 photograph taken during the previous cross-trenching (Photo 26), and did little damage to the stratigraphic record.

Also associated with the Skillman House was the city water line which was, according to the Skillmans (Andrea Foster, personal communication), put in when the house was built in 1924. It ran from the western edge of excavation across three units to the eastern edge only inches south of the well. The pipe trench was 116 Photo 26 - Telephone Line in 28H Area, 1955 Photo 27 - Late Nineteenth-Century Windmill at Peyton Randolph Site 117 approximately a foot wide and 18 inches deep. Another water line led from the eastern edge of the well to the eastern edge of excavation. This pipe was associated with a windmill which stood over the well from the late nineteenth century until 1920. This windmill appears in an archival photograph (Photo 27, previous page) which was taken around the turn of this century. The building serviced by the windmill must be the large structure to the east, but the temporal and functional aspects of the building are unknown. It was located partly where the Skillman House once stood and partly where the parking lot for the life tenant now exists, and was therefore not excavated during any of the archaeological investigations of the back yard. The windmill had four supports placed around the well. Each was dug into the ground some four feet. Only one of the support features was fully excavated, the one on the northeastern edge of the well. Each dimension was four feet. The feature contained many early twentieth-century artifacts.

118

V.B. MARL WALKWAY # 1

Figure 55 - Marl Walkway #1, 28H Area

The archaeological remains of the feature designated Marl Walkway # 1 (Fig. 55, above) were sketchy at best. Recent intrusions such as the Skillman House and associated utilities, hedges, etc., as well as the 1930s and 1950s archaeological work, significantly obscured the original configuration of the feature. The artifacts recovered from the marl walk indicate a TPQ of 1820, though the diagnostic indicator—whiteware—was not recovered in significant amounts (only four sherds out of a total of 1,516 artifacts). Stratigraphically, the feature sealed the c. 1795 Refuse Layer discussed later in this report. The walkway was sealed in turn by Topsoil 2, with a TPQ of 1907. Unlike Walkway # 2, the most prevalent ceramic type recovered was creamware, with pearlware representing a significant second most prevalent type. Delft and white salt-glazed stoneware were ranked forth and fifth in frequency of occurrence rather than first and second as in Walkway # 2. Walkway # 1 probably performed a similar function as that of Walkway # 2, though it was not as extensive and could not be regarded as a yard spread, but rather a simple walkway from the kitchen (Structure L) to the well, Structure G, and Structure H. Though very disturbed in the area near the porch appendage to Structure L, the walk appears to be related to the porch, unlike the older marl spread which was in place when the appendage was constructed. This marl walkway is probably attributable to the Peachy occupation of the house (1800-1860).

119

V.C. MARL WALKWAY # 2

Figure 56 - Marl Walkway #2, 28H Area

Marl Walkway # 2, or the Marl Yard Spread as it was also called, was found below the Refuse Layer described in chapter V.D. This rather well packed layer of marl (Fig. 56, above) was located east of Structure L, between it and Structures P, Q, and R. The purpose of the c. 0.2-0.3 foot thick paving was apparently to provide a stable surface to the area around the main kitchen. The remnants of a brick walk were imbedded in the yard paving leading to the well area. The marl extended northward to the south wall of Structure H, probably abutting the building before this section was removed during the 1930s excavations.

Artifacts recovered from the marl rendered a tentative TPQ of 1820, derived from two small whiteware sherds. If these are disregarded as mechanically displaced, a TPQ of 1779 is indicated. The most frequently encountered ceramic types recovered from the Marl Spread were delft, white salt-glazed stoneware, Chinese porcelain, creamware, and Fulham-type stoneware. Overall, the most numerous artifact types were bottle glass, bone, window glass, and wrought nails (see below for artifact count and percentages). Artifact types did not tend to indicate functional differences in the area of the marl, but rather represented the usual assemblage of domestic debris. Artifact density was also rather low, with a total of only 617 artifacts having been recovered from the layer, which was located in thirteen separate units. The low density is to be expected from a layer which was probably "mined" from a nearby outcropping, spread over the area in a short period, and most likely swept clean of debris on a regular basis. The presence of a significant amount of pearlware (eleven sherds) indicates a deposition after 1779, and probably after the house was sold at auction in 1783. The spread may have been deposited as late as the first decade of the nineteenth century, but this possibility is unlikely since the Refuse Layer covering it was probably deposited at this time, and it would be illogical to pave an area and then almost immediately deposit fill on top. The Marl Yard Spread is therefore attributed to the Hornsby occupation of the house (1783- 1800).

120

ARTIFACT TYPES IN WALKWAY # 2

Type	Count	Percent
Bottle glass	174	28.20%
Bone	159	25.77%
Window glass	58	9.4%
Wrought nails	44	7.13%
English delft	24	3.89%
White salt-glazed stoneware	24	3.89%
Chinese porcelain	18	2.92%
Creamware	16	2.59%
Pipe stems	13	2.11%
Glass	12	1.94%
Fulham stoneware	11	1.78%
Pearlware	11	1.78%
Leaded glass	11	1.78%
Copper alloy	11	1.78%
Shell	8	1.30%
American stoneware	4	0.65%
Redware	3	0.49%
Slipware	3	0.49%
Yorktown	2	0.32%
Whiteware	2	0.32%
Iron objects	2	0.32%
Pewter/lead	2	0.32%
Colono-ware	1	0.16%
Westerwald	1	0.16%
Porcellaneous	1	0.16%
Pipe bowls	1	0.16%
Machined nails	1	0.16%
Whieldon ware	0	0.00%
Refined earthenware	0	0.00%
Bennington stoneware	0	0.00%
TOTAL	617	100.0%

121

V.D. REFUSE LAYER

Figure 57 - Refuse Layer, 28H Area

Located principally in units 110N 40E, 120N 40E, 110N SOE, and 120N SOE, between Topsoil 2 and the Marl Yard Spread, was a layer of secondary fill approximately 0.4 feet in thickness (Fig. 57, above). The soil was a grey brown sandy loam [Munsell: 10YR5/2] containing inclusions of shell, bone, fossils, and charcoal as well as an extremely high concentration of eighteenth- and nineteenth-century artifacts. Cross-mending of the major ceramic types with other layers and features from the site produced no glue-fit mends, indicating strongly that the fill had its origins elsewhere, having been brought to the Peyton Randolph back yard in order to fill in a low area of the site. The exact eastern edge of the layer was not investigated, though it was not encountered in the area around Structures P, Q, and R east of the marl walkway. The layer extended to the west only to within about two feet of the eastern wall of Structure L, and, though extant northeast of the Structure J hearth, was very disturbed. It was not found within the interior of Structure H, but probably abutted its south foundation wall before the 1930s excavations interrupted the stratigraphic sequence associated with the structure.

122

The density of the artifact concentration can be illustrated grossly in sheer numbers. For example, in unit 120N 40E, the refuse layer covered approximately 80% of the square and yielded 9,649 artifacts—1,575 from one 2.5 by 2.5 foot sub-unit alone! Ceramic types included:

• English delft, plain white
• English delft, blue and white
• English delft, manganese
• North Devon gravel coarseware
• Colono-ware
• Staffordshire mottled coarseware
• redware
• Yorktown coarseware
• redware, unglazed
• slipware
• Staffordshire slipware or "dotware"
• Whieldon/Wedgwood cauliflower
• creamware, royal
• creamware, feather-edged
• creamware, molded
• refined earthenware, green glaze
• pearlware
• pearlware, transfer printed
• • monochrome
• Westerwald stoneware
• Nottingham stoneware
• Fulham stoneware
• white salt-glazed, bead reel
• white salt-glazed, hand painted overglaze
• white salt-glazed, debased scratch-blue
• white salt-glazed, plain
• Jackfield ware
• Chinese porcelain, Batavian
• Chinese porcelain,
• • bianco supra bianco
• Chinese porcelain, plain
• Chinese porcelain, hand painted
• • underglaze
• Chinese porcelain, hand painted
• • overglaze
• Chinese porcelain, hand painted
• • under and overglaze
• bone china, porcellaneous

In addition to a vast variety of eighteenth- and nineteenth-century ceramics were glass vessels, nails (both wrought and machine-made), clay pipe bowls and stems, worked stone, brick, mortar, bone, shell, carbonized wood, coal, iron, brass, pewter, and lead. A sample analysis of finds from two units, 28H/299 and 28H/300, represented by 11,125 artifacts, indicated the majority were bottle glass fragments (29.65%), bone (20.78%), and window glass fragments (13.12%). Creamware was the most often recovered ceramic (5.12%).

The material had probably not been seriously trampled. The average sherd diameter was 1.006 inches for all ceramics, somewhat larger than those recovered from the Marl Spread that the layer covered. This fact tends to suggest that the layer was probably deposited as fill, not as a "living" surface.

A complete breakdown of thirty-six artifact types from the two units, in order of frequency, is as follows:

123

ARTIFACT TYPES IN REFUSE LAYER

TYPE	COUNT	PERCENT
Bottle glass	3299	29.65%
Bone	1549	20.78%
Window glass	1460	13.12%
Wrought nails	849	7.63%
Creamware	570	5.12%
English delft	345	3.10%
Chinese porcelain	333	2.99%
Oyster shell	329	2.96%
Leaded glass	236	2.12%
White salt-glaze stoneware	190	1.71%
Coal	184	1.65%
Glass	131	1.18%
Ferrous objects	128	1.15%
Pipe stems	96	0.86%
Slipware	93	0.84%
Copper alloy objects	89	0.80%
Fulham stoneware	87	0.78%
Redware	74	0.67%
Yorktown-type earthenware	58	0.52%
Pearlware	55	0.49%
Pipe bowls	41	0.37%
Pharmaceutical	36	0.32%
Westerwald stoneware	20	0.18%
Pewter objects	17	0.15%
Colono-ware	16	0.14%
Machined nails	16	0.14%
Whieldon ware	14	0.13%
Rouen faience	10	0.09%
North Devon	7	0.06%
Nottingham stoneware	7	0.06%
Staffordshire	6	0.06%
Worked stone	5	0.04%
Refined earthenware	4	0.04%
Porcellaneous	4	0.04%
Buckley	2	0.02%
Jackfield	1	0.01%
Whiteware	1	0.01%
TOTAL	11,125	100.00%

124

Mending within the entire layer produced fragments of 101 unique vessels:

Creamware (30 vessels):

• 2 Royal pattern plates
• 6 feather-edge plates
• 3 feather-edge platters
• 5 tea bowls
• 3 saucers
• 2 bowls
• 1 chamber pot
• 1 plain platter
• 1 basket
• 1 sauce boat
• 1 tea pot lid
• 1 tankard
• 1 large saucer
• 1 milk jug
• 1 small bowl

Delft (16 vessels):

• 4 plain chamber pots
• 3 plain bowls
• 2 plain plates
• 2 blue and white plates
• 1 manganese and white
• 2 blue and white punch bowls
• 1 can
• 1 polychrome bowl

Chinese porcelain (20 vessels):

• 8 blue under glazed plates
• 2 polychrome plates
• 2 plain tea bowls
• 2 blue under glazed tea bowls
• 1 platter
• 1 pitcher
• 1 bowl
• 1 lidded vessel
• 1 can
• 1 saucer

White salt-glazed stoneware (13 vessels):

• 2 bowls
• 2 tea pots
• 1 chamber pot
• 1 sweetmeat dish
• 1 bowl
• 1 bead and reel plate
• 1 basket weave plate
• 1 tankard
• 1 tea bowl
• 1 custard cup
• 1 pitcher

Slipware (5 vessels):

• 1 small bowl
• 1 charger
• 1 bulbous mug
• 1 dotware mug
• 1 pie pan
• 1 tankard

Fulham stoneware (3 vessels):

• 1 jug
• 1 storage vessel
• 1 bottle

Redware (4 vessels):

• 1 bowl
• 1 pitcher
• 1 chamber pot
• 1 flower pot

Pearlware (6 vessels):

• 1 saucer
• 3 tankards
• 1 tea bowl
• 1 plate

Colono-ware (1 vessel):

• 1 bowl

Whieldon (2 vessels):

• 1 small bowl
• 1 plate

Rouen faience (1 vessel):

• 1 storage jar

125

Of the vessel types represented, 83.81% are tablewares. The remaining artifacts recovered from the refuse layer are principally domestic in function, including a large number of faunal remains. The amount and variety of expensive ceramic types indicates that this secondary refuse deposit probably came from an affluent domestic site, of which there are many in the area.

126

127

V.E. "CLEAN" BROWN LAYER

The lowest stratigraphic layer, and the only layer pervasive over the entire 28H excavation area, was a light brownish grey [Munsell: 10YR6/2], medium-grained sandy loam. Virtually devoid of artifactual remains, given its extent and thickness (c. 0.2'-0.4'), it became known as the "Clean" Brown Layer. The layer covered eighteen units in the 28H area and probably corresponds with Brown Sandy Loam 2 in the 28G area, though no direct physical connection was made. Both layers seal subsoil.

A variety of different layers and features were found to seal the Clean Brown Layer. The Marl Spread, also called Marl Walkway # 2, sealed the layer in unit 130N 40E; it was covered by the probable floor and the hearth in Structure K. Within the interior of Structure L, a fill layer of yellow mottled clay, probably dating to the mid-nineteenth century renovation of the structure, covered the layer. The Clean Brown Layer was intruded by the same Marl Spread in the 80N 40E and 80N 50E units, and by the builder's trench for Structure L. The layer was found within the interiors of all structures in the 28H area, usually below any recognizable floor. The north wall of the "hall" section of Structure J was found within the Clean Brown Layer, whereas the eastern and southern walls of the same "hall" rested on the layer. Because of the trenching around the structure in the 1930s, it was impossible to determine whether the layer continued to be deposited around the walls resting upon it. Structure J has been determined architecturally and archaeologically to be one of the first buildings erected on the lot, probably at approximately the same time as the main house (c. 1718). Since part of the structure was built on the already existing Clean Brown Layer, the layer must relate to the early eighteenth-century ground surface. Protected from subsequent deposition by Structure J and later Structure L, the Clean Brown Layer remained covered only by the interior of a building until the building was renovated in the mid-nineteenth century. At that time approximately a half-foot of clay was spread over the interior of what was then Structure L. The purpose of the clay is unknown. It may have functioned to reduce the moisture retention under the structure.

To the east of the Structure J/L complex, the Clean Brown Layer became covered with several deposits described above. As a ground surface on the exterior of the buildings, however, the Clean Brown Layer accumulated some cultural debris.

Ceramic remains within the Clean Brown Layer were sparse. Those recovered from the layer appear to have been trampled. The average size of the ceramics represented in the layer was the smallest of any of the areas sampled (0.814").¹ There are some indications as well that the deposition of artifacts in the Clean Brown Layer occurred in two phases. Approximately 90% of all artifacts recovered from the layer were found in units 100-130N 40E, along the east exterior of Structure J/L, and within the interior of Structure K (Fig. 58, opposite). The ceramic types represented in the area of Clean Brown within the interior of Structure K were delft and Westerwald. Those represented on the northern exterior 128 of Structure K include later ceramics such as Whieldon and faience in addition to delft, Westerwald, Chinese porcelain, and combed slipware. Though evidence is suspect given the paucity of ceramics recovered, this distribution of finds may indicate that an east entrance to the lean-to on Structure J was located near the center of the western side of where Structure K was later added. A path led from the door northward to the dairy (Structure H) and the well, and sweepings from Structure J were trodden into the path. When Structure K was erected and a clay floor laid over the Clean Brown, the exterior door from Structure J became an interior door to Structure K and another door was placed on the north wall of Structure K for access to the well and dairy. This path continued to be deposited with later ceramics during the course of Structure K's use. The Clean Brown Layer was subsequently covered by various deposits until completely sealed by the Marl Spread in the 1780s.

129

V.F. FENCE LINE

Figure 59 - Fence Line

Of the sixty-one post holes excavated in the 28H area, only nine (with possible exceptions) seem to share any kind of relationship with each other, and so constitute a macro-feature. Eight or nine of the post holes appear to reflect the presence of a north-south fence line (Fig. 59) beginning at 60N 40E and continuing to 130N 40E, erected some time in the first half of the nineteenth century. The following is a summarization of each post hole by context number.

133

V.G. STRUCTURE G - DAIRY (?)

PHOTO 28. - Structure G

The first of a series of foundations re-excavated in 1984 was the remains of a building designated Structure G (Photo 28, above). From the James Knight map, drawn after the 1955 cross-trenching, the foundations appear to be those of a dairy. Unfortunately, due to the proximity of the Skillman- House foundations just to the east, the structure was never fully excavated.

The foundations measured ten feet north-south, and probably the same east-west, though the east wall has never been uncovered. The building lay predominately in unit 150N 40E, resting slightly askew of the surrounding buildings. Interestingly enough, however, it is parallel to the existing well-curb.

English bond was attempted in the northwestern corner of the structure. This, however, gave way to rowlock and finally to casual bonding as one moves southward along the western wall. The brick bats in the foundation were, of course, re-used, but the whole brick in the foundation appear to have been new at the time of construction. The foundations have been damaged and repaired in places, and damaged but not repaired in others. The brick used in the floor paving was made from both whole and partial brick, and has undergone extensive damage over the years.

134

As is the case with the majority of the outbuildings at Peyton Randolph, Structure G had been separated from the surrounding stratigraphy by the early archaeological activities of the Foundation. Dating the construction and destruction of the building is therefore not easy. None of the artifacts recovered in or around the structure can be said to come from, or be related to, the structure in any way. Derivation of the function of the building from the artifacts is therefore impossible. The brick floor does not necessarily suggest that the building was indeed a dairy, nor does its ten by ten foot size.

The location of the building may reveal a relative date that the builder's trench (having been destroyed in the previous digging) could not. The building rested approximately within the eastern line of the eastern edge of the west wing of the house, as did Structures K and H. It is logical to assume that all of these buildings were constructed prior to the erection of the mid-section of the main house, i.e., around the mid-1760s. Whereas Structure K, part of the rather organic kitchen complex, probably did not survive the construction of the mid-section, Structures G and H apparently did. The Marl Yard Spread, also known as Marl Walkway # 2, was associated with the use of the latest kitchen (Structure L), and covered Structure K but not Structures G and H. Therefore G and H must have still been in use during the last kitchen period. Both buildings apparently survived into the early nineteenth century as the c. 1800 refuse layer brought in by either Hornsby or Peachy did not cover either structure. The last marl walkway (Marl Walkway # 1), which has a TPQ of 1820, did cover Structure G, indicating it was down by the mid-nineteenth century. Structure G was then either unnecessary or had been replaced by one of the other structures located in the 28B area (N, P, Q, or R). If Structure G had functioned originally as a dairy, it may have been replaced by Structure P. An early twentieth-century photograph (see Photo 30) shows two eighteenth-century style outbuildings in the back yard. One is apparently a smokehouse (probably Structure H), and the other is a dairy which appears to stand where the foundations of Structure P were uncovered.

135

V.H. STRUCTURE H - SMOKEHOUSE (?)

PHOTO 29. - Structure H

The foundation designated Structure H (Photo 29, above) is located approximately two feet south of Structure G in units 130N 40-50E and 140N 40-SOP. The building measures 12.4 feet north-south by c. 12.0 feet east-west (exterior). Well-fired primary-use brick, mortared with shell mortar in the English bond style, were used to construct the foundation. Compared with Structure G, the foundation remains were in excellent condition. Structure H was partially excavated during the 1930s archaeological investigation, then partially excavated again in the 1950s. The 1984 and 1985 excavations completely uncovered only the southern and western walls, leaving the northern and eastern walls only partially dug. Approximately half of the interior was excavated to subsoil. In addition to the four foundation walls, Structure H had a rather unusual appendage attached to the west side. The appendage consists of a row of brick (and one flat stone) laid on edge extending westward 2.3 feet from the southwestern corner, at which point the row brick extended northward eleven feet, where it again made a right angle, extending westward another six feet to the southeastern corner of the well-curb. The purpose of this rather strange addition to Structure H is unknown. The area between the brick row and the foundation walls was either enclosed, or had a special function because the stratigraphy inside the appendage was distinctly different from that outside to the west and from the interior of Structure H.

136

138

V.J. STRUCTURE J - DWELLING/SECOND KITCHEN

PHOTO 31. - Structure J

Structure J (Photo 31, above) is probably one of the oldest and most enigmatic buildings excavated in the 28H area. The layout and extent of the structure has been a topic of a great deal of discussion among architectural historians and archaeologists. Included in this descriptive analysis of Structure J are several foundation walls, either truncated, subsequently built upon, or robbed out. The most rudimentary part of the foundation consisted of two east-west walls eight feet apart, constructed both in casual and rowlock bond. A thirteen-foot north-south wall connected the two on the eastern side, and it also reflects the two different types of brickwork. A similar wall to the west was displaced by a walnut tree and was not excavated because of fear of damage to the tree. These four walls would enclose an eight by eleven foot room or parlour entrance. The thirteen-foot eastern wall was later truncated by the construction of the Structure L double hearth, though remnants of a south wall for Structure J exist along the north edge of the vaulted cellar. Together, these walls would indicate another room south of the parlour measuring approximately eleven by eleven feet square. North of the parlour entrance room, there appears to have been a small room measuring approximately ten feet north-south by eleven feet east-west, containing a five by 139 eight-foot interior hearth. Forty of the 110 available square feet in the room was taken up, therefore, by what appears to be a cooking hearth. Apparently, not too long after the construction of the building, a six foot wide lean-to was built on the east (back) side of the structure. Whether the lean-to ran the entire thirty-two foot length of the building initially is uncertain, but it appeared to do so in later phases of the structure's life.

Structure J grew organically over the years. The interior hearth was abandoned, and an exterior hearth and an oven was constructed on the north elevation. Later, the fourteen-foot-square room designated as Structure K was added on to the building, making an L-shaped kitchen with two hearths and approximately 775 square feet of work space. Additionally, at some time in the eighteenth century, what appears to be a forge was added to the northwestern corner of the second (exterior) chimney base to Structure J. Finally, Structure J was incorporated into Structure L, the last free-standing kitchen to serve the Peyton Randolph House. Whether the, hearth and oven at the north end of the structure remained in use with the building of the large new kitchen is unclear. All archival photographs showing the back yard of the house in the last century were taken from such an angle that the northern part of the kitchen is obscured.

Hearth 1 in the north parlour of Structure J was uncovered in the 1938 excavation. Judging from the brickwork, James Knight gave it a relative date of construction as "first period," along with the main body of the structure. The hearth appears to have been placed in the northeastern corner of the room. Unlike the corner hearth in Structure A, whose firebox was at an angle to the walls of the room, the hearth in Structure J is parallel to the north wall. According to the 1938 map resulting from the archaeology, the hearth was covered with several inches of ash above the brick paving of the firebox, and contained six inches of ash below the brick paving. Unfortunately, nothing is known about the artifacts, if any, recovered from either above or below the paving. None were found in the 1984 excavations. The only layer associated with this hearth appears to be a lens of ash partially inside and extending southward. This deposit is more likely associated with the second hearth. Found in the ash, aside from the creamware which gave it a TPQ of 1762, was a commemorative shoe buckle from Admiral Vernon's victory over the Spanish on November 22, 1739 (Figure 60).

Figure 60 - Admiral Vernon Commemorative Buckle

A small (2.6' interior diameter) oven was discovered two feet east of the later hearth. This oven was the only significant brick structural feature not included on the Knight map following the 1930s excavations. Though intruded later by a modern driveway, it appears the round oven on its five foot square base was at least partially uncovered during the 1930s excavations. Apparently, the foundations were inadvertently left off the map. The layer covering the oven had a TPQ of 140 1880, indicating a possibility that the chimney and oven stood above ground as long as the rest of Structure L.

The small, northernmost hearth or forge built onto the northwestern corner of the second hearth to Structure J measured approximately 5.5 feet east-west and 6.5 feet north-south. The brick was in good condition and appeared to be set in English bond. Ash in and around the hearth has a TPQ of 1762. Exactly what the forge was used for is not known, though small tool repair forges are not unusual, even on urban sites. No slag or melted metal was found in association with the hearth, however. Among the unusually prolific domestic items recovered from the ash layer associated with this forge were 114 nails.

141

V.K. STRUCTURE K - THIRD KITCHEN

PHOTO 32. - Structure K

Structure K (Photo 32, above) was actually an addition to Structure J, probably replacing the former Structure M kitchen some time in the late 1730s or early 1740s. Exterior dimensions are square, 17 feet by 17 feet. It contained an interior hearth along the center of the east wall, with a firebox six feet long and 3.5 feet wide. Placing the hearth in the interior caused two 2.5 foot spaces on either side of the chimney. These spaces could have been ovens, though no evidence of this possibility remained archaeologically. The structure was built from highly-fired, dark-colored brick, incorporating a large number of bats in construction. Both the north and south walls were interrupted about halfway by two post holes from a later fence line. The structure had been fully excavated, in the 1930s, thereby separating it from the surrounding stratigraphy, except in one small area. A layer of undisturbed brown loam [TPQ 1762] lay over a small portion of the southeastern corner of the structure. In addition, layers which existed both on the interior and exterior of the structure, such as the Marl Spread described earlier, had TPQ dates of 1779.

142

The hearth inside Structure K had been left fairly well intact by the earlier archaeology, and was fully excavated in the 1984 season. The first layer in the firebox consisted of highly-burned brick, under which were successive lenses of ash and sand. Artifacts recovered from the firebox had a TPQ of 1720. The fill soil and subsoil below showed signs of prolonged and intensive heat, i.e., the soil was varying shades of red and purple. The floor in the structure was probably earth, as the hearth was not elevated and signs of burning extend over a foot into the interior of the structure. A layer of yellow clay [TPQ 1720] just above subsoil was found over part of the interior. This may have been the remains of the kitchen floor. No other features associated with Structure K were located.

143

V.L. STRUCTURE L - FOURTH KITCHEN

PHOTO 33. - Structure L

"Structure L" is the designation given to the last exterior kitchen to serve the Randolph House, constructed by Peyton Randolph during the 1760s. The remains (Photo 33, above), first uncovered in the 1930s excavations, included a vaulted cellar with a bulkhead entrance, massive support piers, a double chimney base, and support piers for a covered way leading to the back door of the main house. The parts described were built onto already-existing foundations, those of Structure J, resulting in one multi-purpose building some 48 feet long (north-south) and 20 feet wide (east-west). Undoubtedly Structure J was radically altered, if not completely razed, during this major building phase. Describing Structure L part-by-part will facilitate understanding the whole structure.

151

V.M. STRUCTURE M - FIRST KITCHEN

PHOTO 37. - Structure M

The foundations designated Structure M actually consisted of only a brick chimney base and a ten-foot line of badly weathered rowlock brick extending northward off the northeastern corner of the hearth foundations (Photo 37, above). No other walls or associated post holes were located in the 1930s or 1980s excavations. The hearth measured 9.8 feet east-west across the back or southern side, and 4.75 feet north-south along the east and western sides. The interior of the firebox was six feet east-west and 3.5 feet in depth (north-south, interior). The east-west chimney base wall was 1.6 feet wide, and the two north-south walls were 1.2 feet wide. The bonding of the brick was with good-quality shell mortar in a somewhat casual English bond style. The bricks in the chimney base were well-fired and in reasonably good condition, in contrast to the line of rowlocks extending from the above-mentioned northeastern interior corner.

Extending westward 2.2 feet from the southwestern exterior wall were three single-course bricks laid end-to-end in a similar fashion as those brick which made up the facade between piers in Structure A. No corresponding row of brick survived on the southeastern exterior corner. If these brick, and the rowlocks described above, were indeed the remnants of the foundations for Structure M, the fireplace was again located in the corner of a room parallel to the walls as is the case with the first hearth in Structure J. Also, the wall extending northward from the hearth in Structure M and the wall abutting the eastern edge of the first hearth in Structure J were both poor in quality, similar in color, and laid as rowlocks. Both of the structures were interpreted as from the same time period (early colonial) when first uncovered in the 1930s, probably because of the similarities described above.

152

Unfortunately, Structure M was so separated from the surrounding stratigraphy and there was so little of the building remaining, that it is impossible to determine exactly when it was built. The archaeological report on the 1930s excavations (Duke 1939) mentions Structure M only briefly. Structure M was probably the first kitchen for the house, given its location of only twenty-seven feet north of the main house, and its large six-foot hearth. It, along with Structure J, appear to be the only support buildings near the main house built in the same time period as the house itself.

Not only is the construction date of Structure M obscured by the lack of stratigraphic correlation, the destruction period is also somewhat vague. The building was most certainly down by the time Structure L was constructed (1760s), and it was likely down by the time Structure K was added to J (late 1730s or 1740s).

The 1985 excavation of Structure M took place during the summer and was accomplished by one full-time crew member, who was both interpreting the site to visitors and conducting a field school while digging. The excavation of layers which may have related directly to Structure M was not completed. Artifacts recovered from the layers that were dug were few and not generally informative. Little evidence of burning inside the hearth was encountered, unlike the hearth in Structure K which contained a great deal of ash. Even the subsoil under the ash in the interior of Structure K's hearth was burned to a depth of several inches. Either the firebox in Structure M was higher than that in K, or there was shorter and less intensive use of the kitchen.

153

V.N. STRUCTURE N - SHED

PHOTO 38. - Structure N

The foundations of Structure N (Photo 38) were the only outbuilding foundations located during the 1930s excavations from which a reconstruction was based. The presently standing 10 by 14 foot shed was built where a foundation with a brick bat-paved floor was uncovered. It appears the footings for the original building were removed and a new foundation built for the reconstruction. The present building served as a kitchen for the Ball family when the kitchen addition they had been using was torn off the main house and a new kitchen was being built for them in the reconstructed east wing (Gertrude Deversa, personal communication). Indeed, a fork from Mrs. Deversa's grandmother's silver flatware service was found by the excavation crew when cleaning the shed for use as a field office. The piece was returned to Mrs. Deversa.

The only clue about the original foundations for Structure N comes from Duke's Archaeological Report:

Separated by 3 feet from the nearest storehouse foundation and by 13 feet from the marl walk, remains of another outbuilding were found opposite the southeast addition to the second kitchen. Four 9-inch walls were traced, defining a building which measured about 10 x 14 feet. Paving of brickbats was found inside. Earlier work under parts of the north and east walls indicates that the building was destroyed and rebuilt on the same site, or perhaps that the same building was raised to a higher level.

154

Two short 9-inch foundations project from the west wall south of center but are not bonded with the wall. They may have supported a small stoop, or a narrow flight of steps.

The brickwork in this foundation is of as early character as that in the first storehouse foundation. The east wall of this foundation lines with that of the third storehouse. Hence, the life of the building or buildings here indicated may have spanned that of all three storehouses.

This building cannot be positively identified but it may be the shed mentioned in Humphrey Harwood's accounts for 1777 (Duke 1939:10-11).

The storehouses to which Mr. Duke was referring are three buildings (Structures P, Q, and R) superimposed on one another. He also mentions in a footnote that he believes one of the four buildings in the group of structures east of the marl walk was built before the property had been amalgamated in 1723-24. This is based on the position of two resting just west of the lot line. Mr. Duke did not have knowledge of the long-maintained fence line along the lot line which divided the back yard from the time of the construction of the two houses to the eventual destruction of the "east wing" some time in the early nineteenth century. With this knowledge, the position of the outbuildings may relate less to the lot line and more to an endeavor to make the back yard somewhat more symmetrical after having added the mid-section of the main house. It is doubtful, though possible, that any of the four buildings were built before the construction of the mid-section, probably in the 1750s.

155

V.P. 28B AREA

Five ten-by-ten foot excavation units (70N 70E, 70N 80E, 60N 70E, 60N 80E, and 50N 80E) were opened in the area between the reconstructed shed (Structure N) and the main house, east of the central marl walk. No structural foundations were found and no features dating earlier than 1850 were encountered.

The remnants of a mortar and brick bat walkway were uncovered under Topsoil 2. The walkway ran north-south along the eastern edge of units 60N 70E and 70N 70E into the adjacent units. The walkway appeared to have been about four feet wide with brick bats laid over a marl base. An east-west marl walkway located in the central section of unit 50N 80E, also approximately four feet wide, apparently intersected the north-south walk at the same stratigraphic level, i.e., sealed by Topsoil 2. It is unknown what traffic these walkways served.

This area was also highly disturbed by utility lines and sewer pipes. A pipe, which is probably a water line, runs north-south through units 70N 70E and 60N 70E about four feet east of the 70E grid line. Another pipe or electric conduit runs diagonally through the very northwestern edge of unit 50N 80E, across 60N 80E, into the southeastern corner of 70N 80E. In addition to the two utility lines, a series of out-of-service ceramic drain or sewer pipes crosses the area. The longest pipe runs from the edge of excavation at 56.5N 70E east-southeast to the edge of excavation at 52N 90E. This pipe is cut twice by the utility lines. Another small segment of sewer pipe was located in the extreme southwestern corner of 60N 70E, and another in the southwestern corner of 50N 80E. This line may have been connected to old down-spouts from the main house. No other significant features were found in this section.

Another section known to contain brick foundations was opened north of the reconstructed shed. The area between 135N 75-95E and 100N 75-95E contained the remains of three structure foundations superimposed on one another—Structures P, Q, and R (Photo 39).

158

V.S. THE WELL

PHOTO 40. - The Well

The Randolph House Well (Photo 40) was located in the northeastern quadrant of unit 140N 20E and the northwestern quadrant of 140N 30E, approximately twenty feet north of the exterior hearth for Structure J. The interior diameter of the roughly circular well shaft was about four feet. Apparently a good water source, the well was used until 1920, but had been completely filled by the time the Skillman House was constructed in 1924. As wells were commonly cleaned of debris while in use (Noël Hume 1969a), it is not likely that any material relating to the Randolph occupation of the property would be recovered by digging. Recovery of debris from this century would not yield information sufficiently important to risk the endangerment of the excavators or the expenditure of the large amount of money inherent in well excavation. The depth of the well is therefore unknown.

The well was first uncovered in the 1955 cross-trenching of the Skillman's front yard. It was drawn and photographed by Mr. Knight at that time. No other wells have been located on the Randolph property. Either this well served the house and kitchens since the early eighteenth century, when the buildings were constructed, or water was carried from a nearby spring located on lot # 185 (as numbered on the Bucktrout Map). Another well may have been located near the stables, which are believed to be near Scotland Street some 500 feet north of the house.

The well was not constructed of the usual wedge-shaped well brick, but of ordinary rectangular building brick. The five foot square well-curb was added to the original well sometime after its construction. The curb may be the result of Humphrey Harwood's repair of the well, noted in his ledger of August 19, 1785.

159

A single row of brick—different in color and hardness from those used in the well—extended eastward from the southeastern corner of the well curb, six feet from where it makes a right angle southward eleven feet, then eastward again for three feet, intersecting the southwestern corner of Structure H. The purpose of this row of brick or "appendage" is unknown. Whether it was associated with the structure or the well (or both) has not been determined.

The well-curb and Structure G are on parallel lines with each other, but with none of the rest of the outbuildings in the 28H area. This orientation may be a result of Harwood's repairs, but why it was done is wholly conjectural.

The well was excavated to a depth of 4.5 feet during the most recent excavations, approximately a foot and a half beyond the depth excavated in 1955. This was done to make certain the fill contained primarily only artifacts from this century, which was indeed the case.

160

1. PRE-1699 (Fig. 63)

This period could as easily be divided into two sections, examining separately pre-1699 English colonial activities, specifically the Middle Plantation settlement, and the activities of the aboriginal people inhabiting the area for the several thousand years before the English. However, the prehistory of Tidewater Virginia is quite beyond the scope or intention of this report. Indian remains specific to the Peyton Randolph back lot were virtually non-existent. Even the Middle Plantation Era, beginning in 1622 and lasting until the establishment of Williamsburg as the colonial capital in 1699, was represented only scantily on the lot.

Aside from a few residual or displaced projectile points and pottery, no aboriginal artifacts or features were located during the archaeological excavations of the 1930s, 1950s, or 1980s. The only reference to aboriginal remains is in the 1952 Architectural Report of the Randolph-Peachy House: "Indians are buried on the east side of the yard. Mr. Chorley and Mrs. Ball both agree they should not be disturbed" (Kocher and Duke 1952:69). The "east side" apparently refers to the area directly in front of or behind the reconstructed "east wing" of the Peyton Randolph House. The statement was made on July 12, 1938, before the east wing had been reconstructed, and before the National Park Service built the Colonial Parkway tunnel under the Historic Area. Sources for the information regarding the alleged burials are unknown, and no archaeology, other than uncovering the foundations of the so-called "east wing," has taken place in that part of the yard. Construction of the tunnel in 1940 disturbed much of the east side of the yard (and the newly reconstructed "east wing"), as illustrated in Photo 41 (above right). If the graves existed, they may have been destroyed at this time.

Photo 41 - Effects of Tunnel Construction

The only discernible Middle Plantation Era feature represented at the Peyton Randolph site was a drainage or boundary ditch which traversed the site from the northwestern edge of excavation to the area near the east-central edge. The northern and southern extents of the feature were not determined. The four-foot-wide trench was partially excavated in the 28G area of the site, but yielded no 163 Fig. 63 - Periodization Map 1 - Pre-1699 164 significant or even datable artifacts. The feature was found at the lowest stratigraphic level, intruding only sterile subsoil. This fact, and the observance that the feature was not oriented in the usual north-south, east-west plan as are most Williamsburg Era features, place the ditch in the pre-1699 period. The lack of known orientation of the ditch to anything but the landscape may reflect a seventeenth-century attitude in which man adapts to the landscape, rather than attempting to impose order on the landscape as was done with the establishment of the City of Williamsburg with a gridded town plan.

2. 1699-1715 [No Map]

This period was chosen in order to represent the time between the establishment of the City of Williamsburg and the first recorded purchase of the lot under investigation. During this time period, the lot was owned by the city and no apparent activity took place that would leave its mark in the archaeological record.

3. 1715-1724 (Fig. 64)

Lot 207, presumably the lot excavated, was sold by the city to William Robertson in late 1714 along with several other lots. In all likelihood it was Robertson who built the main house, Structure J, Structure M, and Structure A. It is not known whether Robertson lived on the lot or rented to tenants. Structures A, J, and the main house all appear to be dwellings rather than dependencies, however, and Robertson probably rented all three structures to tenants. The tenant in Structure A constructed at least two of the four planting beds uncovered near the building during this period. The main house was definitely the grandest of the three structures, probably having an exterior kitchen (Structure M). Structure J was a rather rude hall-and-parlour structure, which would have probably been less out of place architecturally in mid-seventeenth century Jamestown than in early Williamsburg. Structure A was a simple structure built on brick piers over the Middle Plantation ditch described earlier. All three buildings had fireplaces in the corners of the rooms rather than along a wall, as was more common.

John Holloway, an entrepreneur from Norfolk, purchased lots 237, 236, 207, and 208 from William Robertson in 1723. At that time the lots already contained houses, an orchard, and a windmill (Stephenson 1952). It is doubtful that Holloway changed much on Lot 207 because he sold it to John Randolph only six months after he bought it. According to the York County Deeds and Bonds, John Randolph lived adjacent to the lot he purchased from Holloway at the time of sale. This is presumably lot 237, and the house in which he may have lived was Structure S, the so-called "east wing" of the restored Peyton Randolph House. When Randolph acquired lot 237 is unknown.

Physical evidence recovered, placing these buildings in this 1715-1724 period, is adequate if not substantial. Dendrochronological dating of the rafters and joists in the main house place its construction date at 1718 (A.I.D. 1983). Structure A was erected on a virtually sterile "A" horizon which appears to be the topsoil which existed before any Williamsburg related activities began. It was also constructed over the Middle Plantation ditch which apparently had not yet been filled because a support wall had to be placed across the ditch in order to rest the southwestern 165 Fig. 64 - Periodization Map 2 - 1715-1724 166 pier of Structure A on a solid footing. Structure D, probably a small service building relating to Structure A, was constructed during this period as well. Structure J's rudimentary architecture, and the evidence of its having been expanded several times during its existence, place it in this early period as well. Structure M is assigned to the early period principally because of the placement of the chimney in the corner of a (the?) room as in the other structures. Also, the hearth size seems to indicate a cooking or kitchen function. Structure K served this purpose in the late 1730s or early 1740s, and Structure L immediately after that. Structure M, which is truncated by Structure L, must logically pre-date that building, and probably Structure K as well. It is an obvious candidate for the earliest kitchen on the lot. For these reasons, and that the structure appears to have been of post construction, Structure M can most likely be seen as contemporary with A and J.

The appearance of the lot during this early Williamsburg period reflects a less gracious and affluent arrangement than exhibited in later times. With the recent movement of the capital and all of the government of the largest American colony to a newly-established town, housing was probably scarce and at times less than luxurious. Money could be made by those who could afford to buy land and build anything, no matter how unpretentious, within the prescribed two years of purchase. Subsequently renting the land to tenants would probably aid significantly in this process and would result in little or no elaboration of the lots.

4. 1724-1737 (Fig. 65)

This period began at the time the property came into the possession of the Randolph family and ended at the death of Sir John Randolph. The immediate back lot of the main house, which became his residence, begins to take on the appearance of the "typical" eighteenth-century dwelling with several support buildings. Architecturally, Structure J probably received a lean-to on the east side of the structure. This is based on the finding of an earlier wall under the eastern support wall for the later Structure L. Structures G and H appear to have been constructed as a dairy and smokehouse. Considering the positions of Structures G and H in relation to that of J, it seems that the role of Structure J became that of a kitchen, replacing Structure M, or that of servant quarters, in either case relating directly to the main house. Structure A, on the other hand, appears to have remained a tenant house. The construction of, two additional planting beds near Structure A, similar in composition to the first two, indicates a continuance of the same resident in this structure after Sir John's acquisition of the lot. Occupation layers from around the structure also reflect its use as a residence, at a social or economic station higher than would be expected of a slave, but not as high as that of the Randolph family.

5. 1737-1755 (Fig. 66)

After the death of Sir John, his wife Lady Susannah inherited the property. She retained possession until her death sometime shortly after 1754. It was she, probably with the help of her increasingly-prominent son Peyton, who expanded the influence of the main house into the northern section of lot 207. The planting beds, Structure A, and Structure D almost certainly ceased to be of use during this 167 Fig. 65 - Periodization Map 3 - 1724-1737 168 Fig. 66 - Periodization Map 4 - 1737-1755 169 time period. Both buildings were razed by the latter part of the period, and two others constructed. Structures E and F appear to be support buildings for the main house rather than residences. This brought the rest of the lot under the influence of the main house. Near the house, Lady Randolph probably added Structure K onto Structure J, making a multi-function, L-shaped building. Structure J was altered as well, moving the old interior fireplace on the north wall to the exterior and creating a small bake-oven next to it. Structure K is placed in this time period for several reasons: it necessarily pre-dates the construction of Structure L, it post-dates the construction of the lean-to on Structure J, and in all likelihood its construction pre-dates the addition of the mid-section onto the main house.

By the mid-eighteenth century, the house and grounds seemed to have taken on the appearance of an urban estate with a large kitchen and several service buildings. Structure S, if still owned by the Randolphs, may have been a tenant house. Although a fence line was maintained along the line between the two properties, archaeology revealed that little activity had taken place in the northern section of lot 237. No archaeology has been done in the area immediately behind the reconstructed house.

6. 1755-1775 (Fig. 67)

This is the period in which Peyton Randolph was the sole owner of the house and lot. Peyton inherited the town house from his mother, continuing to live there until his death in 1775. It was probably during this period that the most dramatic changes occurred, both in the yard and in the main house. A recent dendrochronology study done on the mid-section of the present house (Heikkenen 1984 [see Appendix 10]) indicated that the timbers for the roof were cut in 1753. This mid-section addition re-oriented the direction in which the house was facing from the west, toward the Palace, to the south, toward Market Square. Structure S was incorporated in the design of the front of the house, creating the effect of a seven-bay Georgian house. No door, however, existed between the mid-section and Structure S, although they were physically connected. There were indications that the interior trim in the older section of the house was replaced at the same time (Graham 1985). The re-modeling and upgrading of the main house into a grand, if not opulent, dwelling was also reflected in changes made to the exterior service buildings as well. The rather organic, rambling kitchen complex was "cleaned up." Structure K was razed, probably so the view from the large arched window on the staircase of the new mid-section would not be obstructed by the chimney in K. Most of Structure J was either torn down or greatly altered. The south wall of J was removed, and a vaulted wine cellar and cooking room was added to the south end of the structure. Structure L resulted from the building and re-modeling. The building was indeed a grand kitchen, possibly two floors, with a laundry, quarters, and a covered passage-way to the main house. The completed structure was rectangular, twenty feet wide and about forty-eight feet long.

The addition of the mid-section to the house early in this period caused a re-orientation of the back yard as well. A marl walkway was constructed leading northward from the mid-section, possibly to the end of the lot. At least two outbuildings were constructed east of this walk, just west of the 207/237 property line. The function of Structure N is not known, but Structure R was probably a dairy.

170

Figure 67 - Periodization Map 5 - 1755-1775

171

In contrast with the activity demonstrated in the immediate back yard, the back lot in the 28G area appeared to receive little attention. The hole created by the removal of Structure A was filled in the early 1750s. A layer of plaster several inches thick was deposited within the interior of the foundations. It is not unreasonable to assume that the plaster was a result of the construction of the mid-section, subsequently dumped in the hole left by the removal of Structure A's super-structure. Structure B, function unknown, was built shortly after the demise of Structure A.

The re-orientation of the main house to the south, facing Market Square, reflects a change in the attitude and economy of the town. Market Square, rather than the Governor's Palace, became the hub of activity in mid-century Williamsburg. Both the Archibald Blair and the St. George Tucker Houses also faced the Square. The fact that they probably blocked the view of the Palace and Palace Green to the west-facing Randolph House may have also been a factor in the change. Indeed, with the addition of the mid-section, a re-orientation to the south was necessary in order for the house to approach Georgian style. The importance of Market Square is illustrated by the construction of the market house around 1757 and the present court house in 1770. The largest private structure in town, Tazewell Hall, built by Peyton's brother John, faced the opposite side of Market Square from South England Street. Peyton and John then shared opposite ends of the "vista" created by Market Square.

The large-scale building and re-modeling that took place in this period reflects a new attitude towards entertaining and social standing which was exhibited in other buildings in Williamsburg. Raleigh Tavern, Wetherburn's, and the Governor's Palace all acquired large rooms for entertainment purposes (Chappell 1981). During the early part of this time period, Williamsburg attained its apogee of power and influence, a small part of which is reflected in the changes that occurred at the Peyton Randolph House. The changes also reflect a rise in Peyton's personal power and influence as Speaker of the House of Burgesses and his increasing involvement in inter-colonial affairs (Stephenson 1952).

7. 1775-1783 [No Map]

After Peyton Randolph's death in 1775, his wife, Betty, inherited the house and property, which she retained until her death eight years later. Though no major construction or other changes in the landscape of the yard can be assigned to this short period, Humphrey Harwood's ledger indicates that he made several minor repairs to the main house and covered way in 1778. It is not surprising that few physical changes took place during this period. Williamsburg was occupied by the British Army from 1777 to 1781, and lost its status as the capital of the Commonwealth in 1781. The principal objectives of the people of Williamsburg were probably surviving and maintaining, rather than building.

8. 1783-1800 (Fig. 68)

Joseph Hornsby bought the house and property at an auction held after the death of Betty Randolph in 1783. Several minor changes appear to have taken place during his approximately seventeen years of ownership. It is known through 172 Figure 68 - Periodization Map 6 - 1783-1800 173 Humphrey Harwood's ledger (see Appendix 2) that Hornsby had a dairy repaired on November 5, 1784 and the granary underpinned on August 8, 1785. The well was repaired on August 19, 1785. Structure E exhibited evidence of having been underpinned with a series of brick piers constructed on 2.5' centers along the length of the building. Artifacts found in association with one of the piers had a terminus post quem of 1779. In addition to the repair work historically documented, Hornsby appears to have constructed an additional outbuilding in the 28G area. Structure C, the south wall of which also had a TPQ of 1779, appears to have been constructed over the forgotten foundations of Structure D, pulled down at least thirty years earlier. The function of the building could not be determined. Hornsby also extended the walkway from the mid-section of the house past the dairy (Structure R/Q), all the way to Structure C. Other outbuildings may have been constructed along this walkway, but later residential construction in the area obliterated any archaeological record of their existence.

In addition to the changes made in the back lot, Hornsby also appears to have laid a marl paving east of the kitchen complex and built a porch appendage to the east side of the same structure.

9. 1800-1860 (Fig. 69)

The Peachy family bought the Peyton Randolph House and lots some time around 1800, although the precise date is unknown. They retained the property for approximately sixty years. Little archaeological evidence was recovered indicating major changes in the area immediately behind the house during the Peachy period. The covered way was probably removed and the porch tower, presently appearing as a reconstruction on the north elevation of the west wing of the house, was constructed. Structure S must have fallen into ill repair, and was also removed. Any archaeological evidence of its removal in this period, however, was destroyed by the 1930s excavations. Since the structure does appear on the Frenchman's Map of 1782, but does not appear in any of the early photographs of the house, it may be concluded that this is the likeliest period of its demise. Shortly after their acquisition of the property, the Peachys dumped a thick layer of refuse over the marl paving east of the kitchen complex, apparently attempting to improve drainage in the area. This refuse deposit contains a great amount of late eighteenth-century domestic artifacts, but its origin is, unfortunately, unknown. The dairy, Structure P, was re-built in this period, apparently surviving until the early twentieth century.

Structure E, probably the granary repaired by Humphrey Harwood in 1785, underwent some changes during the Peachy occupation. The west wall appears to have been entirely replaced. Sand mortar was used in the construction of the wall, and the bricks were primarily used brick bats. The construction was of extremely poor quality; the wall was not straight nor the bonding in any particular order. This wall may have replaced an earthfast post-construction wall, as no archaeological evidence was found of a previous brick wall, though post holes were evident under the existing west wall. Also added to Structure E during this period were two interior wall support footings at right angles to the east and west walls, as well as a corresponding central support pier. Structure B, function unknown, was probably constructed during this period.

174

Figure 69 - Periodization Map 7 - 1800-1860

175

10. 1860-1900 (Fig. 70)

The beginning date for this period was chosen because the house left the possession of the Peachy family at approximately this time, though the exact date is unknown. It also marks the beginning of an era of poor economic conditions in the Williamsburg area, from the beginning of the Civil War until the beginning of the "Restoration." Williamsburg was occupied, once again, by enemy (Union) troops during most of the Civil War, and the subsequent years were a time of slow growth. The Peyton Randolph House changed ownership four times before the end of the century, and an 1870s photograph (Photo 42, below) shows the house in a less than well-maintained condition.

Photo 42 - South Elevation of House circa 1870

Several changes took place during the period which were documented in documents and photographs, as well as in the archaeology. Structure L, the large kitchen, was altered, receiving a facade of board-and-batten siding and a layer of clay fill in the interior and along the western exterior. The clay may have been an attempt to seal moisture from the building. The old wine cellar under the kitchen was sealed with a layer of cement, converting it into a cistern for water collection. The last half of the nineteenth century also saw the erection of a water-pumping windmill over the well.

All buildings in the 28G Area were torn down during the 1860 to 1900 period. The two pecan trees which still remain on the site were planted, and most of the area north of the present access drive was turned into a plowed field early in this period.

176

Figure 70 - Periodization Map 8 - 1860-1900

177

11. 1900-1920 (Fig. 71)

This period defines the approximate Warburton ownership of the Peyton Randolph House, a time just prior to the beginning of residents' recollections. Structure L was finally removed. An early twentieth-century photograph (Photo 46) indicates that Structure L had been removed by the second decade of the twentieth century, though a past resident of the Randolph House does recollect a hand pump over the cistern in the early 1920s. The same photograph also reveals that two seemingly eighteenth-century outbuildings, a smokehouse and a dairy, were still standing in this period. Both were probably removed during these two decades. The windmill was offered for sale in a 1920 Virginia Gazette. It was probably dismantled the same year. Any remaining outbuildings were removed. A kitchen and maid's quarters was added on to the rear of the main house either late in the last period or early in this one. It was not investigated archaeologically.

12. 1920-1938 (Fig. 72)

This period reflects the last individual ownership of the Randolph property. The Wilson/Ball family bought the house early in the 1920s, after having rented the dwelling from real estate brokers Williamsburg, Inc. since 1919. The original lot had by now been divided into several smaller lots. Two residential structures were built on the sub-divided lots in 1924 and 1926. Both buildings had basements, destroying any archaeological evidence of previous structures or features. City water was connected and telephone and electric cables put in. Use of the cellar as a cistern was discontinued and it was then filled with household debris in circa 1926. The Colonial Williamsburg Foundation acquired the Peyton Randolph House from Mrs. Ball in 1938, though she continued to live in the main house until 1947. The first archaeological investigation of the house and yard was conducted by CWF Architecture Department in 1938. After an archaeological survey was made of the foundations of Structure S, it was reconstructed by the Foundation.

13. 1938-1985 (Fig. 73)

Radical changes took place in this forty-seven year period of Foundation ownership. The main house was fully restored in the early 1940s to its physical appearance (according to the Architecture Department) during the Peyton Randolph period (1755-1775). Structure N was reconstructed to serve as a kitchen for the Ball family while the main house was being restored.

Archaeological cross-trenching took place in 1955 in the back lot, and a replica of Robertson's windmill was built in its present location. Colonial Williamsburg had acquired the remainder of the land associated with the Randolph House by the 1960s. Both of the 1920s houses were removed by the early 1970s. The land between the windmill and the main house was undisturbed until full-scale archaeology began is 1977-1978, and was continued in 1982-1985.

178

Figure 71 - Periodization Map 9 - 1900-1920

179

Figure 72 - Periodization Map 10 - 1920-1938

180

Figure 73 - Periodization Map 11 - 1938-1985

181

Having been acquired by an outdoor museum and becoming a public exhibition building rather than a private residence², the Randolph House has ceased becoming part of any individual's outward manifestation of social standing, self-worth, or idiosyncrasies. But it will no doubt continue to change, both in structure and in the ways it influences people's understanding of the eighteenth century in Williamsburg.

182

184

FIGURE 74

186

FIGURE 75

188

Figure 76

190

FIGURE 77

192

FIGURE 78

194

FIGURE 79

196

FIGURE 80

198

FIGURE 81

200

FIGURE 82

202

FIGURE 83

204

FIGURE 84

206

FIGURE 85

209

PRINCIPAL MEASUREMENTS OF ILLUSTRATED ARTIFACTS

FIGURE 1 (Catalog number 4004-28G-89J) Height of vessel: 8 1/8" (20.8 cm)
Diameter of base: 4 ¾" (12.06 cm)
Diameter of shoulder: 4 ¾" (12.08 cm)
Height of push-up: 1 25/32" (4.5 cm)
Diameter of pontil scar: 2 ¾" (8.9 cm)
Diameter of bore: 5/8" (1.b8 cm)

FIGURE 2A (Catalog number 4006-28G-89N)
Approx. diameter of base: 5 ¾" (14.6cm)

FIGURE 2B (Catalog number 4005-28G-2175K)
Diameter of base: 4 ¾" (12.06cm)

FIGURE 2C (Catalog number 4005-28G-84J)
Diameter of base: 4 ½" (11.43cm)
Maximum height of vessel: 5 15/16" (15.08cm)

FIGURE 3 (Catalog number 4001-28G-170)
Thickness of tile: 6/32" (0.47cm)

FIGURE 4A (Catalog number 3897-28G-89E)
Diameter of saucer: Approx. 4 ½" (11.4cm)
Height of saucer: Approx. ¾" (1.9cm)
Diameter of standing ring: Approx. 2 ¼" (5.7cm)

FIGURE 4B (Catalog number 3900-28G-84N)
Estimated height of cup: Approx. 2 ½" (6.3cm)
Diameter of cup rim: 2 17/32" (6.43cm)

FIGURE 4C (Catalog number 2881-28G-89N)
Diameter of vessel: Approx. 9" (22.8cm)
Height of vessel: 1 9/16" (3.97cm)

FIGURE 4D (Catalog number 3909-28G-2170E)
Diameter of vessel: 2 ¾" (6.9cm)

FIGURE 4E (Catalog number 3880-28G-2175K)
Estimated diameter of plate: Approx. 9" (22.0cm)

FIGURE 5A (Catalog number 3865-28G-89N)
Estimated diameter of vessel: Approx. 10" (25.4 cm)

FIGURE 5B (Catalog number 3848-28G-89E)
Diameter of vessel: 5" (12.7cm)

FIGURE 5C (Catalog number 3849-28G-89K)
Diameter of plate: Approx. 9" (22.9cm)

FIGURE 5D (Catalog number 3852-28G-89K)
Diameter of plate: Approx. 9" (22.9cm)

FIGURE 6A (Catalog number 4009-E.R.2173P-28G)
Diameter of rim: Approx. 15 ½" (39.2cm)

FIGURE 6B (Catalog number 4007-28G-108H)
Diameter of rim: 4 ½" (11.4) cm

FIGURE 6C (Catalog number 4008-28G-108H)
Diameter of vessel: Approx. 10" (25.4cm)

FIGURE 7A (Catalog number 2211-28G-2067)
Diameter of standing ring: 1 17/32"

FIGURE 7B (Catalog number 2209-28G-2066)
Diameter of vessel: 4 ½" (11.4cm)
Diameter of standing ring: 2 ¼" (5.7cm)
Height of vessel: 7/8" (2.2cm)

FIGURE 7C (Catalog number 2210-28G-1548F)
Estimated diameter of plate: 9" (22.8cm)

FIGURE 9 (Catalog number 2251-28G-1075)
Estimated diameter of base: 4 ¼" (10.7cm)

FIGURE 10A (Catalog number 2238-28G-1074)
Diameter of rim: 1 7/8" (4.7cm)
Height of vessel: 1 11/16" (4.2cm)

FIGURE 10B (Catalog number 2245-28G-1075)
Diameter of rim: 2 ½" (6.3cm)
Height of vessel: 1 15/16" (4.9cm)

FIGURE 10C (Catalog number 2218-28G-1069)
Diameter of base: Approx. 7 ½" (19.0cm)

FIGURE 10D (Catalog number 2225-28G-1533G)
Maximum diameter of base: Approx. 8" (20.3cm)

FIGURE 10E (Catalog number 2215-28G-1069)
Estimated diameter of base: Approx. 7" (17.7cm)

FIGURE 11A (Catalog number 2261-28G-2068)
Estimated diameter of vessel: 13 ½" (34.3 cm)

FIGURE 11B (Catalog number 2257-28G-1069)
Diameter of rim: Approx. 13" (33.0cm)
Height of vessel: 2 5/8" (6.66cm)

FIGURE 11C (Catalog number 2264-28G-2066)
Estimated diameter of vessel: 15 ½"

FIGURE 12 (Catalog number 4003-28H-748)
Maximum diameter of seal: 1 3/8" (3.49cm)

210

ARCHAEOLOGICAL REPORT (PEYTON RANDOLPH)
Compiled by Francis Duke, September 1, 1939

OUTBUILDINGS WEST OF WALK

MARL WALK AT CENTER

Northward from a point near the center of the present house runs a path about 6 feet wide, edged with brick on the west and surfaced with layers of several substances, predominantly marl. This walk was probably in use for many years, as it was built up at various times with marl, crushed, ashes and other debris. Its thickness ranged from 5 to 9 inches. The original walk was of marl.

OUTBUILDINGS EAST OF WALK

EAST WING OF HOUSE

A foundation was uncovered of brickwork 14 to 15 inches thick and about 20 x 35 feet in extent. On the south side its face recedes about 4 inches from the face of the present house, but on the north its face projects nearly a foot beyond the house.³⁸

The east wall of the house may be built upon the west wall of the "wing" as its foundation;³⁹ but investigation is not yet complete. What appears to be a 4-inch brick veneer covering the house wall is set back 4 inches from the east face of the foundation wall.⁴⁰

Brick	9 x 4 ½ x 2 ½.	Reddish salmon. Common bond.
Mortar	Shell.
Condition	Unfit for reuse.

CHRONOLOGICAL SUMMARY

Before 1724 the west lot probably contained the westernmost of the original dwellings, the first kitchen and perhaps one or more of the group smokehouse-storehouse-shed. The east lot contained the dwelling and perhaps no outbuildings. It was probably not long before the two dwellings were joined, or long after that before the east wing (the old east dwelling) disappeared.⁴⁴

Some time before 1782, probably before 1775,⁴⁵ the first kitchen gave way to the second⁴⁶ and the second to the third. The third kitchen was connected with the house for a time (probably still before 1775) by the Covered Way. The Covered Way disappeared at a time unknown but while the third kitchen was still in use, for other service entrances were provided to replace it.

The subsequent history of the outbuildings is one of progressive desuetude, decay and disintegration.

CHRONOLOGICAL TABLE, BASED ON REPORT OF THE DEPARTMENT OF RESEARCH AND RECORD
July 26, 1938

1714	8 lots47 deeded to William Robertson with proviso requiring erection of buildings within two years.
1715	Philip Ludwell buys two southeastern lots.48
1716	Houses presumably completed, since Robertson and Ludwell retain titles.
1723	John Holloway buys 4 lots,49 with houses and a windmill from Robertson for £80.
1723/24	John Randolph buys one of Holloway's two southwestern lots.
1724	Randolph buys the second of Holloway's lots.
1737	Randolph dies, mentioning "house in Williamsburg" in his will. His son Peyton Randolph inherits the property.
1775	Peyton Randolph dies.
1776	Inventory mentions lumber in storehouse .
1777	"Shead" mentioned in Humphrey Harwood's accounts.
1778	Covered Way mentioned by Harwood.
1783	Betty Randolph dies. House bought from estate by Joseph Hornsby.

SOURCES

• 1.

  Fragments:

  Few fragments, and none of importance, were found among the outbuildings. The east wing excavation yielded some china fragments in unusually good preservation, some o f them being almost whole pieces. But nothing was found of particular architectural interest. Fragments were consigned to the Educational Department (Archaeology).

• 2.

  Photographs:

  5818	Wine cellar from east
  5819	Southeast kitchen addition from south
  5820	Close-up of kitchens from southwest
  5821	North kitchen chimneys from north
  5822	Wine-cellar steps from north
  5823	"Storehouse" from northwest
  5824	"Shed" from south
  5825	"Shed" from northwest
  5901	"Storehouse" from northwest
  5902	North kitchen chimneys from south
  5903	Forge and chimneys from northwest
  5904	Service entrance foundations from northeast
  5905	South chimney from north
  5906	Service entrance foundations from west
  5907	East wing basement stair from southwest
  5908	East wing chimney from south
  5909	Southwest corner of east wing from south
  5910	East wing basement stair from northwest
  5911	Basement stair east of service entrance foundations from north

  Points from which photographs were taken are indicated on Archaeological Survey Drawing. Prints may be found in record books in offices of Colonial Williamsburg.

• 3.

  Archaeological Field Notes

  In separate file

• 4.

  Archaeological Survey Drawing

  A print is filed with the original of this report.

• 5.

  Research Report, "Randolph-Peachy House," July 26, 1938

  A copy is filed with the original of this report.

• 6.

  Report of Conference, July 12, 1938, between Mrs. Ball and Messrs Kendrew and Reed

  A copy is filed with the original of this report.

Photo 5818 - Wine Cellar from East

Photo 5819 - Southeast Kitchen Addition from South

Photo 5820 - Close up of Kitchens from Southwest

Photo 5822 - Wine Cellar Steps from North

Photo - 5823 - "Storehouse" from Northwest

Photo 5824 - "Shed" from South

Photo 5901 - "Storehouse" from Northwest

Photo 5902 - North Kitchen Chimneys from Northwest

Photo 5903 - Forge and Chimneys from Northwest

Photo 5904 - Service Entrance Foundations from Northeast

Photo 5905 - South Chimney from North

Photo 5906 - Service Entrance Foundations from West

Photo 5907 - East Wing Basement Stair from Southwest

Photo 5908 - East Wing Chimney from South

Photo 5910 - East Wing Basement Stair from Northwest

Photo 5911 - Basement Stair East of Service Entrance Foundations, from North

Photo 6113 - Porch Foundations on West Side of Main House

Photo 6134 - Porch Foundations in Front of Main House

Photo 6480 - Rear Mid-Section, Porch Foundation and Bulkhead Entrance

PEYTON RANDOLPH HOUSE
1978
INTERIM NOTES

BY
ERIC KLINGELHOFER

INTRODUCTION

From January 13 to March 22, the Department of Archaeology conducted excavations at the Peyton Randolph House, Lot 28, Area G, as part of a long term project to investigate and restore the outbuildings and gardens on the property. The work was supervised by William Henry and was restricted to the remains of an outbuilding located at the northwest corner of the property. Only a temporary grid was established for the area as assistance from C.W. surveyors was unavailable.

Excavations carried out by archaeologist James A. Knight in 1955 had located the foundations of this building and others along the western property edge. This structure measured 20'4" by 16'6" and had a corner chimney at the northeast wall angle. As the 1978 excavations were restricted to the west half of the building, we were unable to examine the chimney, but we did find that the layer of a gravel bed for a bus turn-around in 1975 apparently completely destroyed a chimney foundation which was attributed to the colonial period by Mr. Knight. It is possible, though unlikely, that evidence for this later occupation of the spot will have survived in the area immediately east of the 1978 trenches.

The excavation found four different groups of features and layers: the body of the building, with demolition debris within the walls; layers of occupation and construction material associated with the building; a series of fence post holes cutting through the building remains; and an earlier ditch ascribed to the Middle Plantation period. These groups will be described in detail within the body of this interim report.

All work was carried out under the protection of a wooden shed (originally constructed for repair of the windmill in 1977). The open sides of the shed were covered with plastic sheeting, and an electric/kerosine heater served to thaw the frozen ground and permit digging in relative comfort. Two discoveries were unwelcome intrusions into the excavation area. A C.W. lightning rod cable was found running inside the north edge of the trench. It protected the large tree near the site, and ran to utility pipes in North England Street. This has been moved to the north where further excavation in that direction may find it. Secondly, a utility trench which was ascribed to VEPCO (though it may have been a phone line) was found when excavations expanded 4'0" west of the property line. This cable was untouched and its trench was rapidly backfilled.

The excavation units for the 1978 season consisted of E.R. 2169-79-28.G. The site started with two ten foot square, E.R. 2169 and 2170, which were joined when the balk, E.R. 2171, was removed. E.R. 2172 was used for spoil. The excavation later expanded 4'0" to the west, with E.R. 2173, 74, and 75 representing the western areas. E.R. 2176 was used for a VEPCO utility trench, and E.R. 2177 and E.R. 2178 replaced E.R. 2170 and E.R. 2169 respectively, after the last letters were used up. E.R. 2179 was given to a layer of pipe stems found within the stratigraphic sequence west of Structure A.

At the end of the excavation, all layers west of the building were taken down to natural, if their removal would not undercut the wall remains. Boarding was firmly placed against the exterior face of the west wall to give it additional support. Finally, all exposed surfaces were covered with plastic, secured by nails.

LATER FEATURES

EARLIER FEATURES

There was only one feature which could be determined to have been in existence before Structure A. A ditch ran northwest-southeast across the site, continuing under the wall foundation of the building. The ditch was 4'0" wide and 1'02" deep with sloping sides and a curved bottom. It was filled with humic grey loam (E.R. 2173U,W), containing only a few nail fragments. The ditch can be no later than the first quarter of the century, after which the office was constructed. Indeed, the fact that the cutting runs totally askew to the orientation of the colonial city, suggests that it preceded the layout of the town, and so was probably a Middle Plantation field ditch.

LAYER CORRELATION CHART

2175A	=	2169A; 70A; 71A; 73A; 74A
2175E	=	2173H; 2174B
2175H,F	=	2174C
2175J	=	2179Y; 71E; 73K; 74F
2175K	=	2169J; 70E; 71F; 73M; 74G
2175L	=	2170J; 71G; 73N; 74H
2175M,N	=	2169Z; 71H; 73P, R; 74J; 77D; 78A
2175P	=	2177E
2175T	=	2171K; 73S; 74K; 77G; 78D
2175V	=	2173T; 77H

POLLEN AND PARASITE ANALYSIS
OF THE PEYTON RANDOLPH SITE
WILLIAMSBURG, VIRGINIA

Karl J. Reinhard
Laboratory of Parasitology
Palynology Laboratory
Department of Biology
Texas A&M University
College Station, Texas 77843

INTRODUCTION

The utility of palynological analysis in archaeological research has long been recognised. Ecological changes, both dependent and independent of human activity, have been traced through palynological analysis (Martin 1963, Euler et al. 1979, Wycoff 1977). Dietary habits have also been elucidated by pollen studies (Adams 1980, Gish 1981, Williams-Dean and Bryant 1975, Reinhard 1985a). More innovative studies have focused on identification of agricultural plots (Berlin et al. 1977) and community growth (Hill and Hevly 1968). The utility of palynological research, however, is dictated at any given site by pollen preservation dependent on environmental conditions, pedology, and biological degradation (Helvy 1981, Bryant and Holloway 1982, Dimbleby 1957).

More recently, parasitological study has emerged as a new aspect of archaeological research. The theoretical background for such study has been clearly defined (Kliks 1983) and many studies have been accomplished (Fry 1977, Reinhard 1985b, Reinhard et al. 1985a). The utility of parasitological analysis is, however, effected by aspects of degradation, just is as the case with palynology.

With these thoughts in mind, palynological and parasitological surveys of soils from the Peyton Randolph site by the Williamsburg Foundation were undertaken. The primary goal of the analyses was assessment of the value of such research at the site.

METHODS

The basic procedures of pollen analysis have been established (Faegri and Iverson 1975). In its simplest form, palynological analysis involves the chemical digestion of soil constituents, leaving behind only the most durable remains which include pollen grains. Carbonates are dissolved by hydrochloric acid, most silicates by hydrofluoric acid, cellulose and hemicellulose by an acetolysis treatment (Erdtman 1960), and humic acids (partly decomposed organics) by potassium hydroxide or sodium hydroxide.

Beyond these basic techniques, there can be several modifications (Bryant and Holloway 1983). For example, sonication will break up some organics that do not normally dissolve in chemical treatment. Highly concentrated hydrofluoric acid can be used to more effectively dissolve silicates, and heavy density separation can be used to float pollen out of sediments that respond neither to chemical treatment or sonication.

To determine the number of pollen grains held in a soil sample, tracer spores of the fern Lycopodium are added. By knowing the number of tracer spores per gram of sediment, the proportion of spores to pollen grains provides a way to quantify the pollen yield per gram of soil. In the Williamsburg analysis, tracer spores were added to each soil sample processed.

The soils of Williamsburg are very resistant to pollen processing. Even though the basic procedures outlined above were applied, the amounts of inert organics and silicates were too great for effective analysis. A second processing was done involving more extensive treatment in concentrated (70%) hydrofluoric acid with initial sonication. This process also proved ineffective. Finally, a third extraction was performed utilizing heavy density separation in zinc chloride. Only in this final extraction did the soil samples clean sufficiently for analysis.

The extraction procedure found most effective with the Williamsburg soil samples is outlined below:

• 1) Two lycopodium tablets containing approximately 11,200 spores were added to 25 grams of soil.
• 2) The soil was treated with dilute hydrochloric acid until reaction with the carbonates stopped.
• 3) The soil/dilute hydrochloric acid mixture was transferred to a 150 ml beaker, the sediment was thoroughly agitated, and then allowed to sit for 30 seconds. Then the supernatant was screened through first a 0.5 mm screen and then a 0.15 mm screen.
• 4) The screened sediment was transferred to a 500 ml beaker and hydrofluoric acid was added. The sediment stayed in the hydrofluoric acid for 30 hours.
• 5) The sediment was processed through three distilled water washes.
• 6) After the washes had thoroughly cleaned the sediment of hydrofluoric acid, it was transferred to a 50 ml centrifuge tube and zinc chloride flotation mixture was added. (The mixture consists of a 10% solution of hydrochloric acid in which zinc chloride is dissolved until a specific gravity of 1.9 is obtained). The sediment was centrifuged at about 1,950 r.p.m. for 30 minutes.
• 7) From the upper portion of the sediment, polleniferous fluid was pipetted into a 50 ml beaker, and the fluid was diluted with distilled water.
• 8) A standard distilled water wash, followed by a glacial acetic acid wash, was performed.
• 9) Acetolysis mixture, consisting of one part sulphuric acid and 9 parts acetic anhydride, was added to the centrifuge tube containing the sediment and the tube was placed in boiling water for 20 minutes.
• 10) Standard acetic acid and distilled water washes were performed.
• 11) A final treatment of the sediments in 5% potassium hydroxide was done.
• 12) A series of standard water washes was were performed until the supernatant was clear.
• 13) The sediment was transferred to vial for study.

For statistical purposes, it is necessary to count a minimum of 200 pollen grains per soil sample. Most Williamsburg samples were found to contain too few to effectively make a 200 grain count. Nonetheless three microscope preparations were counted from each sample of extracted sediments. Usually, a 200 grain count can not be made form samples containing less than 1,100 pollen grains per gram of soil due to the tremendous time investment in isolating the identifying enough gains for statistical purposes as well as to erosional and degradational factors which destroy certain pollen types more easily than others. This results in pollen counts that are not representational of the original pollen rain deposited in the soils. Identification of grains was based on published keys and on comparison with the pollen collection on file in the Laboratory of Palynology, Texas A&M University.

Parasite analysis is relatively new as applied to archaeological sites. Techniques for the extraction of parasites (primarily the eggs of tapeworms, roundworms, and flukes) are varied. Some analysts have effectively applied clinical techniques to prehistoric soils and coprolites. When parasite eggs are partially decomposed or severely dessicated, I have found clinical techniques inadequate. It is necessary to dissolve carbonates in the soil that would otherwise cement the eggs in the soil matrix. It is also necessary, in certain types of soil, to dissolve the heavier silicates to allow effective flotation. My technique, which I now find applicable to both coprolites and archaeological soils, employs treatment of the soils in hydrochloric acid and, depending on silicate content, sometimes hydrofluoric acid. The soils are then reduced to the point at which the zinc chloride heavy density flotation fluid will effectively levitate even poorly preserved ova.

RESULTS

Of the 16 soil samples submitted for pollen study, only one contained sufficient pollen to make a 200 grain count. The samples submitted for analysis are listed in table I with the number of pollen grains present per gram of soil.

Table 1: Number of pollen grains contained per gram of soil as calculated by tracer spore to pollen grain ratios.

Sample #	Provenience	Number of grains per gram
28-G-
1037 III	Planting Bed I, 2nd level	40
1069 III	Planting Bed I, 4th level	124
1044 II	Planting Bed II, 2nd level	108
1075 I	Planting Bed II, 3rd level	38
1484 A	Planting Bed III, (refuse)	1,684
1596 R	Planting Bed III, (refuse)	42
1969, ABEF	Planting Bed IV, fill	155
2067	Planting Bed IV, (refuse)	29
1626	Control (subsoil between PB III&IV)	88
183 R	Dark brown soil, interior Str. A	66
1311	Control (subsoil within Str. A)	19
28-H-
334	Marl Layer	231
300B	Control (refuse layer)	82
692	Control (clean brown layer)	107
703	Ash layer	10
811	Control, (subsoil east of Str. K)	38

As can be seen, only sample 28-G-1484 A contained sufficient pollen for a complete 200 grain count. The pollen grains found in for all other samples are presented in Table 2. The pollen count from sample 28-G-1484 A is presented in Table 3. Common names of the plant taxa found in the pollen analysis are presented in Table 4.

Table 2: Pollen grains found in samples containing insufficient pollen for a 200 grain count.

Sample #	Pollen Grains Found
1037 III	1 Pinus, 1 Ligulafloreae
1069 III	1 Pinus 3 Ligulafloreae, 1 ChenoAm, 1 Betula
1044 II	5 Pinus, 5 ChenoAm
1075 I	1 Ligulafloreae, 1 ChenoAm, 2 Primulaceae, 3 low spine composite
1596 R	1 Pinus, 1 ChenoAm, 1 Betula, 2 low spine composite
1969 ABEF	1 low spine composite
2067	1 pinus, 1 Salix, 1 Ostrva/Carpinus
1626 P	2 Pinus
183 R	3 Pinus
1311	1 Pinus, 1 low spine composite
334	13 Pinus, 3 unidentifiable grains
300 B	2 Pinus, 3 ChenoAm, 1 Artemesia, 1 Carva, 1 low spine composite
703	1 Pinus
811	3 Pinus, 1 ChenoAm

Table 3: Pollen count from 28-G-1484 A.

Taxon	Grain #	Percentage of Total
Acer	4	1
Betula/Myrica	5	2
Carva	1	<1
Castanea	1	<1
Celtis	1	<1
ChenoAm	6	2
Diospyrus c.f.	1	<1
Fraxinus	2	1
Ilex	1	<1
Juglans	4	3
Juniperus	7	3
Ligulafloreae	3	1
Liquidambar	20	7
Penstamon c.f.	1	<1
Ostrva/Carpinus	8	3
Pinus	141	53
Platanus	1	<1
Portulaca	4	1
Quercus	11	4
Rosaseae	11	4
Umbellifereae	1	<1
Vibirnum	2	1
Viola c.f.	1	<1
Unidentifiable	19	7
Unknown	9	3
Total	265	100

Table 4: Common Names of taxa identified through pollen analysis.

Scientific Name	Common Name
Acer	Maple, probably Sugar Maple
Betula/Myrica	Birch or Sweet Gale
Carva	Pecan and/or Water Hickory
Castanea	Chestnut
Celtis	Hackberry
ChenoAm	Plants in the Goosefoot or Pigweed Family
Diospyros	Common Persimmon
Fraxinus	White Ash
Ilex	Holly
Juglans	Walnut
Juniperus	Cedar or Juniper
Ligulafloreae	Probably Dandelion or Chicory
Liquidambar	Sweet Gum
Penstemon	Probably Beard-Tongue
Ostrva/Carpinus	Hop-Hornbeam or Blue Beech
Pinus	Pine
Plantanus	Sycamore
Portulaca	Purslane
Quercus	Oak
Rosaceae	Rose Family (Apple, Chokecherry, Avens)
Umbellifereae	Carrot Family
Viburnum	Arrow-wood
Viola, C.F.	Violet
Unidentifiable	Grains too eroded, collapsed, or torn for recognition
Unknown	Grains well preserved but unknown to author and found neither in references nor in comparative collections
Salix	Willow
Artemisia	Wormwood, sage
Low Spine Composite	Wind pollenated plants in the Composite Family such as Grondsel and Ragweed

No evidence of parasite remains was found although all samples were tested.

DISCUSSION

REFERENCES

• Adams, K.R. (1980) Pollen, Parched Seeds, and Prehistory: A Pilot Investigation of Prehistoric Plant Remains from Salmon Ruin, a Chacoan Pueblo in Northwest Mexico. Eastern New Mexico University Contributions to Anthropology 9.
• Berlin, G.L., Ambler, J.R., Hevly R.H. and Schaber G.G. (1977) Identification of a Sinaqua agricultural field by aerial thermography, soil chemistry, pollen/plant analysis, and archaeology. American Antiquity 42:588-560.
• Bryant V.M., and Holloway, R.G. (1983) The role of Palynology in Archaeology, chapt. 5 in Advances in Archaeological Method and Theory, vol. 6, (M. B. Schiffer, ed.) Academic Press, Inc. New York.
• Dimbleby, G.W. (1957) Pollen analysis of terrestrial soils. New Phytologist 56:12-68.
• Dutton, J.P. (1979) Plants of Colonial Williamsburg, The Colonial Williamsburg Foundation, Williamsburg.
• Erdtman, G. (1960) The acetolysis method: a revised description. Svensk Botanisk Tidskrift 54:561-564.
• Faegri, K. and Iverson J. (1975) Textbook of Pollen Analysis. Hafner Press, New York.
• Fry, G.F. (1977) Analysis of Coprolites from Utah. University of Utah Anthropological Papers 97.
• Gish, J.W. (1981) Pollen Studies in the Palo Verde Archaeological Project Hohokam settlement at the Confluence. Museum of Northern Arizona Research Paper 20.
• Goldstein, F. (1968) Degradation of pollen phycomycetes. Ecology 41:543-545.
• Havinga, A. J. (1964) Investigations into the differential degradation of pollen and spores. Pollen et Spores 6:621-635.
• Hevly, R.H. (1981) Pollen production, transport, and preservation: potentials and limitations in archaeological palynology. Journal of Ethnobiology 1:39-54.
• Hill, J.N. and R.H. Hevly (1968) Pollen at Broken K Pueblo: some new interpretations. American Antiquity 33:200-210.
• Holloway, R.G. (1981) Preservation and experimental diagenesis of pollen exine. Ph.D. Dissertation, Department of Biology, Texas A&M University, College Station.
• Klilks, M. (1983) Paleoparasitology: on the origins and impact of human-helminth relationships. Chapt. 11 in Human Ecology and Infectious Diseases, Academic Press, New York.
• Martin, P.S. (1963) The Last 10,000 Years. University of Arizona Press, Tucson.
• Reinhard, K.J. (1985a) Helminth Examination of Coprolites: The Cultural-Ecology of PreHistoric Parasitism on the Colorado Plateau. M.A. Thesis, Northern Arizona University, Flagstaff.
• (1985b) Parasitism at Antelope House, a puebloan community in Canyon de Chelly. In Health and Disease in the Prehistoric Southwest (Merbs, C., ed.), Arizona State University Anthropological Research Papers 39.
• Reinhard, K.J., J.R. Ambler, and McGuffie, M. (1985a) Diet and parasitism of Dust Devil Cave. American Antiquity, in press.
• Reinhard, K.J., Mrozowski S., and Orloski K. (1985b) Pollen and Parasites: the biological nexus in historical archaeology. Paper presented at the annual meetings of the Society for Historical Archaeology.
• Tschundy, R.H. (1969) Relationship of palynomorphs and sedimentation. In Aspects of Palynology, pp. 79-96, (Tschuny, R.H. and Scott, R.A. eds.). Wiley Press, New York.
• Williams-Dean, G. and Vaughn M. Bryant (1975) Pollen analysis of human coprolites from Antelope House. The Kiva 41:97-112.
• Wycoff, D.G. (1977) Secondary forest succession following abandonment of Mesa Verde. The Kiva 42:215-232.

A REPORT OF THE ANALYSIS OF MACROFOSSIL REMAINS
FROM THE PEYTON RANDOLPH HOUSE,
WILLIAMSBURG, VIRGINIA

By
Stephen A. Mrozowski
Archaeobotanical Consultant

INTRODUCTION

The analysis of seeds and other plant macrofossils has been a major part of archaeological research since its beginnings as a discipline during the nineteenth century (Kunth 1826; Heer 1866). Since that time, archaeologists seeking answers to questions concerning prehistoric plant use and the origins of agriculture have relied upon the microscopic remains of plants for evidence (e.g., Dimbleby 1978). Within the past two decades two very important trends have emerged in the study of plant remains in archaeology. The first of these has been the application of these data to reconstruct past landscapes (e.g., Jones and Dimbleby 1981). The second has been the emergence of contextual studies, the main goal of which has been to link intra-site patterning to human behavior (e.g., Dennell 1972, 1976; Hubbard 1976; Hillman 1973, 1981). The "behavioral approach" as Jones (1985:108-111) has recently noted is now the main focus of most archaeobotanists' work in Europe.

While the importance of incorporating plant studies into, archaeological research has a long tradition among pre-historians and ethnobotanists in North America (see Ford 1978), the same is not true for historical archaeologists. Perhaps because of the emphasis placed 2 upon the study of the documentary record, historical archaeologists have ignored many of the analytical avenues traveled by pre-historians. This may prove unfortunate because recent studies suggest the potential this approach holds for providing information concerning the reconstruction of past landscapes and diet (Kelso and Schoss 1983; Mrozowski 1983).

A major stumbling block to the analysis of plant remains from archaeological contexts is the problem of differential preservation (e.g., Green 1982; Minnis 1981; Keepax 1977). These difficulties include differentiating between seeds deposited in antiquity as the result of human activity and those that are the product of post-depositional disturbance. One answer to this dilemma is to accept only charred seeds as being culturally relevant (Minnis 1981; Keepax 1977). While this may be appropriate in some cases, there are archaeological contexts that are characterized by aerobic environments that promote organic preservation. These include features such as wells, drains and privy pits. Seeds recovered from Roman and Viking sites in York (Hall et al. 1980; Hall and Kenward 1982) have shown remarkable preservation.

Urban sites, like those in York, often contain archaeological features such as wells and privies that can be counted upon for well preserved organic remains. The predominance of features like privies and drains on urban 3 sites reflects the need for such facilities brought on by the spatial constraints of city life. One unfortunate by-product of this situation is that urban sites are given priority by analysts because of their productivity. What is unfortunate is that rural sites and non-water-logged deposits in cities are often ignored. This state of affairs is unlikely to change without some experimenting with contexts such as those just noted. The Peyton Randolph House provides just such a testing ground.

The main focus of our research at the Peyton Randolph House was to reconstruct the site landscape. This involved reconstructing the plant communities that characterized the yard during the eighteenth century, as well as determining function of four features that may have served as planting beds. This report provides a tabulation of the seeds identified from the four planting beds in the yard and an analysis of those findings. As the results indicate, the value of this kind of research on sites like the Peyton Randolph House reflects the overall integrity of the site itself. As the findings suggest, the site appears to be heavily disturbed although some interesting information was obtained from the material recovered form the planting beds. While evidence of wide-scale disturbance was quite clear, data pertaining to the use of yard space and possible dietary information was recovered as a result of this analysis.

4

METHODS

Soil samples were collected by the archaeologists from features located in the yard. These were matched by a like number of control samples. These samples were then processed at Colonial Williamsburg by flotation (see other section). The floated samples were then scanned for non-seed material such as roots and nitrogen nodules. An initial sorting was then carried out at Colonial Williamsburg. Once this was completed samples of sorted specimens were provided the author. Each specimen was subjected to microscopic analysis to determine the identity of seeds at the family level. This initial phase was followed by a second round of analysis in hopes that species identification could be obtained. These identifications were compared with seed specifications provided by manuals and type collection samples. In some instances it was possible to identify seeds to the species level. All seeds were quantified precisely with the exception of the genus Rubus, which is a member of the Rose family and includes blackberry and northern Dewberry. Representatives of this genus are often recovered in large numbers from archaeological contexts. In this instance approximate numbers are provided along with other information in Table 2.

5

RESULTS

This discussion will focus on archaeological contexts as opposed to individual samples. Table 1 contains a precise tabulation of all the seeds identified from each sample. By concentrating on archaeological contexts as opposed to discussing individual samples emphasis will be placed upon the interpretation of the floral material as it pertains to the overall site. It is hoped that this will be more fruitful than a discussion of the plants alone.

Planting Beds

The vast majority of seeds that were recovered from the Peyton Randolph House were from the four features alleged to be planting beds. These features were found at the rear of the house lot in close proximity to one another. Instead of discussing individual samples from the planting beds. All samples from each bed will be dealt with as a feature assemblage. Furthermore, the four beds will be considered as a group due to the noticeable similarity between them in terms of what seeds they contained.

Discussion

Overall, the plants represented in most of the archaeological contexts at the Peyton Randolph House are familiar to those found growing in the area today. This should not come as a surprise, however, given that plant communities have probably changed very little in the last two hundred years. Most of the information that can be gleaned from these data relate to the general setting of the yard. Based upon the presence of both Oxalis (sorrel) and Physalis (groundcherry) it wold appear that the yard was relatively well shaded. The idea that the planting beds were used for asparagus or other legunainous species is supported indirectly by a community of plant remains that grow well in nitrogen rich soil and in shade, conditions both suitable for asparagus. This, coupled 13 with what may be asparagus seeds, provides at least some support for this possibility.

The presence of Cucurbit remains in contexts that may have been associated with food preparation or disposal indicates that members of this family were being consumed by the site's inhabitants. Their presence in the planting beds suggests they could have been cultivated in the beds or that they represent refuse used as a mulch or compost. The latter of these two possibilities is the most likely given the size of the beds and the way Cucurbits grow. Members of this family send out long runners that can quickly dominate a garden plot. Furthermore, nothing so elaborate as the planting beds is necessary for growing plants like squash or pumpkin. Asparagus, on the other hand, would require a well controlled, moist, rich environment in which to flourish. Randolph (1826:3-5) suggests the use of dung or other waste products in the propagation of asparagus so the presence of some plant material should be no surprise. In fact, many of the weed seeds recovered from the planting beds represent plants that would have been a nuisance to gardening, but that could have been added to compost used as fertilizer. Ethnoarchaeological research recently conducted in Massachusetts indicates that garden weeds such as Pokeweed, Chenopodium or Rumex (Dock) as well as many 14 creepers (Rubus and Physalis) are removed only to be redeposited as fertilizer.

If we compare the seed assemblages of the four planting beds it is possible to discern some differences. The overall breakdown is presented below in Table 1. With the exception of the Cucurbit and the possible Asparagus all of the seeds found in the planting beds represent weeds. There are, however, some subtle differences between beds 1 and 2 and beds 3 and 4 that may be important. One distinction is the lack of Cucurbit in either bed 3 or 4. The other difference is the presence of Trifolium and Taxacum in beds 3 and 4. Both clover and dandelion are grass plants that would not be anticipated in gardens under cultivation. This leads to the inference that the assemblages of beds 3 and 4 represent post-cultivation communities while those in 1 and 2 may reflect at least some of the plants that were present while the beds were being used. This is a tenuous assertion that must be approached with caution. Oxalis stricta, for example, is a plant that would be anticipated in cultivated soil. What is most likely is that beds 3 and 4 have been more disturbed than either l or 2, and that some original community members are represented.

Other contexts within the site provide further information, but like the planting beds, they too present difficulties concerning their interpretation. The four 15

Table 1.

PLANTING BED 1*	PLANTING BED 2*	PLANTING BED 3	PLANTING BED 4
LEVEL 1	LEVEL 1	Rubus	Trifolium (clover)
Rubus	Possible Asparagus	Pokeweed	Rubus
Chenopodium	Acalapha	Taxacum (dandelion)	Polygonaceae
Brassica		Hackberry	Oxalis
		Physalis	Pokeweed
LEVEL 2	LEVEL 2	Rosaceae	Acalepha
Cucurbit	Cucurbit		Physalis
	Rubus		Possible Asparagus
LEVEL 3	Possible Asparagus
Pokeweed
Cucurbit	LEVEL 3
Rubus	Pokeweed
	Acalapha
LEVEL 4	Tulip tree
Acalapha

16 Prunus seed fragments from the black ash around a hearth certainly represent food (cherries) eaten or prepared in close proximity to the feature. The same can be said, with some assurance, for the Sambucus (Elderberry) and Cucurbit seeds from the refuse layer overlying the yellow marl. Elderberry bushes take a number of years to produce berries and their presence on a site, unlike Rubus, usually can be attributed to cultural activity. What use the Elderbery was put to is unknown, but it could have served as a food, drink or food preservative.

The only additional family of note that is represented in the site assemblage is the Caryphyllaceae found in the brick bat and shell filled ditch as well as in the layer of yellow clay. The Caryphyllaceae family known commonly as "Pinks" is a pleasant flowering runner that is often promoting along fence lines, walls or the sides of houses.

On the basis of the information as a whole it would seem that the four features within the yard did function as planting beds. Compost was used to fertilize the beds that contained domestic refuse and weed remains probably collected in the yard. Answers to the, question what was grown in the beds remain elusive. The large, asparagus-like seeds may indicate asparagus was cultivated in at least two of the beds. If not asparagus, then we can at least assume some kind of large legume, a bean perhaps. 17 The presence of Oxalis suggests a shaded area, which would be conducive for asparagus. No definitive answer can be offered unfortunately.

CONCLUSION

It is difficult to draw any sound conclusions from the seeds recovered from the Peyton Randolph House because it would appear that the site is extensively disturbed. This contention is supported by both the archaeology and the plant remains found on the site. The presence of Rubus seeds in almost every context is testimony to the activity of humans in shaping the site. Plants are among the most sensitive indicators of human activity. The challenge is discerning what the activity was and when it took place. This is particularly true of the planting beds. As much as Randolph (1826:3-5) exhorts his readers to rid their beds of weeds, this is virtually impossible. Therefore the presence of active colonizers such as Chenopodium, Rubus, Oxalis and Physalis does not necessarily indicate post cultivation succession. Rumex is often found in non-active areas while plants such as Chenopodium are, more often found in soil under cultivation. However, beyond this, little is known about what weeds were present in eighteenth century gardens.

The results of the analysis have been somewhat 18 disappointing in that the majority of seeds are from what appear to be mixed contexts. The seeds suggest this quite strongly. The mix of uncharred weed seeds with those of charred legumes in the planting beds may indicate successional shifts or disturbance. The conditions indicated by the plant remains are not unlike those present on the site today with the possible exception of the Tuliptree seed. This tree is quite common, but whether its presence represents recent disturbance or is more indicative of past conditions is difficult to determine.

One recommendation that arises from this study is that this kind of analysis might be best applied to contexts in which good organic preservation is anticipated. In York, Allen Hall, the botanist for the University of York Environmental Archaeology Unit, informed me that he was limiting his efforts to waterlogged deposits only. This realization is probably applicable to Colonial Williamsburg as well in the future, physical analysis should be contingent upon a set of criteria that would place priority upon those contexts such as trash deposits, wells, privies or any deposit in which organic preservation will be promoted. Experiments like this are necessary, however, so that we can set a course for research and analysis in the future.

19

Table 2. Quantitative Breakdown of Seeds From Peyton Randolph House.

PROV	DESCRIPTION	FAMILY/SPECIES	# FRAGMENTS
28H 780	Feature, black ash around N. hearth	Prunus	4
28H 551	Layer/Yellow clay	Caryophyllaceae	2
		Gramineae	3
28H 393A	Shell and brick bat ditch	Caryphyllaceae	3
		Rubus	1
28H 313B	Layer refuse layer over marl	Sambucus	1
28H 300C	Layer refuse layer over marl	Wood fragment Charred curbuibita	1
28H 119	Feature: Large dark circular pit in cellar
28G 2096	Feature: Dark brown 7	Compositeae ?	1
		Physalis, ground cherry	1
		Charred: asparagus	1
		Rubus	100+
28G 2070	Feature: Planting bed 4	Trifolium; clover	1
		Rubus	6
		Unidentified charred	2
28G 1971 N-P	Feature: Planting bed 4	Charred legume	1
		Poss. asparagus	1
		Rubus	2
		Trifolium; clover	1
28G 2032 E7	Feature: Planting bed 4	Rubus	20+
		Rumex	1
28G 2035	Feature: Planting bed 4 extension	Rubus	5
		Trifolium (clover)	1
28G 2036	Planting bed 4; subunit north end extension	Rubus	3
28G 2065
28G 2067	Subunit, planting bed 4	Oxalis stricta	2
21
28G 826 II	First level of planting bed 1	Charred No. I.D.
28G 826 I	First level of planting bed 1	Exterior fragments
28G 927 II	First level of planting bed 2	Possible asparagus or legume	1
28G 987 III	First level of planting bed 2	Charred Acalypha virginica	1
		Insect part	1
28G 989 I	first level of planting bed 2	Charred Frag. No I.D.	1
28G 989 II	First level of planting bed 2	Two strange ones that may not be seeds	2
28G 1037 II	Second level of planting bed 1	Cucurbit	1
28G 1037 III	Second level of planting bed 2	Possible watermelon No I.D. Crude seed	1
28G 1037 IV	Second level of planting bed 2	Wood frag. and charred No. I.D.	2
28G 1038 I	Second level of planting bed 2	Charred No I.D.	1
		Rubus	1
28G 1038 II	Second level of planting bed 2	Charred No I.D.	1
		Rubus	1
28G 1042 I	Second level of planting bed 2	Charred large seed asparagus size	1
28G 1042 II	Second level of planting bed 2	Rubus	2
28G 1043 I	Second level of planting bed 2	Charred No I.D.	1
28G 1043 II	Second level of planting bed 2	Charred - No I.D. Seed ?	1
28G 1050 I	Third level of planting bed 1	Pokeweed	1
22
28G 1051 I	Third level of planting bed 1	Insect part	1
		Curburbit	1
28G 1051 III	Third level of planting bed 1	Rubus	2
28G 1051 IV	Third level of planting bed 1	Charred achene	1
28G 1052 I	Third level of planting bed 1	Rubus	10
		Charred achene	1
28G 1052 II	Third level of planting bed 1	Rubus	8
		Poss. Insect Part	1
28G 1069 I	Fourth level of planting bed 1	Rubus	1
		Copper leaf	1
28G 1069 III	No Data	Hackberry Shell frag.	1
		No I.D.	1
28G 1069 IV	Fourth level of planting bed 1	Acalypha virginca	1
		Acalpha virginca	1
28G 1074 I	Third level of planting bed 2	Lirodendrom tulipiton (tulip tree)	1
28G 1074 II	Third level of planting bed 2	Pokeweed	1
28 1075 II	Third level of planting bed 2
28G 1075 III	Third level of planting bed 2	Acalypha	1
28G 1212 N	Plowzone	Rubus	1
		Physalis	1
		Oxalis	1
28G 1420	Builder's trench, exterior east wall of structure E	Rubus	1
		Physalis	1
28G 1468 F	Light brown layer with brick & mortar frag.	Charred Physalis or Salanum	1
23
28G 1480 D	Planting bed 3	Charred legume	1
28G 1482 C	Planting bed 3	Charred seed	1
28G 1482 F	Planting bed 3	Oxalis	1
28G 1483 C	Planting bed 3	Rubus	2
28G 1483 H	Planting bed 3	Rubus	1
		Oxalis	1
28G 1483 R	Planting bed 3	Rubus	1
		Charred achene (nutlet)	1
28G 1484 A	Planting bed 3	Charred achene (nutlet)	1
28G 1484 N	Planting bed 3	No I.D.	1
28G 1484 R	Planting bed 3	Pokeberry	1
28G 1533 G	Layer planting bed 3	Rubus	3
28G 1533 H	Layer planting bed 3	Taxacum (Dandelion)	1
28G 1533 L	Layer planting bed 3
28G 1534 E	Layer planting bed 3	Rubus	1
		Oxalis	1
28G 1534 P	Layer planting bed 3	Hackberry	1
		Pokeberry	1
28G 1534 R	Layer planting bed 3
28G 1584 E	Planting bed 3	Pokeberry No I.D.	1
28G 1584 R	Planting bed 3	Charred nut	1
28G 1544 N	Planting bed 3	Physalis	1
		Rosaceae	1
28G 1619 F	Layer of marl in exterior of Structure E	Rubus	1
28G 1752	Layer: brick, plaster and mortar rubble	Rubus	200+
24
28G 1795	Robber's trench extended earthwork from 1785 pit features	Rubus	100+
28G 1846	Interior Structure B soft dark brown	Rubus	300+
28G 1862	Feature	Rubus	250+

25

REFERENCES CITED

Betts, E. (editor) -

1981

Thomas Jefferson's Garden Book. American Philosophical Society, Philadelphia.

Dennell, R. W. -

1972

The Interpretation of Plant Remains: Bulgaria. In Papers in Economic Prehistory, edited by E. S. Higgs, pp. 149-159. Cambridge University Press, Cambridge.

1976

The Economic Importance of Plant Resources Represented on Archaeological Sites. Journal of Archaeological Science 3:229-247.

Dimbleby, G. -

1978

Plants and Archaeology. Humanities Press, Atlantic Highlands.

Fernald, M. L. -

1970

Gray's Manual of Botany, eighth edition, D. Van Nostrand, New York.

Ford, R. I. -

1978

Ethnobotany: Historical Diversity and Synthesis. In The Nature and Status of Ethnobotany, edited by R. I. Ford, pp. 33-49. Anthropological Papers of the Museum of Anthropology, University of Michigan, No. 67, Ann Arbor.

Green, F. J. -

1982

Problems of Interpreting Differentially Preserved Plant Remains From Excavations of Medieval Urban Sites. In Environmental Archaeology in the Urban Context, edited by A. R. Hall and H. K Konward, pp. 40-46. Council for British Archaeology Research Report No. 43, London.

Hall, A. R., H. K. Kenword and D. Williams -

1980

Environmental Evidence From Roman Deposits in Skeldergate. The Archaeology of York 14/3. Council for British Archaeology, London.

Hall, A. R. and H.K. Kenward (editors) -

1982

Environmental Archaeology in the Urban Context. Council for British Archaeology Research Report No. 43, London. 26

Hillman, G. -

1973

Crop Husbandry and Food Product: A Modern Basis for the Interpretation of Plant Remains. Anatolian Studies 23:241-244.

1981

Reconstructing Crop Husbandry Practices From Charred Remains of Crops. In Farming Practice in British Prehistory, edited by R. Mercer, pp. 123162. Edinburgh University Press, Edinburgh.

Hubbard, R. N. L. -

1976

On the Strength of the Evidence for Prehistoric Crop Processing Activity. Journal of Archaeological Science 3:257-265.

Jones, M. -

1985

Archaeobotany Beyond Subsistence Reconstruction. In Beyond Domestication in Prehistoric Europe, edited by G. Berker and C. Gamble, pp. 107-128. Academic Press, New York.

Jones, M. and G. Dimbleby (editors) -

1981

The Environment of Man: The Iron Age to the Anglo-Saxon Period. B.A.R. British Series #87, British Archaeological Research, Oxford.

Keepax, C. -

1977

Contamination of Archaeological Deposits by Seeds of Modern Origin with Particular Reference to the Use of Flotation Machines. Journal of Archaeological Science 4:221-229.

Kelso, G. and J. Schoss -

1983

Exploratory Pollen Analysis of the Bostonian Hotel Site Sediments. In Archaeology of the Bostonian Hotel Site, Massachusetts Historical Commission Occasional Papers in Archaeology and History #2, pp. 67-76, Boston.

Kunth, C. -

1826

Examine Botanique. In Catalogue Raisonne et Historique des Antiquities Decourvertes en Egypte. J. Passalacque, Paris.

Minnis, P. -

1981

Seeds in Archaeological Sites: Sources and Some Interpretive Problems. American Antiquity 46(l): 143-151. 27

Mrozowski, S. -

1983

Examining the Urban Environment Through the Analysis of Hord Remains. The Newsletter of the Conference on New England Archaeology 3(2):46-52.

Informant Descriptions

Lot 4: Mrs. Angle Cowles, Apt. 1, January 1939 - February 1956

Mrs. C. When did I move into the Donegan house? Miss Castle was living across the street in the Donegan [house] and she was going to leave. That's when I moved over there. I had the side of the house down next to the… windmill. It was on the end not closest to the Peyton Randolph [house].

I remember one great big [tree], not very far from the edge of the tunnel. The tunnel that goes under Williamsburg. That was built while I was [living in the Donegan house].

It was a nice house, a nice apartment to bring up a child because it had a nice yard. The boys could play football out on the side of the house. They had a real good time playing out there. They had this lot that was next to it where they were excavating for the windmill at the time [1953].

It was a large, large house. The rooms were large and the ceilings were high. It had a nice living room, nice big living room. And no dining room. And a large kitchen. Then a back porch to each apartment that was screened in. It had two bedrooms upstairs. A bath, and one large bedroom on the third floor. That had dormer windows. You can imagine the looks of the house because there were two sides just like that. Each side was exactly alike.

Ms. F. Today would that third floor be considered an attic? Was there anything above the third floor?

Mrs. C. No. Nothing above it. But it had those little closets underneath those dormers. You had to bend over double to get into them. 'Course I remember getting all of my son's toys and things out of one of those closets.

[There was a fireplace] in the living room.

Ms. F. You had bats come down the chimney?

Mrs. C. Uhuh… I killed 40 bats in my house. Then they send down and gassed the chimney to be sure to get all the bats out. Then they put wire over it so they couldn't come in… C.W. [put wire over the chimney]. Ohh. But those bats almost drove us crazy for a little while there … It was in the summertime. Yes. I remember.

I had a broom that I kept handy. I had something to cover my hair [also]. You've always heard about them, getting into your hair. So, I had something to put over my hair. If I'd hear a bat I'd go.

The furnace was in the basement. It was a coal furnace.

[The garage] didn't have any doors. It had a little storage area to one side. That part was enclosed. It was just room for two cars. [The storage area] was made onto the garage. They couldn't have put that on the side when we were living in the Historic Area. So, you see, that was very nice. When the Tuckers moved in, Mr. Tucker put the [basketball] hoop on a pole.

Ms. F. Did you enjoy living in that house?

Mrs. C. Yes. I did. It was on the fringes of the restoration and yet we could do things there that you couldn't have done [if you lived in the Historic Area]. For instance, the boys couldn't have played football in the backyard of any of those other houses. They used to go up and play on the Courthouse Green.

I can't remember what all was [in the yard]. But I remember when we all had to have Victory Gardens [1942]. We had a little one out there. It was in the duplex yard back there where the windmill is. We just had it that one year … . It was back from the street, not quite at the edge of the tunnel. It was when the war was going on and everybody was asked to have a garden. We just had a small garden. Lettuce… .

The front yard had a walk down the center. I guess it must have been brick, but I don't remember. I don't think the front yard was ever used very much. I don't think it had shrubbery, which made it a bare sort of looking place. We had flowers round the back. I used to have a bed of flowers in front of the screened porch.

Ms. F. Who else lived in the duplex?

Mrs. C. Mr. Campbell, who worked in the Architectural Department. They were there when we moved in. They had a son. The Pilands lived there for a while. Raymond Piland. They had a son. Several people lived there while I did. Ohh. The Tuckers. The father was Pete Tucker [who] worked for Colonial Williamsburg. And they had a son who was Pete also. Mr. [Armond] Campbell was in charge of the Colonial Williamsburg Laundry. They lived there. There were a good many families who lived on the other side of the duplex.

Ms. F. You mentioned the Campbell who was the architect, Raymond Piland, and Pete Tucker. Do you remember any other names?

Mrs. C. The Langtons. L - A - N - G - T - 0 - N. In the Accounting Department. That just about does it… Mr. Langton died there.

Ms. F. Really. He died in the house?

Mrs. C. Mhm.

Ms. F. Did you know the Donegans.?

Mrs. C. No. I didn't have any idea who they were.

The Millers were still living where they were [Lot 5 and 6]. I think the Skillmans still lived there [Lot 7]. They owned the house back of the Peyton Randolph [house] and the Millers. The Millers had a little cottage out in their years. [They] were mighty good neighbors. We enjoyed them. So were the Skillmans. We didn't see too much of them. You know, in those days you just didn't have much time.

Ms. F. Do you remember what year you moved out of the house?

Mrs. C. I was told while I was living at the Donegan house that it would be a good idea to keep my eyes open because they were going to build the windmill on that spot. When the Greenhow Kitchen became vacant [I moved there]. That's behind the spinning and weaving shop [former]. I lived there for the last 11 years [before retiring] on the second day of January 19 and 67.

Ms. F. So, you lived in the duplex from sometime in the late 1930s [1938] to 1956.

Mrs. C. The Donegan house was taken down. Someone bought it I think. It was my understanding they bought it, marked the pieces and built it somewhere else … . But I don't know. It was a nicely planned house.

Ms. F. I thought I had read something about it being demolished…

Mrs. C. Well, maybe it was. I just vaguely remember somebody saying that someone bought it and moved it. Of course, they had to tear it down to move it.

Lot 4: Mr. Armond D. Campbell, Apt. B, May 1955 - August 1956

Mr. C. May 1955 to August 1956

Ms. F. How did you come to be in the Donegan house?

Mr. C. I left Williamsburg in '43 and came back in '44, out of the Service. When I came back John Gree, who was Vice President and General Manager for the hotels and restaurants, put me in the little cottage on Tazewell Street. I believe they sold that property to Dr. Clarence Hall. I had to get out of there and they let me have a part of the Donegan house (apartment].

The biggest thing that I can remember about that house, well, I didn't have any maintenance on the yard. It was kept up by C. W. When you walked up to the front entrance it looked to me like they had taken the house and cut it right smack in two and partitioned it off. There was a door on the right hand side [after you went through the front entrance] and a door on the left hand side.

As you go in the door there was a living room with great big tall windows and a fireplace in the downstairs. The next room back was the kitchen. And then in the summertime the whole length of the kitchen area on the back was a porch. A nice screened-in porch.

I can't remember whether there [were] two or three bedrooms. There must have been three bedrooms. There [were] two on the second floor and one on the third floor. At that time I had three children. I had two girls, and a boy. The oldest girl's name was Joan. The other girl was named Barbara and the boy's name was Armond. There were five of us.

I hated like heck to move from there. Rent was cheap. I had no maintenance. As near as I remember, I only paid $35.00 a month for that apartment. I remember getting along with Angie Cowles … . It sticks in my mind that we had a central heating system and a sharing plan with the oil and all. We enjoyed it there. We were restricted to some degree. Your activities, what you could do in the Restored Area, but it didn't bother us.

Other than Mrs. Cowles we never knew or associated with anybody else over these that I can remember. At that time, in 1955, I was 40 years old. We were playing hard. We were fishing, hunting, all that in our spare time. The children played around the neighborhood. There was a family directly across the street that had some children. I can't remember who they were. I remember the children playing around there. I had no trouble with mine getting down on the Park Service.

Ms. F. What do you remember about your yard?

Mr. C. It was flat and green and well-kept. I remember these two sheds [in photo] out here behind the house. [One] could be used for a garage. At that time I didn't have an automobile. I can't remember what exactly was in there.

Now that you mention it, [Mrs. Cowles's son] was away at college. There wasn't anyone there but Angie. My son was born in '43. He would have been 12 years old. That brings back old memories … At that time, as I say, I was very active fishin', and huntin'. The York River was full of fish and the woods were full of game. My cousin and I salted a great big ol' beef hide. I can't even remember whatever happened with that thing. But we were just foolin' with that thing. We had a ton of salt on it, I know that!

Ms. F. When you were living in the Donegan house was the garbage taken away or did you dump it someplace?

Mr. C. No. The garbage was taken away. It was picked up by the town. I remember putting the garbage out front, the cans out front.

Ms. F. Do you remember when the house was taken down?

Mr. C. Well, I should. Really, I can't remember much about it. I just lost all interest in the site when I left there. It wasn't on my way home, back and forth to work.

Ms. F. Is there anything I haven't asked that you remember about the neighborhood?

Mr. C. No. You see, we had our own little crowd of people that we went around with. And actually, due to the fact that I lived in the Restored Area, most of our activities were out of the area. Because of the confinement. but, we enjoyed very much living there.

Ms. F. Do you remember what some of those restrictions were? Were they written down?

Mr. C. No. They were just taken for granted. You were just on your good behavior when you were in the Restored Area. I guess with original and life time tenants, it's a little different. But for an outsider to come in, like I was, you pretty well played according to their rules. Just plainly understood.

Lot 4: Mr. Frank Jacobs

Mr. J. From there we went to Miss Van Garrett's bungalow (behind the Archibald Blair House]. Approximately 1931, for about a year. It was a one story, three room bungalow with a little small porch on front with a board fence across the front.

The person I can remember who lived in the Donegan house while I was living [in the bungalow], was "Bennie" Meisel, who was the Town Manager, City Manager. Bennie, B-E-N-N-I-E Meisel. I always called him Bennie Meisel. [A. L. was in the records]. I can only remember him having one daughter and she was killed in an automobile wreck coming back from Richmond one time. . I think I had moved from the area. I was living on South England Street when that happened.

Mr. Pete Tucker lived there one time, same time Angie Cowles lived there. Floyd Tucker. We called him Pete. The Langton boys, I used to pal around with those, two of them. But at that time I was living over on Capitol Landing Road … . We went to school together. It was Charles and Jack Langton. And there father, Mr. Langton, worked in the Accounting Department, I believe. He died of a heart attack and then they went back to New York where they came from.

Mr. Minor Wine Thomas

Okay. I knew George Small, slightly. He was a professor at the college. I knew of Mr. Meisel, city manager. I knew G. T. Brooks, although he didn't live in (the Donegan house) at the time. I knew him later years after he'd moved out of that house. Oh. Wanda Castle and Miss Croxton. Wanda became personnel manager of CW's secretaries. I haven't thought of them for years … When I was a kid my grandfather ran a dairy farm, and in the summertime I used to take over the milk delivery route. So, I knew the back door of practically every house in Williamsburg. I remember Miss Castle and Miss Croxton simply as people I delivered milk to. Angie Cowles and her son, I knew.

Oh. I see, this is the other half of that house (timeline) … Wash Reed I knew. He was with the Architectural Department with CW … George Campbell was a CW architect I knew slightly. Raymond Piland, I remember simply as a name I delivered milk to. Armond Campbell was the fellow who used to run the CW Laundry.

CHAPTER 4

In Chapter 4 the informants discuss what they recall about Jimmy Knight and his crews during excavation activities around their properties. Lots Four, Five, Six and Seven were excavated between 1955 and 1956 (Lot Four after August 1955). Mr. Moreau Chambers then collated an artifact analysis from the material saved during those excavations (analysis dated April 1956). In 1938, the east wing of the Peyton Randolph house was uncovered in excavations. At the same time, part of the backyard was cross-trenched. Excavations of Lot 10 were completed in 1940 during which time the house was being renovated. In 1978, part of Structure A on Lot Five was excavated, and a larger strip on that property was dug from July to December of 1982. Work has continued southward onto Lots Six, Seven and Ten since then.

-32-

Peyton Randolph House, Phase I

The cross dating of the Peyton Randolph House, Phase I, is based on tulip-poplar samples removed from the ceiling of the second floor. The tulip-poplar trees used to construct the joists, sleepers and a valley plate had completed the growing season of 1715. The fresh condition of the wood and lack of evidence of bark-infesting insects suggest the tulip-poplar trees were felled, hewed, sawn and put into the house frame within a short period of time. Thus the ceiling of the second floor is believed to have been constructed during the spring/summer of 1716.

Roof Structure: The unusual structure of the roof consists of three parallel frames with the valleys above two oak gutters. The two oak gutters were hewed from rapidly growing red oaks; thus the number of annual rings provided a pattern of only 54 years. The two oak joists were combined with the gutters to form a KY pattern 38 years having 6 +KY and 8 -KY (P = .13., n = 7) (Fig. 18). The combined oak KY pattern from the Peyton Randolph House, Phase I, was cross dated against the following combinations of area KY patterns:

Key Year Patterns	Year of Best Fit
	1	2	3
Oaks	1718	1948	1841
Pine (Geddy)	1773	1690	1711
Oak/pine	1718	1948	1690
Oak/pine/tulip-poplar (PRI)	1948	1841	1911
Tulip-poplar (PRI)	1736	1707	1651
Tulip-poplar (PRII)	1824	1776	1842

-33-

The varying years of best fit suggest that the oak KY pattern from the Peyton Randolph House is extremely limited both in sample size and in length of pattern.

The last year of tree growth for the oak gutters cannot be established with any degree of certainty using the currently available tree-ring data.

Shingles: The shingles used to cover the W-shaped roof structure are cypress (Taxodium spp.). None of the shingles had edge; thus the annual rings were not measured for this study. However, of importance is the amount of surface loss on the exposed butt ends, the maximum amount being one-eighth inch on each of four shingles. The "rule of thumb" for weathering (non-abrasive) of wood is one-fourth inch per 100 years. Accounting for the extreme durability of cypress to weathering, the amount of surface loss suggests the shingles were exposed to the weather for a considerable period of time.

Rafters: The rafters (R² ) forming the present roof of the Peyton Randolph House, Phase 1, are believed to have been added (spliced above plate [P]) to the rafters (R¹ ) forming the original W-shaped roof.

The wood samples used in this study (n = 5) were taken from what are believed to be the original rafters (R¹ ). There were three wood samples removed form the later addition (R² ) to form the present roof; however, two were of tulip-poplar heartwood and thus provided no data regarding the closing of the roof structure. The third wood samples was of tulip-poplar, had a bark edge and was cross dated with the tulip-poplar rafters in the Peyton Randolph House. Phase II.

Thus the closing of the W-shaped roof of Phase I may have occurred during the construction of Phase II of the Peyton Randolph House.

-34-

The original tulip-poplar rafters (R1) from the Peyton Randolph House, Phase I, provided five suitable samples for a key-year pattern (Fig. 19). This KY pattern could not be cross dated with the tulip-poplar KY pattern of the joists/sleepers/plate (Fig . 20), suggesting the trees were not felled in the same year. Further evidence the trees were not felled in the same year is suggested by the drastic reduction in the number of KY years when the two patterns are combined at the 0 year (Fig. 21).

The KY pattern for the rafters (R¹ ) from the Peyton Randolph House were then cross dated against the following combinations of area KY patterns:

Area Key-Year Patterns	Year of Best Fit
	1	2	3
Oak	1716	1703	1707
Pine (Geddy)	1744	1758	1716
Oak/pine	1707	1763	1977
Tulip-poplar (PRI)	1725	1656	1703
Tulip-poplar (PRII)	1819	1789	1821
Oak/pine/tulip-poplar
1982-1607	1707	1963	1977
1900-1800	1882	1819	1823
1800-1700	1707	1703	1716

When the tulip-poplar KY pattern for the rafters (R¹ ) of the Peyton Randolph House, Phase I, was cross dated against the various combinations of area KY patterns, the most significant year of best fit for the last year of tree growth was 1716.

Area Key-Year Patterns	Year of Best Fit	Chi-Square Value	Kappa Value
Oak	1716	8.0	1.0
Pine	1744	5.0	1.0
Oak/Pine	1707	7.0	1.0
Tulip-poplar (PRI)	1725	5.0	1.0
Tulip-poplar (PRII)	1819	3.9	.72
Oak/pine/tulip-poplar
1982-1607	1707	8.0	1.0
1900-1800	1882	5.0	1.0
1800-1700	1707	8.0	1.0

-35-

Thus the key-year patterns for the structural members for the ceiling and roof of the Peyton Randolph House, Phase I, were combined at the year of best fit with the area oak KY pattern as follows: Also, see Fig. 22.

I)	Joists/sleepers/plate:	1715
II)	Rafters (R1):	1716
III)	Gutters/joists:	1718

The combined tulip-poplar/oak KY pattern was then cross dated with the area oak/pine KY pattern. The year of best fit was 1718 (P = .05, n = 14) (Fig. 23).

Possible Construction Years: The construction of the second floor and the roof of the Peyton Randolph House, Phase I, based on the years of best fit of the tulip-poplar and oak KY patterns when cross dated with the area oak and area oak KY patterns may have been as follows:

Second floor ceiling	Last year of tree growth
Joists/sleepers/plate	1715
Roof
Rafters (R1 )	1716
Roof Structure
Gutters	1718

-36-

Peyton Randolph House, Phase II

The cross dating of the Peyton Randolph House, Phase II, is based on tulip-poplar samples removed from the ceiling of the second floor and the roof. The KY pattern developed for the rafters, purlins, joists and summer was cross dated with the area oak/pine key-year pattern at 1821 (pg.24).

Also, the key-year pattern was cross dated with the following combinations of KY patterns:

Area Key-Year Patterns	Year of Best Fit
	1	2	3
Oak	1821	1900	1890
Pine (Geddy)	1767	1788	1770
Oak/pine	1821	1900	1890
Tulip-poplar (PRI)	1719	1684	1738
Oak/pine/tulip-poplar	1821	1900	1890

The most significant year of best fit was 1821 for the oak, oak/pine, and the oak/pine/tulip-poplar KY patterns.

Area Key-Year Patterns	Year of Best Fit	Chi-Square Value	Kappa Value
Oak	1821	14.5	.81
Oak/pine	1821	14.5	.81
Oak/pine/tulip-poplar	1821	14.5	.81
Pine	1767	8.0	1.0
Tulip-poplar	1719	6.0	1.0

The possibility was investigated that the tulip-poplar KY pattern for the Peyton Randolph House Phase II, cross dated with the area oak/pine KY pattern at the year 1767 (Fig. 24). The goodness of fit at the year 1767 was not significant.

REPORT ON A BRICK FROM WILLIAMSBURGH - OXFORD ref: 381h19
28G - structure D. Peyton - Randolf site

A brick from Williamsburgh was submitted to the Oxford Laboratory for a dating feasibility study. The most obvious way of assessing the problems was to attempt to date the brick but limiting the number of measurements with the aim of producing an approximate date within 12-15% error limits.

Two methods of analysis were used: fine grain dating and pre-dose dating on quartz inclusions.

The procedures are described below and the results are tabulated. Our conclusions are recorded at the end.