At the southwest corner of the foundation, evidence matching that at the northwest corner was found. Here, again inset 4 inches from the line of the main south foundation wall, were to be seen the tops of red-sandstone slabs like those found at the north end (fig. 36), in this case with one tie rod still in place. The driveway obscured the point to which the corner of this extending structure could presumably be projected. Subsequent construction against the sandstone slabs had covered their surfaces with a rubble of brick and mortar that appeared to be the foundation for masonry steps (fig. 35). Projecting out from the southwest corner of the foundation was a rectangular red-sandstone block which appeared to be the corner of these superimposed steps.

Although situated under the driveway, it was apparent by projection that Wall B-I joined the southwest corner of Wall C. It will be demonstrated from surviving records that Wall C, with its connecting sections, was the foundation of a full-length veranda.

The belief which persisted for a time that Structure B might have been the courthouse was dispelled by doc.u.mentary evidence showing that it was John Mercer's mansion.

[Ill.u.s.tration: Figure 35.--SOUTHWEST CORNER OF STRUCTURE B, showing molded-sandstone trim with added brickwork in front. Bricks also covered red-sandstone block, lower right. (Diagonally placed bricks at left are not part of structure.)]

SIGNIFICANT ARTIFACTS a.s.sOCIATED WITH STRUCTURE B

_Date _Artifact_ of Manufacture_ _Provenience_

2 rim sherds from ca. 1730 Beneath flagstone in brown-banded; porch ap.r.o.n north "drab," stoneware of Structure B.

mug (USNM 59.1754; fig. 67b)

Iron candle-snuffer 1730-1750 Debris at south end (USNM 59.1825; ill. 62) of Structure B.

Small crescent-shaped Debris at south end chopping knife of Structure B.

(USNM 59.1837; fig. 85a)

Silver teaspoon ca. 1730-1750 Wall debris near (USNM 59.1827; fig. 86d) north end.

In addition, there was the usual variety of 18th-century delftware, Nottingham and white salt-glazed stoneware, pieces of a Westerwald stoneware chamber pot, and much miscellaneous iron, of which only a hinge fragment and a supposed shutter fastener probably were a.s.sociated with the house. None of this material has provenience data, nearly all of it having turned up in the process of trenching. Little of it, therefore, throws much light on the history of the structure. The most important artifacts found in and around Structure B are those of an architectural nature, and these will be considered primarily in the following section.

ARCHITECTURAL DATA AND a.n.a.lYSIS OF STRUCTURE B

That the "manor house," as Thomas Oliver called it in 1771, was an extraordinary building is both revealed in the Structure B foundation and confirmed by the insurance-policy sketch of 1806. Long, low, and narrow, fronted by a full-length veranda and adorned with stone trim for which we can find no exact parallel in 18th-century America, it was as individualistic as John Mercer himself. Yet, far from being a vernacular anachronism or a mere eccentricity, it was apparently rich with the Georgian mannerisms that made it very much an expression of its age.

[Ill.u.s.tration: Figure 36.--SOUTH WALL OF STRUCTURE B, looking east. Base of veranda extends to bottom of picture at left. Molded-sandstone trim appears through brick rubble that has been attached to it, evidently as base for steps.]

The measurements made of the foundation when excavated, as we have seen, show a length of 108 feet and a width of 28 feet for the main structure, with an overall width, including the projecting Wall C, of 37 feet 6 inches. The insurance policy states a length of 108 feet 8 inches and a width of 29 feet 6 inches for the main foundation, plus a separate width for the "portico" (as the structure above Wall C was called) of 8 feet 4 inches. These small discrepancies probably lie in the differences between measuring a standing house and a foundation.

Despite the fact that the foundation was far from fully excavated because of the presence of trees and highway, it is clear, nevertheless, that two cellars of unequal size were situated within the main foundation, separated by sections where there were no cellars. These findings correspond with the notation on the insurance-policy plan, "a Cellar under about half the House."

[Ill.u.s.tration: Figure 37.--CELLAR OF STRUCTURE B, showing remains of original cross wall at left and added cross wall at right. Mercer probably referred to the latter in 1749 in his account with Thomas Barry: "Underpinning and altering the cellar."]

The partly destroyed cross wall extends about midway across the foundation, acting as a retaining wall. As described above, this cross wall was found to be tied into the brick pavement that ab.u.t.ted it on the south side.

The bricks in the main foundation walls and in the partly destroyed cross wall and pavement, on the basis of sample measurements, show a usual dimension of about 8-1/2 by 2-3/4 by 4 inches. An occasional 9-inch brick occurs--about 10 percent of the sample.

In contrast, the bricks in the second cross wall are all 9 inches long, except two that are 8-1/2 inches and one that is 8-3/4 inches. Similar sizes prevail in the bricks exposed in the "portico" foundation (Wall C) at the south end. The significance of these brick sizes will be discussed later.

It is clear that Wall C was the foundation of the "portico," and that by "portico" the writer of the insurance policy meant veranda or loggia.

The policy also shows a "Porch 10 by 5 f." extending from the middle of the veranda. The highway now covers this spot.

In the s.p.a.ce between the two parallel cross walls within the main foundation, the debris yielded a large section of a heavy, red-sandstone arch, 14 inches wide, 9 inches thick, and 3 feet 2 inches long. This arch was roughhewn on the flat surfaces and on about half of the outer curved surface, or extrados. The inner surface, or intrados, and the remainder of the extrados are smoothly dressed (fig. 38). At the south end of the main foundation another curved red-sandstone piece was recovered. This piece curves laterally and has a helically sloped top surface. It is 25 inches long, 14-1/2 inches high at the highest point, and 9 inches thick. Presumably, it was part of a flanker for a formal outdoor stair or steps (fig. 39). Also at the south end was found a cast-mortar block with grooves on the back for metal or wooden fastenings (USNM 59.1823; fig. 40). This was perhaps part of a simulated ashlar doorframe. A few gauged or "rubbed" bricks occur that are slightly wedge shaped.

[Ill.u.s.tration: Figure 38.--SECTION OF RED-SANDSTONE ARCH found in cellar, presumably from an arcade surrounding the veranda.]

Turning to the doc.u.mentary evidence, one may recall that an item dated September 1747, "By building part of my House," appeared in David Minitree's account in Ledger G. Two years later, in 1749, several items related to the house appeared in the account of Thomas Barry, "By Building the Addition to my House/ By 22 Arches/ By 900 Coins & Returns/ By a Frontispiece/ By Underpinning & altering the Cellar." In 1749 and 1750 William Copein was paid for mason's work.

[Ill.u.s.tration: Figure 39.--HELICALLY CONTOURED red sandstone, possibly a flanker for the steps at the south end of the veranda, near which it was found.]

[Ill.u.s.tration: Figure 40.--CAST-CONCRETE BLOCK, probably part of a rusticated door enframement. Found at south end of Structure B. (See ills. 1 and 2.)]

[Ill.u.s.tration: Figure 41.--DRESSED RED-SANDSTONE SLAB (originally in one piece), molded on both edges. Although last used as a doorstep in Structure E, this slab was probably designed as trim for the sides of steps connected with the main house (Structure B).]

[Ill.u.s.tration: Ill.u.s.trations 1 and 2.--Front and back of cast-concrete block, probably part of a rusticated door enframement (fig. 40).

One-fourth. (USNM 59.1823.)]

[Ill.u.s.tration: Figure 42.--FOSSIL-EMBEDDED black sedimentary stone, used for hearths and fireplace surrounds in the mansion.]

There is a clear sequence here. "Building part of my house" referred to the basic brick structure built in 1747 by Minitree on the main foundation. The work of William Monday, the carpenter, followed in 1748.

This doubtless included building the roof, setting beams, laying floors, and building part.i.tions. Then in 1749 Barry built the "Addition to my House"--almost certainly the veranda. The item for 22 arches is difficult to understand unless one relates it to the veranda and divides the figure in two. The veranda was probably an arcade having 11 arched openings, with arched facings of rubbed brick both inside and outside the arcade. Thus, for the bricklayer, each actual arch would have required two arches of brick. The intrados, or undersurfaces, of the arches were probably red sandstone, like the fragmentary arch found in the site; the basic element of the arch was then faced on each side with bricks also arranged in an arch formation. The arcade at Hanover courthouse seems to have been built in a somewhat similar fashion, except that there the brick facing appears on the exterior of the arch only. The "900 Coins and Returns" probably are gauged bricks, that is, bricks ground smooth on a grindstone to provide a different texture and richer red color to contrast with the ordinary wall brick. They were widely used in Virginia mansions of the 18th century for corner and arch decoration. At Marlborough over 600 rubbed bricks would have been required to trim the piers of 11 arches, while the remainder may have decorated the porch. The porch, we may be sure, was the "Frontispiece."

[Ill.u.s.tration: Ill.u.s.tration 3.--Iron tie bar used to secure dressed red-sandstone slabs to each other. One-fourth. (USNM 59.1833.)]

[Ill.u.s.tration: Figure 43.--FOUNDATION OF PORCH at north end of Structure B, surrounded by flagstone pavement.]

The item for "Underpinning & altering the cellar" probably refers to the knocked-out original cross wall and the added parallel cross wall, although the reasons for the change will always remain a mystery. As has been noted, the average brick sizes in the main foundation, on the one hand, and those of bricks in the new cellar cross wall and in the veranda were mostly different. Probably the distinctions represent the differences between Minitree's and Barry's bricks.

[Ill.u.s.tration: Figure 44.--PLAN OF MANSION HOUSE drawn on a Mutual a.s.surancy Society of Virginia policy of 1806 after the house was acquired by John Cooke. (_Courtesy of Virginia State Library._)]

The detailed sequence of joiners', plasterers', and painters' work during the 1748-1750 period has already been given attention in the historical section, enough to indicate that the mansion was one of luxurious appointments. The insurance policy describes it as a "Brick Dwelling House one Story high covered with wood." In modern parlance this would be called a story-and-a-half house with a wood-shingled roof.

The veranda, probably in the form of an arcade, was trimmed with dressed red sandstone and perhaps paved with the squares and oblongs of this material found scattered around the site. The small projecting porch mentioned in the insurance policy provided a central pavilion. The appearance of the house from here on must be left wholly to speculation with only hints to guide us. We know, for instance, that a considerable amount--three books--of gold leaf was employed. Was there, perhaps, a small gilded cupola to break the long expanse of roof line? Were the 162 ball.u.s.ters, purchased from George Elliott towards the time of completion, made for staircases indoors or for a bal.u.s.trade along the roof? Or did they border the roof of the veranda? To these questions there can be no answer. Another question is whether the house, described as one story high, was built over a high bas.e.m.e.nt or near ground level.

Here we have evidence pointing to the latter, since the foundation had two separate cellars, equalling "a Cellar under about half the House." A high or English bas.e.m.e.nt, by contrast, would have been continuous.

Furthermore, the veranda was at, or near, the ground level. The ground floor thus might have been as much as 3 feet higher, reached by steps from the veranda--but not a whole story higher. The depth of the cellars, ranging from about 4 to 5 feet below ground level, implies that the first floor was not more than 3 feet above ground level.

Suggestions as to details of trim and finish are made here and there, again in fragmentary hints. Several broken pieces of a dark-gray, fossil-embedded marble survive from the "chimney-pieces" and hearths of fireplaces (fig. 42). They may be the "hewn stone from Mr. Nicholson"

paid for in 1749. A piece of plaster cyma-recta cornice molding shows that some rooms, at least, had plaster rather than wooden ceiling trim (USNM 59.1829, ill. 4). Thomas Oliver's statement that "the Manor house wants lead lights in some of the windows" suggests an unparalleled anachronism, since the term "lead light" is an ancient one referring to cas.e.m.e.nt sashes of leaded gla.s.s. But it is inconceivable, in the context of colonial architectural history, that this house should have had leaded-cas.e.m.e.nt windows, and it is very probable, therefore, that the semiliterate Oliver was indulging in a rural archaism to which he had transferred the meaning of "sash lights." The latter term was used commonly to denote double-hung, wooden-sash windows, such as Georgian houses still feature. In support of this inference is the complete lack of archeological evidence of leaded-gla.s.s windows.

[Ill.u.s.tration: Ill.u.s.tration 4.--Cross section of plaster cornice molding from Structure B. Same size. (USNM 59.1829.)]

The cellarless areas of the foundation may have provided the footings for chimneys. These probably stood several feet from the ends, perhaps serving cl.u.s.ters of four corner fireplaces each, for each floor. One may surmise that there was a hip roof, with a chimney rising through each hip. A porch at the north end had a rectangular brick base 4 by 6 feet, surrounded by a flagstone area 16 feet wide and 8 feet 5 inches in extent from the house. This evidence, however, differs from the figures given in the insurance plan which shows a "Porch 8 by 6 feet."

The mansion embodied some characteristics which are traditional in Virginia house design and others which are without parallel. The elongated plan indicated by the foundation was more frequently encountered in Virginia dwellings of the late 17th and early 18th centuries than in the "high Georgian" mansions of the 1740's and 1750's.

Turkey Island, for example, built in Henrico County in the 17th century, was 103 feet long, 5 feet less than Marlborough.[149] The additions to Governor Berkeley's Green Spring Plantation, built during the late 17th century, consisted of an informal series of rooms, one room in depth for the most part. Waterman is of the opinion that Green Spring was "in a sense an overgrown cottage without the real attributes of a mansion."[150] The excavations conducted in 1954 by Caywood have altered the basis for this opinion somewhat, but, with its 150-foot length, Green Spring remains an early example of the elongated plan.[151]

Aside from being elongated, Marlborough derives from the ubiquitous informal brick cottage of Virginia. So indigenous is this vernacular form that it is often found in houses of considerable pretension, even in the 18th century. Such are the Abingdon glebe house in Gloucester County, Gunston Hall in Fairfax, and the Chiswell Plantation, known as "Scotchtown," in Hanover. Robert Beverley noted the Virginians' fondness for this style, commenting that they built many rooms on a floor because frequent high winds would "incommode a towering Fabrick"--an explanation as delightful as it is absurd.[152]

That these one-story houses could be completely formal is demonstrated in the unique early 18th-century addition to Fairfield (Carter's Creek Plantation) in Gloucester County, which burned in 1897. This dwelling had a full hip roof, with dormers to light the attic rooms, and a high bas.e.m.e.nt. Its cla.s.sical cornice was bracketed with heavy modillions, while a ma.s.sive chimney protruded from the slope of the hip.[153]

Gunston Hall, on the other hand, reverted to the gable-end form.

Although essentially a Virginia cottage, it is richly adorned with Georgian architectural detail. Completed in 1758, only eight years after Marlborough, and owned by Mercer's nephew George Mason, this building may be more closely related to Marlborough than any other existing house.[154]

[Ill.u.s.tration: Figure 45.--THE VILLA of "the magnificent Lord Leonardo Emo" at "_Fanzolo_, in the _Trevigian_;" ill.u.s.trated in _The Architecture of A. Palladio_ (Giacomo Leoni, ed., 3rd edition, corrected, London, 1742). Palladio's was one of the works owned by Mercer and probably used by Bromley. The arcaded loggias of the one-story wings of this building may have contributed to the inspiration of Marlborough. (_Courtesy of the Library of Congress._)]

Of all the one-story Virginia houses that have come to our attention, only Marlborough has a full-length veranda. To be sure, there are multiple-story houses with full-length verandas, the most notable being Mount Vernon. Elmwood, built just before the Revolution in Ess.e.x County, is another, having a foundation plan similar to Marlborough's.[155] The Mount Vernon veranda is part of the remodeling of 1784, so that neither house reached its finished state until a quarter of a century after Marlborough's completion. Marlborough may thus at the outset have been unique among Virginia dwellings in having such a veranda. However, full-length verandas on buildings other than dwellings were not unknown in Virginia prior to the construction of Marlborough, for they occurred in an almost standard design in the form of arcaded loggias in county courthouses. Typical were King William and Hanover County courthouses, both built about 1734 (figs. 5 and 61).

The arcaded loggia is Italian in origin and is traceable here to Palladio, whose influence was diffused to England and the colonies in a variety of ways. We know that _The Architecture of A. Palladio_ was one of four architectural works acquired by Mercer in 1748 and apparently lent to his "architect," joiner William Bromley. The direct influence of this work on the overall plan of Marlborough probably was negligible.