Colonial Williamsburg Digital Library

a. First Period Fenceline

The earliest or first period fence apparently had been the most substantial. The extremely large postholes (See Plate V) ranged from 3'10" x 4'0", 4'7" deep, to 3'0" x 3', 21" deep, and averaged 3'3" x 3'3", 2'6" deep; and the post molds, some still retaining parts of the posts, ranged from 6" to 1'0" in diameter. The postholes at the corners were 3'6" x 3'8" deep, compared to an average depth of about 2'6" for the holes in between. Most of the first period posts had been set on 9'3" centers on the average with a few exceptions; i.e., gatepost locations (see below), and along each Plate V. View from the north of feature CG 1539 showing second and third period garden fenceline postholes cutting into first period garden fenceline postholes cutting into first period posthole. (See also Fig. 28 foldout rear jacket). (72 INH 898) Plate VI. View from the northeast of trenches CG 1596, CG 1555, showing first and second period garden fenceline postholes cut into original grade of terrace flank and north-south crosse section through trash layers that had been deposited over and around the fenceposts (CG 1555M, CG 1596F) below the brick rubble. (See also Fig. 28 foldout rear jacket) (72 INH 904) Plate VII. View from the south of trench CG 1540 showing partially excavated second period garden fence posthole (CG 1540C) with 8" diameter roughly finished post in situ . (See also Fig. 28 foldout rear jacket) (72 INH 900) Plate VIII. View from the south of trench CG 1525 showing first and second period garden fenceline postholes as they ascend the first tier of the terrace. Note wooden posts in situ . (See also Fig. 28 foldout rear jacket.) (72 INH 899) 11. side of the terrace where some holes were as close as 4'0" (See Plate VI). Surviving sections indicate that the posts had been made of cedar and were left, at least below ground, in a rough, unfinished state. (See Plate VII)

The archaeological evidence points to a post-1740 date for the construction of the first period fence; i.e., the backfilling of the postholes. Artifacts recovered from the postholes consisted primarily of wine bottle glass (CG 1353A, 1496A, 1497B, 1602A, 1596Q, 1596K) and pipe bowls (CG 1353A, CG 1550) made no earlier than 1740, and a vast quantity of artifacts found in trash layers sealing the postholes along the west terrace flank (CG 1555P, CG 1596F) also indicates a post-1740 installation-use date. A wine bottle neck (CG 1498A) found in a first period post mold (presumably having fallen into the cavity after the post rotted or was removed) suggests that the first fenceline was abandoned during the period 1760-1780.

b. Second Period Fenceline

Postholes for a second period fenceline were also found (See Plate VIII). They were smaller and not as deep as the first period holes, perhaps indicating that the second fence was less substantial. The holes ranged in size from 1'7" x 1'8" to 2'0" x 2'6" averaging 2'0" x 2'0" 1'1" deep. The post molds averaged 9" x 9" and were placed on 9'0" centers. The second fence generally followed the same line as the first on the west side and northwest return, but from the southwest corner to about the middle of the east line, the smaller postholes gradually moved outside (to the south 12. and east) of the earlier post line. However, the sections of the east line excavated on the east terrace flank, in the south front yard, and next to the kitchen wing of the mansion indicated that both periods of postholes had merged into the same line.

There is no doubt that the smaller holes were dug and filled after the larger holes, because, in many cases, the small holes actually cut into the fill of the larger holes (e.g., CG 1433, CG 1539) (See Plate VIII). How much time elapsed between the two periods, however, is almost impossible to determine. Sherds of creamware, a ceramic type apparently not used in Virginia until after 1769, were found in a second period posthole (CG 1536D). This suggests that the second period fence was installed sometime after 1769, perhaps two decades after the installation of the first period line. Moreover, common sense suggests that one would not make the effort to put in approximately 160 postholes and posts a second time unless the first fence had deteriorated to the point that it had ceased to function. In fact, the first fence must have been gone entirely on the west line, where both periods were dug on the same line. On the south and most of the east line, however, the earlier fence may have still been in existence during the second fence installation, necessitating the placement of the second period line outside of the first. Also, a ten to twenty year lifespan for the first fence does not seem unlikely in the Virginia climate.

Apparently the second fence, in turn, was abandoned in the first quarter of the nineteenth century. A trash spread (CG 1536B, CG 1536C) found sealing a second period post and mold was found southeast of the kitchen and contained artifacts made after about 1835.

c. Intermediate Fenceline

An intermediate fenceline, forming a 55' 0" x 245' 0" enclosure at the south end of the garden, was also located (See Plate II). The posthole size, averaging 2" x 2", 10" deep, the hole spacing (9'0" centers), and the alignment with the second period posts at each end suggest that this enclosure was added to the plan at the same time or after the second period fenceline. No datable artifacts were found in association with this fenceline.

d. Third Period Fenceline

A third period fenceline was also found, but only along the north half of the main west line (See Plate IX). These holes varied in size from 1' 2" x 1' 2", 2' 4" deep, to 2' 0" x 3' 0", 2' 10" deep, had molds 8" in diameter on the average, and were spaced on an average of 9' 9" between centers. In some cases, these postholes cut into the first period postholes, clearly indicating that they were dug later. However, no datable artifacts were found in the third period posthole fill; therefore, it is impossible to determine when this last fence had been installed. The third period line is also problematic in that it did not appear to join into the office wing of the mansion, nor did the postholes show up near the southwest edge of the terrace. For this reason, one could speculate that the third line was installed only where sections of the Plate IX. View from the south of trench CG 1553, CG 1586-89 showing three periods of garden fenceline postholes after excavation. (See also Fig. 28 foldout rear jacket.) (72 INH 901) 14. line was installed only where sections of the second period fence had deteriorated.

e. Gates

Five, and possibly six, gate locations, indicated by abrupt changes in the fenceline posthole spacing, were found. The central gate on the south fenceline had posts on 4' 6" centers in the first period fenceline and posts on 3' 6" centers in the second period fenceline. Moreover, the easternmost first period gate posthole (CG 1360) was 4' 11" x 3' 0", had been dug to a depth of 3'0" BNG, and had a 9" diameter post mold. Perhaps the increased dimensions indicate that this hole supported the pivot post for a heavy gate. However, if the size and depth of the holes do indicate a heavy gate, then the second period central gate must have been much smaller. The later postholes (CG 1409, CG 1462) were only 1' 4" x 1' 7". Also, the second period central gate included a third posthole located 3' 6" south of the fenceline, probably having supported a 6" diameter post for a gate weight.

Both periods of fences had gates at the east end of Crosspath 1 (See Plate X). The first period gate at that location had 9" diameter post molds, 4' 9" apart, set in one large posthole 5' 0" x 9' 0", while the second period gate had 9" diameter postmolds, 4' 6" apart, set in separate holes. The hole spacing at the west end of Crosspath 1, 5' 6", suggests that a gate existed there, but only during the life of the Plate X. View from the west showing first (CG 1492-93) and second (CG 1479-80) period garden fenceline gate postholes. (See also Fig. 28 foldout rear jacket.) (72 INH 902) 15. second period fence. The larger, earlier holes remained on the standard 9' 3" centers. The hole spacing of 3' 6" also suggests that a gate existed at the center of the intermediate fence. Additional gates may have been located next to each of the mansion outbuildings. The hole spacing diminished to 7' 0" next to the west wall of the office and next to the east kitchen wall, perhaps indicating that a wide gate had stood there.

f. Path System

The area to the south of the terrace enclosed by the main fence system was divided into at least four equal rectangular garden plots (each approximately 100' 0" x 75' 0") by the installation of a system of sand paths. The central path, l' 1" wide, ran 344' 0" from the main south fenceline gate to the foot of the terrace and from there 107' 0" up an 11° - 17° sloping ramp, finally becoming a 6'0" wide crushed shell path running across the south front yard toward the door of the mansion. Paths 6' 0" to 8' 0" wide were also placed along the inside of the east and west fenceline running from the south intermediate fenceline up to and across in front of the base of the terrace. In addition, at least two east-west crosspaths 10'0" wide were placed at about 75' 0" intervals, connecting the central path with the sidepaths.

The first period central gate postholes cut through the sand fill in the central path indicating that the central path pre-dated the fence system. The central path may have also extended out into the field beyond the south fenceline suggested by a 14' 0" gap in the two outside planting(?) ditches (See below).

The path system had been constructed by either leaving some of the natural sand strata undug between garden plots (See below ) inside the flanks of the terrace or by digging trenches and filling them with sand (See Figure 6).

g. Garden Plots

One of the aforementioned rectangular plots, located just south of the terrace, was excavated, showing that the natural sand subsoil had been dug out to a depth of 2' 0" and subsequently filled with a rich brown sandy loam suitable for planting. Also, near the terrace, strips of natural sand were left between the plots to serve as the base for the paths. It is reasonable to assume, but not assured, that the other plots were similarly excavated and prepared for cultivation.

The 105' 0" x 256' 0" area at the south end of the enclosed area was stripped of the modern plow zone, revealing scattered areas filled with the same type of rich brown soil found in the plots, but mostly revealing only the sterile white sandy subsoil (See Plate II). No clearly defined plots or planting areas were found. The intermediate fenceline suggests that most of this area had been used for a paddock.

FIGURE 6 — EAST-WEST SECTION THROUGH GARDEN FENCELINE
Note: See Figure 28 for location.

h. Ditches

Three sets of what appeared to be planting ditches were also found in association with the enclosed area: two 2' 7" wide, 10" deep, 115' 0" - 120' 0" long features (CG 1517, CG 1518), running parallel to and outside of the main south fenceline; two 2' 0" wide, 1' 3" deep, about 176' 0" long ditches on the central path (CG 1590, CG 1593); and, three 5' 0" wide shallow ditches running parallel to and between the east and west fencelines and sidepaths.

Although no planting holes as such were found within the 5' 0" excavated sections of the south fenceline ditches, their plan suggests that they had been dug for planting purposes and not for drainage. Both features ended abruptly at the corner of the main south fenceline and just south of the end of the central path, rendering them useless for drainage. on the other hand, the ditches along the central section of the east and west fencelines could have been used for planting or drainage. Not enough of their profile survived modern plowing to be sure of their purpose. The two ditches on the central path, however, contained distinct planting holes toward the bottom. The sections of these two ditches that were excavated (CG 1590, CG 1593) showed a series of planting holes spaced 2' 0" - 4' 0" apart. However, a nineteenth-century fragment of bottle glass (CG 1593B) found in one of these planting holes suggests that these features were not part of the original eighteenth-century garden plan.

i. Terrace Ramp

As mentioned above, the central path led through the garden to the base of the terrace where it became a sloping ramp. The ramp was tested archaeologically for a distance of 84'0" (See Plate XI). A photograph of 1907-1910 (See Plate XII) shows that the ramp was in use in modern times. An east-west area, 42'0" in length, was also tested on the terrace, and this digging revealed planting holes (CG 1542D, CG 1554B) about 8'0" east and west of the edge of the ramp. The holes, about 2'11" in diameter, 1'2" - 1'6" deep, were backfilled sometime after the second quarter of the eighteenth century (CG 1542D) and after 1785 (CG 1554B). In modern times, perhaps during the renovation work carried out by the McCreas,¹² the ramp was filled in and shaped to match the elevation of the terrace tiers.

j. Terrace

Machine test trenches dug into the terrace and the flanking topography showed that a tremendous amount of soil was quarried out of the original hillside to create the terrace, its flanks, and the lower level field between the terrace flanks. In all probability, the spoil from the original excavation was deposited on each flank to create a knoll on the west and a gradual slope on the east. Artifacts found beneath the spoil on the west terrace flank date this operation to c. 1740-1760 (See page 29). Also, spoil heaps located at the base of each terrace Plate XI. View from the southeast showing terrace ramp (center) original terrace grade and curb (?) (right center) and junction of central garden path with crosspath 3 (at base of terrace). (See also Fig. 28 foldout rear jacket.) (72 INH 903) Plate XII. 1907-1910 photo from the south showing crude dirt terrace ramp still in use. (Valentine Museum, Cook Collection in Mary Stephenson, Carter's Grove Plantation, page 339). 19. flank must have come from the quarrying (i.e., the orange sandy gravel fill matches the natural subsoil strata of the hillside).

A test trench into the first two tiers of the terrace showed that the upper and middle tiers were made up of shaped natural subsoil, while the lowest step was formed partially of slightly mixed natural sand fill above graded natural subsoil. It could not be determined whether or not the actual tiers were put in when the original hillside was quarried or if they had been dug out later. However, it seems logical that both operations occurred at the same time.

There is little doubt that the main garden fenceline and garden paths were put in after the original hillside quarrying operation. In some sections of the fenceline (i.e., along the east and west lines), both periods of postholes cut through the aforementioned terrace spoil piles (CG 1471A, CG 1524E). Also, the east and west fencelines narrowed to 234' 0" apart east-west ostensibly to run along the base of the already existing terrace flanks, and some of the fenceline postholes showed up on the man-made grade of the terrace. Moreover, in some cases, layers of trash dating c.1740-1760 were found sealing the fill in the first period fenceline postholes as well as the shaped terrace flank grade (CG 1525E, CG 1555P, CG 1596F) (See Plate VI), suggesting that the creation of the terrace and the installation of the original garden fenceline occurred at 20. approximately the same time — after c.1740-1760.

k. The Historical Evidence

The historical record confirms the Carter's Grove garden archaeological evidence and dates its first period construction to 1765-66. In June 1765, Burwell's Ledger shows that "1500 Garden Pails [pales or pickets]" were purchased and in 1766 Peyton Randolph was paid for "840 fence Rails" and "921 Stakes."¹³ From this, there can be no doubt that a considerable amount of fence construction was being carried out at Carter's Grove in 1765-66 and that a huge garden was being enclosed by a pale or picket fence. Including the area between the original outbuildings flanking the mansion, the fenceline around the garden found by archaeology measures some 1400'0". It is reasonable to assume that 1500 pales would be needed to secure a fenceline of that length.¹⁴ Also, it may be that some of the 840 rails were used to support the pales, and perhaps some of the stakes were used as the rail supporting posts.

Another Burwell Ledger entry may be the key to a more precise installation date for the second period garden fenceline. In 1774, Burwell sold James Hubbard of Williamsburg 100 cedar garden posts. It could be that these posts were made at Carter's Grove during the reconstruction of the garden fenceline. Also, in 1781, Burwell planted three rows of trees outside the "garden".¹⁵ Symmetrical gardens of the Carter's Grove plan seem to be typical of Tidewater plantations along the James, at least those owned by the Burwell family. A 21. military map of 1781 (See Figure 7) shows that neighboring Kingsmill Plantation (owned by Colonel Lewis Burwell) had no less than two patterned gardens, one (near river) having been there long enough for the river to erode half of it away.¹⁶

Finally, the following description of a strikingly similar garden in Maryland, written about 1800, suggests that the Carter's Grove garden was probably partially decorative, partially utilitarian and perhaps typical of its period:

Md.—Mr. Pratt's [Ca.] 1800

Tuesday 15. The exercise much the same as yesterday — more intoxication.

Saturday 19th … Mr. Pratts garden for beauty & elegance exceeds all that I ever saw —It is 20 rods long — and 18 wide — An ally of 13 feet wide runs the length of the garden thro' the centre — Two others of 10 feet wide equally distant run parrallel with the main alley. These are intersected at right angles, by 4 other alleys of 8 feet wide — Another alley of 5 feet wide goes around the whole garden, leaving a border of 3 feet wide next to the pales — This lays the garden into 20 squares, each square has a border around it of 3 feet wide — Likewise the border of every square is decorated with pinks and a thousand other flowers, which is impossible for me to describe. The remaining part of each square, within the border, is planted with beans pease, cabbage, onions, Betes, carrots, Parsnips, Lettuce, Radishes, Strawberries, cucumbers, Potatoes, and many other articles —Without the pales stand a row of trees upon three sides of the garden — These consist of Pear, Peach, apple, cherry, & mulberry trees. Within the pales, on the out border, are planted, Quince, snoball, Laylock, and various other small trees, producing the most beautiful flowers—

The beauty, Taste, & elegance which attends it, is perfectly indescribable —
It is an enchanting prospect, and carries the spectator, into an extacy, which he cannot describe — The effluvia arising from the various flowers — sweetens the air.¹⁷

Figure 7 — Simcoe Map of 1781 showing gardens at Kingsmill Plantation (arrow) similar in plan to that found at Carter's Grove.

(a.) Pit A

A circular pit (Pit A), 6' 6" in diameter and 4' 8" deep, was found in the north end of the expanded trenches (See Figure 10 and Plate XVIII). The pit had been backfilled with: (1) an initial deposit of domestic trash (CG 460G); (2) clean brown sandy loam (CG 460F); (3) a layer of trash, woodash, and oystershell (CG 460D); and, (4) another layer of trash (CG 455B, CG 460A) below the topsoil (See Fig. 12). The artifacts and cross-mending suggest that this feature had been backfilled in the period 1750-60.

(b.) Pit B

Another pit (Pit B), about 5' in diameter but only 18" deep, was found 7' southeast of Pit A (See Fig. 11). It primarily contained a concentration of broken wine bottle glass, but occasional sherds of colonial ceramics were also found, some of which cross-mended with ceramics recovered in Pit A. Thus, both Pit A and Pit B were probably filled at the same time (c.1750-60) with material coming from the same place.

FIGURE 10

MAP OF CARTER'S GROVE PLANTATION ON THE JAMES RIVER

Plate XVIII. — View facing north of icehouse pit (?) Pit A (CG 460) on west terrace flank (background and cross-section (left) through sandy fill (CG 439B) from terrace construction (?). (See also Fig. 24 for location and Figs. 10 & 11.) (72 INH 910)

(c.) Terrace Fill

The fill in Pit B (CG 460H, CG 439F) had been deposited directly on top of the northern edge of orange clay fill (CG 439B) which, in turn, had been deposited on the natural slope of the hill to the south (See Figure 11). The orange clay itself had been deposited on a spread of brickbats (CG 439C) mixed with original topsoil (See Plate XIX). Wine bottles, probably made in the period 1730-1745, were found with the bricks (See Plate XX). Thus, the orange clay fill had been deposited between the deposition date of the bricks and bottles (after c.1730-1745) and the deposition date of Pit A and Pit B (after about 1750-1760). Also, there is a strong possibility that the broken bricks were left over from the construction of the nearby mansion and/or outbuildings, either the office or the kitchen. A few of the fragments were glazed and glazed headers were used in the outbuildings, but not in the mansion, No mortar was found on or with the bricks, further suggesting that they were left over from construction.

It is also logical to assume that the orange clay fill had come from the construction of the nearby terrace (i.e., the spoil left over from cutting away the natural clay hillside) (See page 18). If that, in fact, is the origin of the fill, then it follows from the above mentioned archaeological evidence that the terrace was constructed during Plate XIX. — View facing south of original topsoil level (CG 439C) containing brick rubble (from construction of mansion outbuildings (?) found beneath sandy fill (CG 439B) from terrace construction (?). Note cross-section through Pit B (on left) (CG 439F, 460H). (See also Figs. 10 & 11). (72 INH 911) Plate XX. — View facing south of wine bottle, ca. 1730-45 (CG 439 B) in situ on brick rubble and original topsoil spread (CG 439C) south of icehouse (?) pit (Pit A). (See also Figs. 10 & 11). (72 INH 912) 30. or immediately after the construction of the outbuildings flanking the mansion or perhaps during or soon after the construction of the entire mansion-outbuilding complex, c.1749-1755.

No structural evidence was found related to Pit A; therefore, its function must be based on speculation. It could have been dug for use as a shallow icehouse pit or cold storage cellar beneath a small storehouse built on simple wooden sills. If such a structure were razed, it is possible that nothing archaeologically would remain except the circular pit. The discovery of goose eggs and a complete chicken skeleton at the bottom of the pit fill (CG 460G) could be considered evidence that the pit had been used for an icehouse (i.e., because of the perishable nature of these items). However, no wash layers were found at the bottom (the usual occurrence in icehouses occasioned by the melting ice). At any rate, if the structure was an icehouse, then the destruction date suggested by the artifacts (c.1750-1760) may indicate that the building served a main house pre-dating the mansion (perhaps the "old house" mentioned in the records), but was razed from view when the "new" mansion was built. This interpretation, of course, raises the question about the location of the "old house" itself, which must have been nearby. The 2' 0" x 2' 0" test holes failed to indicate its presence in the immediate vicinity. 31. It may be that the "old house" was located either where the terraces were constructed or where the mansion now stands. But on this score one is left to contend with a 1770 reference to the "old house" repairs.

a. Tanning Vats(?) Pits A and B

Two of the series of rectangular pits, Pit A and Pit B, set apart from the others, need special mention in that they were the largest and had been lined with wood. (See Plate XXIII). Pit A (CG 715) measured 9' x 6' on the outside extremities and contained a similar wooden box in the center measuring 6'7" x 46". Builder trenches (CG 715B, CG 716G) resulted between the outside edges of the boxes and the original edges of the pit excavation, which had been subsequently backfilled with gold colored clay at the top and mottled yellow clay (Pit B) or sandy fill beneath (Pit A). Both features contained traces of the wooden boxes and where these FIGURE 12 Plate XXII. — View from the northwest of tanning (?) pits in north end of field 13. Note earlier, ditch (CG 719A) across site (left). (See also Fig. 1 for location and Fig. 12). (72 INH 914). Plate XXIII. — View from the southeast of wood lined tanning pits (?) (Pit A foreground, CG 715), (Pit B background, CG 716) in field 13 before excavation. Note lighter clay fill (CG 715B, CG 716G) builder's trenches (?) around the edges and postholes at the ends of each feature. (See also Figs. 13 & 14) 34. had rotted away, the nails that had held them together were left in place in the dark brown loam (See Figures 13 and 14). More structural evidence remained in Pit B than in Pit A. Impressions in the sides of the builder's trenches and lines of nails suggested that the boxes had been made up of twelve 2" x 3" "wall studs", four on each narrow side and two on each long side. The studs on the narrow sides were placed one in each corner and one from 3" to 1' 1" from each corner stud or post. The studs on the longest walls were placed from 1' 0" to 1' 6" from each corner and 3' 6" apart on the north wall and 4' 0" apart on the south wall. These four studs were apparently connected to similar sized timbers which acted as floor joists or sleepers supporting the floor boards. The faint traces remaining of the floor and wall boards suggest that they had been from 6" to 1' 0" wide and 5/8" to 1½" thick, depending on evidence based on the apparent thickness of the wood traces or evidence based on the distance between the nail heads and the "clinch" on the spikes found in situ . Although it was extremely difficult to tell for sure, it is likely that pine was used in the construction of the boxes. The spikes found in line with the stud and joist impressions, and in one case in a section of preserved joist, were 3 ¾" long, had hand-made rose pointed heads, and had been driven in from the outside of the box towards the center. Therefore, the boxes FIGURES 13 & 14 35. had been constructed outside of the pit, then lowered into place. Subsequently, the "builder's trench" between boxes and pit walls was filled with clay.

Pit B was 3' 3" deep and Pit A was 3' 6" deep below the subsoil level; therefore, the studs must have been at least that long. Adding the 1' 0" of plow zone (old topsoil), it is logical to surmise that the pits and boxes must have originally been about 4' 6" deep. It is also logical to assume that Pit A had been built with lumber of the same dimensions as that used in Pit B, although less remained of the studs and joists. What did remain vaguely suggests a similar construction. Both pits had been built in line on a northwest to southeast axis and both had postholes at the center of each narrow side suggesting that some kind of a rack or frame had been built over the boxes.

The stratigraphy in both pits was similar (See Figure 14, Plate XXIV & Plate XXV) in that they had apparently been entirely backfilled with dark brown or black loam (CG 715A, CG 716A) after 2" to 3" of black loam or sandy wash had accumulated at the bottom (CG 716E, F) directly on the wooden floor boards.

Pit B contained lenses of lime and oystershells and some traces of lime were also found in the filling of Pit A. Pearlware sherds found in the upper level of fill in both pits (CG 715A, CG 716A) suggest that the backfilling had Plate XXIV. — View from the northeast of wood lined tanning (?) pit (Pit A, CG 715) showing north-south cross section through fill (background) and domestic trash in situ (CG 715B & C) on floor boards. (See also Figs. 12-14). (72 INH 916) Plate XXV. — View from the east of wood lined tanning (?) pit (Pit B, CG 716) showing north-south cross-section through fill (background) and domestic refuse in situ (CG 716E) on floor boards. (See also Figs. 12-14). (72 INH 917) 36. occurred after 1785. Crossmends of pottery between the upper backfilling levels of Pit A and Pit B show that they had probably been backfilled and abandoned at the same time.

b. Smaller Rectangular Pits

Eleven smaller backfilled rectangular pits were found in the area east and north of the wood-lined pits (See Figure 12). They ranged in size from 2' 8" square, 4" deep (Pit F, CG '708) to 6' 2" x 4' 1", 1' 10" deep (Pit L, CG 643) to 5' 6" x 4' 3", 2' 8" deep (Pit N, CG 721). The clay floors of Pit M (CG 704) and Pit N (CG 721) had been scorched by fire. Pit C (CG 702) contained bricks stacked along the south edge (See Plate XXVI) and Pit L (CG 643) included an associated posthole (CG 736A) 3' 0" from the west edge.

Generally speaking, the pits were backfilled with a brown loam fill containing some woodash, some brickbats, and some domestic trash. Pits D thru L (CG 703, 706, 708, 709, 710, 712, 713) contained sherds of creamware (See Plate XXVII) dating their backfilling to after c.1769; Pit H (CG 709) contained a Virginia copper halfpenny (issued in 1775) and a British(?) 7th regimental officer's military button; and a total of ten pewter spoons were found in the pit fill, one (CG 706) (See Plate XXVIII) still retaining untrimmed mold marks indicating that it had never been used.

Pit N (CG 721) was dug on the same compass axis as the other rectangular pits. However, it is located 55' 0" from the Plate XXVI. — View from the north of Pit C (CG 702), field 13 showing bricks in situ along south wall of feature. (See also Fig. 12 (72 INH 918) Plate XXVII. — Typical creamware types from the small rectangular pit area, field 13. (Top: Royal Pattern plate, CG 713C, Pit G, bottom: bowl, CG 643A, CG 643B, CG 706A, Pit E and Pit L), both after ca. 1769. (72 NJI 267) Plate XXVIII. — Two of ten pewter spoons recovered from the fill in the small rectangular pits, field 13. Note: untrimmed mold flaw on tip of spoon on the right (CG 706A) showing that it was never used and molded maker's mark (unreadable) on back of handle of spoon on the left (CG 643A). 72 NJI 274) 37. other relatively closely spaced rectangular features. Also, as mentioned above, a fire had scorched the clay floor and much care had been taken to make the sides vertical. The pit contained four separate layers of fill (See Figure 12) inset) and Plate XXIX: (1) mottled clay at the top (CG 721A); (2) a brown loam layer (CG 721B) containing a considerable quantity of brickbats and whole bricks; (3) 6" of concentrated woodash containing most of the artifacts (CG 721C); and finally, (4) a burnt clay floor scorched to a depth of about 1" (CG 721D). The presence of a hand-painted pearlware cup base (CG 721C) suggests a backfilling date of some time after 1785.

c. Relationship and Interpretation

Because of the subsequent plowing, there was no way to relate the pits stratigraphically. However, artifact crossmending, dating of the fill and spacial considerations together suggest:

• (1)They have a related function.
• (2)They stood open at the same time.
• (3)They were backfilled at the same time (c.1780-1790).
• (4)They were backfilled with material from the same source (i.e., domestic refuse from the same building).

Although the artifacts suggest that a temporal relationship had existed among all the pits, the presence of Plate XXIX. — View from the southeast of Pit N field 13 (CG 721) showing north-south cross-section through fill above scorched clay floor. Note iron trowel (CG 721C) projecting from bottom layer of fill. (See also Fig. 12). (72 INH 919) 38. pearlware in the two wood-lined pits (A and B, and in Pit N) seems to indicate that they were filled last. Ceramics crossmended between Pit A and Pit B, but not directly with the others. However, pieces of what appear to be from the same Chinese porcelain plate were found in Pit B and in Pit J (CG 710). Pit J, in turn, contained a sherd of Colono-Indian pottery identical to Colono-Indian pottery found in Pit H (CG 709) (shown above to have also contained the Virginia halfpenny of 1773). Thus, regardless of the pearlware in Pit A and Pit B, it is still likely that the large wood-lined pits were used at the same time as the nearby smaller rectangular pits.

Before the entire area, now in woods, surrounding Field 13 can be excavated, there is probably little value in drawing rigid conclusions as to the function of the pits. However, the excavations so far strongly suggest certain possibilities.

The wooden linings and flooring, the postholes at each end, and the presence of lime in the fill can all be considered evidence that Pit A and Pit B were used in tanning leather. Historical documents seem to fit the archaeological evidence. The preparation of the pelts required repeated soakings in lime pits or vats, and the tanning process itself required long-term soaking of the prepared pelts in a series of vats containing tannins. Presumably the extraction of the organic 39. tannin would require cooking of leaves or bark over a low steady heat. Perhaps the fire pits (M and N) were used for this function while Pit A and Pit B were used for pelt preparation and/or tanning. Diderot shows a sketch of a tannery with round vats ranging in function from lime soaking to fire pits used in heating cauldrons of tanning solutions.²⁰ Moreover, a sketch of W. H. Pyne (1808, See Plate XXX) shows tanning being done in rectangular wood-lined vats with recessed corner posts strikingly similar to Pit A and Pit B at Carter's Grove. Pyne also shows "clothes-lines" in the background used as drying racks, which may have been the purpose of the postholes (See below) found near the vats at Carter's Grove. Finally, entries in the Burwell Ledger in the years 1742-64, 1765, 1773, 1774, and 1782, shows that leather was being tanned at Carter's Grove.²¹ In fact, Carter Burwell was so adapt at the craft that he sent detailed instructions concerning hide curing to Thomas Jones in 1749.²²

Another possible interpretation of the rectangular pits suggested by the nine smaller pits, the military button recovered in Pit H (CG 709), and the lack of structural evidence is that the field was the site of a military encampment — the smaller pits being merely a series of backfilled latrines. However, one military button is hardly enough evidence by itself to show a military function of the features Plate XXX. — Rectangular wood lined tanning vats as sketched in 1808. Note recessed corner posts matching slots found at corners of Pit B (CG 716) at Carter's Grove. (See also Fig. 13). 40. and the permanent nature of the wood-lined pits seem to rule it out. Also, one must always keep in mind the possibility that excavations in areas to the west and north of the field will produce structural evidence.

The presence of ten pewter spoons in the backfilling of the pits, one still retaining the untrimmed mold flaw on the tip of the bowl (CG 706A), might be considered further evidence that the pit area had been used for crafts (See Plate XXVIII). However, one would have expected to find considerable quantities of lead and pewter waste, particularly around the fire pits (Pit M and Pit N). None was recovered. Moreover, there does not seem to be any consistency in style among the spoons. Two different types were found (See Plate XXVIII). It is logical to assume that if spoon manufacture was being carried out in the area, the styles would probably not be so varied. Also, two different maker's marks were found on the spoons (CG 643A, CG 707A).

Archaeological conclusions must be based on the positive evidence at hand. However, in the case of the features found in Field 13, the unusual absence of wine bottle glass in the backfill of the pits may be significant. Fragments from no more than five wine bottles were found in the fill of the smaller rectangular pits. The lack of wine bottles, although it is negative evidence, might help in determining the origin of the fill itself. It is logical to assume 41. that for obvious reasons, the issuance of liquor to slaves would be kept at a minimum or prohibited altogether.23 If this were the case at Carter's Grove, then there is every reason to believe that domestic refuse from slave quarters would contain very little wine bottle glass.

d. Other Features

Pit M (CG 704) had been dug through a shallow, narrow, backfilled ditch (CG 719) which ran at least 86' across the pit area in a southwest-northeast direction. It is certain, therefore, that the ditch had been abandoned and backfilled before the area began to be used for pits. About 21' 0" from the northeast end of the ditch, a smaller connecting ditch had been dug at about an 80° angle to the longer ditch which gradually became shallower and disappeared about 17' 0" from the junction. At 61' 0" from the northeast end, the ditch (CG 719) had also been cut through by the digging and backfilling of a small trash pit or tree hole (CG 740E), the artifacts suggesting a deposition date in the period 1760-1770. Only two artifacts were found in the long ditch fill (CG 719A) — a minute sherd of Chinese porcelain and a fragment of a pipe bowl, both not specifically datable. But it is reasonable to suppose that the ditch had silted up and had been abandoned before c.1760-1770, the deposition date of the trash in feature CG 740E.

Additional ditches of uncertain purpose were also found northeast of the pit area (CG 740A-C, 740D-E) and three irregular holes, presumably backfilled tree holes (CG 700, CG 707, CG 701) and a borrow pit, were found in the area of the rectangular pits. The borrow pit (CG 714) had the rough appearance of a rectangular pit at first, but it had extremely irregular sides and had been dug into an area where the subsoil consisted of an outcropping of gold clay. It is extremely probable that the clay from this pit was used to fill in around the wooden boxes in Pit A and Pit B (See page 33 ).

Numerous post holes were found in the area of the rectangular pits, six of which (CG 724, 725, 726, 729, 730, 732) may have been contemporary with each other and may have been dug to support the posts of a crude structure or drying rack. However, with the absence of stratigraphy in the area, there is little evidence to relate the holes, apart from their spacial arrangement. The presence of creamware in some of the backfilling of these holes dates their backfilling after 1769.

About 100' 0" east of the rectangular pit area, an irregularly shaped deposit of dark fill (CG 717) was found measuring 36' 0" x 20' 0" x 9' 0" (See Figure 12). Time and weather stopped complete excavations of this feature, but 16' 0" on the south end was tested. An irregular hole sloping into a ditch 43. 2' 9" deep was found. The ditch had been backfilled primarily with brickbats, some of which had come from poorly fired or overly fired bricks. No mortar was found with the bricks. The artifacts found in the fill suggest that the feature had been backfilled after 1769 and crossmending suggests a relationship with the rectangular pits.

Paving made of unmortared brickbats was found on the east side of the feature measuring roughly 5' 0" x 3' 6". It appeared to have a straight side along the west edge and another straight side along the southwest edge. Time and weather halted further excavations in this area.

Also, surface surveying in the edge of the wooded ravine 70' 0" west of the pit area located a depression 20' 0" wide shaped like an inverted cone (See Figure 25). It is probable that this feature was a well abandoned and allowed to erode in at the top. The feature was not excavated.

a. Brick Clamps

Clamp 1

Clamp 1, located 1000' 0" southeast of the mansion, (Foldout Figure 15) consisted of a rectangular area of burnt clay subsoil measuring 43' 0" x 17' 0". The clay had been scorched black, roughly in alternating sections or strips 9' wide on an average. Less black scorched earth and some brick dust were found in the intervening areas vaguely forming strips 3' 9" wide on an average. These alternating deposits may represent fire channels and brick "benches", but the edges of each were extremely difficult, if not impossible, to define exactly.²⁵

No datable artifacts were found in association with Clamp 1.

Clamp 2

A second burned area (Clamp 2) was found parallel to and 10' 0" to the west of Clamp 1 (See Plate XXXI). The burned area was smaller, measuring only 18' 0" x 12' 0". This feature did not appear to have been burned in strips, but in the FIGURE 15 Plate XXXI. — View from the northeast of brick clamp 2, field 8. Note scorched black clay (background center) and symmetrical alignment of postholes. (See Fig. 15) (72 INH 920). 45. northeast corner, the fire had been hot enough to burn the clay subsoil to a brick-like consistency and color. A row of 5 postholes was found running northeast-southwest through the burnt clay area (1' 1" x 1' 2", 1' 7" deep on the average). The holes were spaced at 9' 6" intervals and one contained the remains of a charred post 6" in diameter. Also, paired sets of smaller postholes(?), averaging 8" x 10", 4" deep, were found about equi-distant (10' 0" southeast and northwest) from the larger central holes, making the clamp 20' 0" x 40' 0". It is probable that the central posts supported the ridge pole and the side holes the eaves of a crude roof or cover.

Clamp 2 also included a 11 0" wide, 5" deep, 15' 8" long ditch (CG 1181A) running in a southeast to northwest direction along the northeast edge of the burned area. Also, no datable artifacts were found in association with Clamp 2.

Clamp 3

Clamp 3 (See Plates XXXII and XXXIII), located 85' 0" west of Clamps 1 and 2, consisted of a strip of: (1) burnt clay subsoil 190' 0" x 14' 0" - 16' 0"; (2) a 289' 6" long row of 16 central postholes ranging from 10" x 10" to 1' 10" x 2' 7", 11" to 2' 4" deep, containing charred posts 6" to 10" in diameter (See Plate XXXIV);); (3) narrow shallow ditches along each side of the burnt area; and, (4) a series of smaller postholes (averaging 10" x 1' 0", from 6" to 2' 1" deep) Plate XXXII. — Aerial view of brick clamps 3 (right) and 4 (upper left) from the south. Note alignment of central and side postholes and parallel ditches along northern section of clamp 3. See also Fig. 15 and Plate XXXIII. (71 F D 581-11). Plate XXXIII. — View from the north of brick clamp 3 partially excavated. (See also Fig. 15 and Plate XXXII. (72 INH 921) Plate XXXIV. — View from the northwest showing a charred central supporting post in situ in posthole CG 1203A brick clamp 3. (See also Fig. 16 for photo angle). (72 INH 922) 46. ranging from 7' 0" to 15' 0" apart along the outside of both ditches. The smaller postholes varied from 22' 0" apart (east-west) at the north end of the feature to 23' 0" apart in the center, and to 17' 6" apart at the south end on the average. A third ditch, parallel to the ditch along the east side of the clamp, was also found measuring 1' 0" wide at the north end to 6" wide at the south end.

The plan of this enormous feature (See Figure 15) suggests that it was made up of at least three separate clamps (units), set up on roughly the same line but probably burned at separate times. For instance, the central postholes are not exactly in line throughout, but align in three sections which, in turn, generally correspond to slight differences in the other associated features. More specifically, the first 87' 6" (or the first 6 central postholes from the north) align and the side ditches are parallel 13' 6" to 14' 0" apart (See Figure 15, Unit A). Then the center line of the charred posts shifts slightly to the east for the next 100' 6" (6 central posts), and the side ditches widen (16' 6" to 17' 0"), remain parallel and end about 100' 0" to the south ( Unit B). Beyond that point, the alignment of the remaining 5 central holes (77' 0") shifts again slightly to the east and the side ditches disappear (Unit C). Thus, it appears that the three individual units within Clamp 3 from north to south would have measured about 87' 0" x 14' 0", 100' 0" x 17' 0" and 77' 0" x 14' (?) respectively.

A total of six datable artifacts was found in situ in association with Clamp 3. A sherd of English Buckley coarse earthenware of c.1750 (CG 1136A) and a pipe stem possibly of the period c.1710-50 (CG 1136A) were found in the burnt clay area, and a pipe stem of the period c.1710-50 (CG 1225B), sections of a Yorktown pot of post 1730 (CG 1225B), and a burned wine bottle bottom, probably of the period c.1740-60, were contained in the fill of the east side ditch (CG 1225, CG 1225B). Together these artifacts suggest that the clamp was used about c.1740-60, and perhaps after 1750.

Also, a wine bottle neck of c.1740-60 (CG 1237A) found in a shallow brick rubble depression 4' 0" west of the clamp and an eighteenth-century brass spigot found in a similar but larger brick rubble deposit (CG 1234A) located 4' 0" from the first tend to substantiate a mid eighteenth-century date for the use of the clamp.

Clamp 4

A fourth clamp overlapping Clamp 3 and cutting into the west side was also found. The east half of this clamp and the north side of the remaining 75' 0" contained burnt clay. A line of 10 postholes ranging from 1' 0" x 1' 0" to 1' 9" x 1' 10", 1' 8" - 1' 9" deep, ran down the center of the burnt area. Four of the 10 central postholes contained charred posts 5½" to 7½ " in diameter. A north-south ditch 1' 2" wide, 48. 12' 0" long, 5" deep, 14' 0" from the east end and an east-west ditch 11 0" wide, 21' 0" long were also included in the plan of Clamp 4 along with a series of side postholes ranging from 5" x 5", 3" deep to 2' 6" x 2' 10", 2' 4" deep. Thus, for all practical purposes, Clamp 4 included the same features as Clamp 3; i.e., charred central posts, burned clay interior, side ditch, and side postholes. Also, it may be that Clamp 4 had been divided into more than one unit because of the alignment of the central holes. The first 7 holes from the east align (which would make a 90' 0" x about 14' 0" unit) and the remaining 3 holes roughly align, which could have been in the center of a second section, 40' 0" x 14' (?).

No datable artifacts were found in association with Clamp 4. However, as shown above, it was built after Clamp 3; therefore, it is probable that it was used about c.1750.

Clamp 5

The final clamp was found 135' 0" to the north of Clamp 1 and Clamp 2. The burned area measured 62' 0" x 14' 0" and, like three of the other clamps, included a line of central postholes containing burnt posts (holes ranging from 9" x 1' 1" to 1' 4" x 1' 4", posts 5" in diameter) spaced irregularly from 10' to 1' 9" intervals. Also, sections of narrow shallow ditches were found running east-west along each side of the burned area; and outside of the ditches, side postholes about 9" x 9" were found in line with these of the central holes. It is probable that this clamp was set up and burned as one 62' 0" x 14' 0" unit.

None of the features of this clamp were excavated; therefore, it could not be dated.

b. Clay Pits

Three large but shallow backfilled pits were also found, two southeast of Clamp 1 and Clamp 2 and one (Pit A) (CG 1091) 10' northeast of Clamp 1 (See Figure 15).

Pit A (CG 1091) measured 62' 0" x 30' 0", the north and east edges being extremely vague. At its deepest point along the southwest edge, the pit was only 1' 6" below natural grade and it contained a brown clay fill with numerous underfired and overfired brickbats (CG 1091C) above a spread of light grey clay (CG 1091D). Both layers, in turn, overlaid a shelf(?) of yellow clay (CG 1091E) with a concentration of underfired brickbats running along the southeast edge. Actually the bottom of the pit sloped toward the shelf(?) which had the appearance of a ditch in profile 1' 7" wide.

Two shallow irregular ditches (CG 1188, CG 1323) led from the pit to a brick-lined well located 4' 0" to the northwest (See pages 55-59). Therefore, it is reasonable to assume that the pit had been dug to catch water drawn from the well.

A wine bottle neck found in the upper layer of fill in the pit (CG 1091C) dates the backfilling to after c.1740-55. Also, the neck joined onto a bottle base (CG 1111J) found some 32' 6" BNG in the well, suggesting both the pit and the well were open, then backfilled at the same time.

South and southeast of Pit A (CG 1091) two other large shallow pits were found, Pit B (CG 1322) and Pit C (CG 1464). Pit B (CG 1322) was 8" deep and contained a mottled clay fill with pockets of brickbats, and Pit C (CG 1464) was 1' 3" deep at a maximum and contained similar fill. No ditches were found in association with these pits. Their purpose is uncertain. Also, no artifacts were found in the fill; therefore, it could not be determined whether or not they were contemporary with Pit CG 1091. However, since the same type of brickbats were found in all three features, there remains the strong possibility that they were backfilled at about the same time; i.e., c.1750.

c. The Wells

Two well shafts were discovered some one thousand feet south of the mansion (Field 8) in the vicinity of the brick manufacturing debris (See Figure 15). Although both wells had apparently been filled in about the same time, the fact that one had been robbed of its brick lining (assuming it had been lined with bricks) led to the conjecture that it was the first to be constructed.

Well A

Perhaps due to continuous erosion this earlier (?) well resembled a bomb crater at the top (See Plate XXXV) because it had a diameter of approximately 17' 0". However, the dimension Plate XXXV. — Elevated view showing the relative proximity of the two wells which were presumed to have been used to provide water for the brickmaking operation. The dark circular area in the wooden platform (foreground) is atop the shaft of Well B and the partially excavated Well A is in the background. Notice the eroded sides in the latter. Photographed from the southwest. 71-BK-877 51. narrowed to a more normal 4' 10" at a depth of 8' 7" below natural grade (See Figure 16) and remained fairly constant to the bottom of the shaft (27' 9" below natural grade). Thereupon, the sides sloped to a maximum depth of 29' 10" (BNG) near the center. As noted above, no bricks were in evidence around the sides and no more than 10 brick fragments were found in the filling, suggesting that the lining was either very carefully robbed or that none had ever been installed. The fact that the remains of a wooden casing were encountered in the lower portion of the shaft seems to imply that the well had not been lined with bricks.

The wooden liner (See Plate XXXVI) consisted of four red oak corner posts (average measurements 7' 11" x 3" x 2-5/8") with wide yellow poplar boards nailed to them from the outside. The tops of the posts were first encountered at a depth of 20' 8" below natural grade and the lower edge of the bottom boards was positioned on level with the bottom of the shaft (27' 9" BNG) with the corner posts continuing down some 10" lower. The poplar timbers used in the wall construction had average dimensions of 2' 11" x 1' 1¼" x 11/8". As already noted, the side pieces had been nailed from the outside (see Plate XXXVII) and, since there was not enough distance between the corner posts and the circular shaft to perform this operation, the casing must have been built above ground and then lowered into its final position.

The fact that the corner posts were found intact with little or no damage to the sawed ends, led to the conclusion FIGURE 16 Plate XXXVI. Upper portion of the wooden lining found in Well A. This framework was first encountered at a depth of 20' 8" below natural grade. (71-FD-712) Plate XXXVII. — The northwest corner of the wooden lining found in Well A. Notice that the vertical post is very close to the edge of the shaft. The side pieces had been nailed from the outside so, since there was not enough room for this operation within the shaft, the construction work must have been performed above ground. (Photographed from the southeast) 71-FD-713. 52. that only a 7' 1" section of the shaft could have been protected by this contrivance. It probably would not have been strong enough to withstand a major cave-in, but this may not have been of primary concern because the sides of the well had cut through tightly-packed marl beds resulting in walls that seemed quite stable. Consequently, it was conjectured that the framework was used as a guide to keep the bucket away from the marl sides as it ascended from the water.

The well's primary fill (CG 1315N) consisted of gray silt with small pieces of marl, but was void of any artifactual remains. The layer began on line with the bottom of the wooden casing (27' 9" BNG), sloping downward to a maximum depth of 29' 10" below natural grade near the center. Directly above the initial deposit was a stratum of silty gray clay (CG 1315M) which was first encountered at 201 0" BNG. This level contained an iron key fragment, the socket and rib portion of an iron hoe blade (See Plate XXXIX), at least 7 pieces of an oval-shaped handle made of oak, and 5 bones from a small animal (An analysis of all bones found in Wells A and B appears in Appendix VIII.

The majority of wooden finds (other than the well's casing) were recovered from a layer of mixed brown clay (CG 1315L) at a depth of between 18' 0" and 20' 0" below natural grade. These included an oak paddle 4' 4" in length, which may have been used in the brickmaking operation (See Plate XXXVIII); no fewer than 13 oak (bucket or tub?) staves l' 3" in length, with surviving clenched nails which indicated Plate XXXVIII. — A portion of the oak paddle, first discovered at a depth of 18' 1" below natural grade, which may have been used in the brickmaking operations. Also visible are the tops of two bucket or tub stoves. Photographed from the northwest. (71-BH-882) Plate XXXIX. — Two fragments of an iron hoe blade found in different levels of Well A's filling. The lower portion was recovered at a depth of 18' 11" below natural grade and the other was encountered at 28' 2" BNG. Notice the maker's mark "AD" stamped four times near the end of the hoe's spine. 71-INH-788 53. how the four rows of binding were attached and also that each was ½" thick; one stave with two ½" holes near the top to suggest that a rope handle had been used to carry the objects; a piece of oak binding 2" wide and ½" thick, which had been secured by ¼" dowels; and a large section of oak binding (3' 9" x 4" x ¼") presumably from a barrel. Only one other object was found in the deposit and it was the end of a hoe blade with the initials "A D" stamped four times (See Plate XXXIX). This piece cross-mended with the socket and rib portion described above.

The well's filling from 5'3" to 18' 0" below natural grade produced just two artifacts: a large bone (CG 1315K) tentatively identified as a tibia from an animal smaller than a horse but larger than a dog or sheep (16' 9" BNG), and a small iron grubbing hoe blade (CG 1315J) from a depth of 8' 5" below natural grade. There was no material in the shaft from 5' 3" to 29' 10" (BMG) to even suggest an abandonment date and, therefore, the objects from higher levels were utilized for determining when the well ceased to exist.

The strata in the upper part of the shaft (0" to 5'3" BNG) were well defined, but all seemed to have been deposited at about the same time. The mottled brown and yellow clay (CG 13151) which produced the small hoe blade (See page 52) was partially sealed by a layer of dark brown loam (see Figure 16) with marly shell (CG 1315H) which, in turn, had been covered by a rather extensive deposit of sandy gray clay (CG 1315G). This last stratum contained an English coarseware pan fragment, two nails, and the majority of one wine bottle along with pieces of at least two others from the period c.1740-65. Thus, 54. this date range serves as the terminus post quem for the well's abandonment.

Immediately above the sandy gray clay were layers of washed brown and yellow clay (CG 1315F) and a deposit of sandy brown loam with pebbles and shell bits (CG 1315E). The latter produced 15 beverage bottle side fragments, 2 nails, an unidentified iron strip, and a very thin piece of sheet brass. The washed layers (CG 1315P) contained a vitrified rim section of a Buckley jar, two wine bottle fragments (one of which cross-mends with another found in CG 1315G), 11 nails, and various animal bones including two pieces of a fossilized whale vertebra. A 4"-9" thick deposit of shell (CG 1315C) sealed the above two layers and extended most of the way across the hole (See Figure 16). It contained no fewer than 8 fragments of Colono-Indian pottery (perhaps from a bowl), pieces of a vitrified brown stoneware jar and Buckley pan, numerous wine bottle fragments, 2 vitrified stones, along with miscellaneous bones and nails.

A layer of dark loam with large brickbats (CG 1315B) had been dumped on top of the shell stratum and it, in turn, was covered by a wedge of mottled yellow clay (CG 1315D) which only produced one nail fragment. On the other hand, the loam deposit (CG 1315B) contained 9 pieces of Colono-Indian pottery (bowl?), 3 sherds of a vitrified brown stoneware jar, 2 fragments of English coarseware, a vitrified stone and brick fragment. There were also 3 bases and one neck from wine bottles of the period c.1740-60. The well's top layer consisted of brownish-gray loam with brick flecks (CG 1315A) containing 2 Colono-Indian 55. pan fragments, a piece of a brown stoneware storage jar, and two vitrified sherds — one from a Buckley (jar ?) and the other from a brown stoneware jar.

The fact that at least four crossmends were found among objects in the top 5'3" of the shaft suggests that the layers were dumped in at about the same time. As already noted, the wine bottle fragments from c.1740-60 implied that the well had been abandoned after that period. Nevertheless, it seems quite possible that the well ceased to exist much earlier because the upper 5'3", from which all dating evidence was obtained, could represent filling thrown in to level the area after the shaft's contents had settled.

WELL B

Only 57'6" (center to center) southwest of Well A were found the remains of another abandoned well which, presumably, had also been used to provide water for the brickmaking operations in the area. It was rather crudely constructed in that the upper 27'11" of lining was formed from rectangular instead of compass bricks. This would logically result in a much weaker wall because there was nothing to prevent the sides from caving. As a matter of fact, the top 5'6" of lining showed signs of falling-because the natural clay around it had begun to push the bricks into a much tighter 2'0" diameter and all of them were leaning down toward the center of the shaft (See Plate XL). Nevertheless, the courses below the 5'6" level seemed to have suffered little or no damage, perhaps due to Plate XL. — The upper lining in Well B, made from regular bricks, which showed signs of imminent collapse because the natural clay had begun to push the bricks toward the center of the shaft. Notice how most of the bricks are leaning downward. Photographed from the southeast. 71-BK-878 56. the fact that most of the shaft had been dug through tightly-packed marl beds (See Figure 17).

The lining of the bottom 6'10" (27'11" to 34'9" below natural grade) was built of compass bricks (See Plate XLI) with average dimensions of 7-¾" x (5¼" x 3-¾") x 2-3/8". It should be noted also that these bricks appeared to have been over-fired since they were a much darker red than normal. In this section, the inside diameter of the shaft was 3'5", with each course consisting of 31 bricks. A rather perplexing situation was encountered at a depth of 31'6" (BNG). Five compass bricks had been wedged vertically behind the well's lining on the east side (See Plate XLII), but no explanation was forthcoming; however, it leaves little doubt that the well's lining was constructed from the bottom. The last course was resting on a wooden curb constructed from four shaped boards (white oak) held together by nails. Since the east and west pieces overlapped those on the north and south sides (See Plate XLIII), the wood surface upon which the brick lining rested was not flat. To remedy this situation, a piece of yellow pine had been placed atop the southerly portion of the curb and secured with two nails; however, no matching filler was found overlying the northerly board and no nails pierced it.

The primary fill (CG 1111K) consisted of gray silt with small fragments of marl and decaying wood, along with persimmon and willow oak seeds. Unfortunately, since there were no datable artifacts in the layer, it was impossible to FIGURE 17 Plate XLI. — Typical 18th century compass brick-lined well shaft but this type of construction was only used in the bottom 6' 10" of Well B. (71 BH 880) Plate XLII. — Five compass bricks wedged vertically behind Well B's lining which left little doubt that the curbing was constructed from the bottom of the shaft. Photographed from the west. (71-BH-881) Plate XLIII. — Crudely constructed wooden curb found beneath the last course of bricks in Well B. It consisted of four shaped boards fastened together by nails. The two at the left and right overlapped those at the top and bottom. 71-BH-879 57. determine how long the well had been in use. The initial deposit began near the bottom of the wooden ring and sloped to a maximum depth of 35'7" (below the natural grade) in the center.

Directly above the initial filling was a 6'9" thick deposit of silty gray clay with occasional brick fragments (CG 1111H & J). The artifacts recovered from this layer (28'0" to 34'9" below natural grade) were far more plentiful and informative than any others encountered in the well. Included in the assemblage were no fewer than 21 identifiable rabbit skeletons, along with the remains of several other small animals (see Appendix VIII for a complete listing). The fact that entire skeletons were found seems to imply that they do not represent the scraps from a colonial meal. Indeed, it would appear more likely that the animals were accidentally killed by drowning after falling down the open shaft. Of course, this would not be possible if the hole had been equipped with a conventional wellhead, but an extant drawing (See Figure 18), depicting an eighteenth-century brickmaking operation, shows a well with no lining above ground level.²⁶ According to eighteenth-century practice, when a site was to have several "furnaces", the well would be lined with masonry "to avoid excessive upkeep".²⁷

The iron objects recovered from this same deposit included a hoe blade with the initials "W K" stamped three times near the end of the rib (See plate XLIV); an axe blade with the maker's mark "W. DEAS" also stamped three times (See Plate XLV); and a 8'4" chain, perhaps from the well bucket, which had a hook at FIGURE 18 — 18th-century working sketch of a brickyard showing the clay preparation and molding sections (see Footnote 26 for source). Plate XLIV. — Iron hoe blade with the maker's mark "WK" stamped three times near the end of the rib. This object was recovered from a deposit of silty gray clay (CG 1111J) at a depth of 32' 6" below natural grade in Well B. 71-INH-787 Plate XLV. — Iron axe blade with the maker's mark "W. DEAS" stamped three times near the center. On the opposite side, there is a heart-shaped emblem with an A (G or C) embossed within it. 71-INH-792 58. one end, a swivel near the center, and a 31/2" diameter ring at the other end.

Unfortunately, no pottery fragments were discovered in the silty gray clay, but the digging did produce pieces of 6 wine bottles from the period c.1740-55 (See Plate XLVI). This find was most important because it provided a terminus post quem for the well's abandonment. A clay marble and a bowl piece from a clay tobacco pipe were also recovered from this level. Of particular interest to the landscape architects was the evidence of plant life. The collection of seeds included at least 171 persimmon, 2 pomegranate, 2 peach, 4 black gum, 1 cherry laurel, and a hickory nut fragment, along with twigs from mulberry and loblolly pine trees and a vine section from a wild grape. The wooden objects (see Figure 19) from this same level (28'0" - 34'9" below natural grade) were most intriguing but are, at present, unidentified as to use or purpose. These included a 2'5" rough yellow poplar log, pointed at one end, with a 1" x 5" finished projection at the other, which may have been intended to support a sill (1); a windlass-like object fashioned from locust, 5'7" long with a fractured tongue on one end, a lathed section 7½" long by 3" in diameter some 1'3¼ from the tongued terminal, and a 2½" x 3" oval hole located 1'3" from the opposite end (2). Also recovered was an oak board beveled at both ends (measuring 2'½" in length, 2-3/8" wide and the width tapering from 3-7/8" to 2½") with a 4½" x 1½" rectangular slot cut into it 4" from one end and a 1¼" hole FIGURE 19 Plate XLVI. — Wine bottles of the period ca. 1740-55 (CG 1111J, left and CG 1091C, 1111J, right) which provided dating evidence for the brickyard. (72 NJI 270) 59. (retaining a portion of doweling) located 5" from the other (3). The deposit also yielded two 1-1/16" thick boards which had apparently been part of the same construction since the nail holes in both were on line (4) and an oak log 4'1", in length by 6" in diameter which had a flat surface (5). There were also several pieces of pine which showed evidence of having been bored into by insects.

The well's upper filling (0" to 28'0" below natural grade) consisted almost entirely of broken bricks (presumably the debris from the manufacturing activity in the area) with average widths of 4¼" and thickness of 2½". The longest fragment encountered was 5". The artifacts from this deposit were far from plentiful and were relatively unimpressive. At a depth of 10'8" several pieces of very thin iron were found and lower (14'0" to 16'0") a wine bottle base of the period c.1740-55 was encountered. The only fragment of pottery from the shaft was recovered at a depth of from 16'0" to 18'0" below the natural grade. It was a pan sherd made of Yorktown lead-glazed earthenware and, since the fractured edges showed signs of having been glazed, the object must have been broken and its pieces left either in or near a kiln. The other interesting finds in the upper filling were five long, tapering poles (the longest being 10'7½") of uncertain purpose, but perhaps waste from rail splitting.

d. Small Pits

Four smaller and deeper circular pits were uncovered to the east of Pit A (CG 1091), two of which contained mixed clay fill with brickbats and water-worn cobblestones at or near the bottom (CG 1196, CG 1198). The function of these pits is also uncertain.

e. Interpretation

There is considerable historical evidence concerning brickyards which aids the interpretation of the archaeological features found at Carter's Grove, (Figures 18, 20, and 21) show the plan and two working sketches of an eighteenth-century brickyard which includes: a clay mixing area at one end of the site complete with a well without a wellhead (R) and a ditch (S) to carry water from the well to catch basins or pits (E) from which water was thrown onto piles of clay (L) for mixing with hoes. Well B (CG 1111), with its two ditches leading to Pit A (basin?) found by archaeology at Carter's Grove generally follows this same plan²⁷ (See footnote 26).

The eighteenth-century plan and sketches also show a system of narrow ditches (M), rectangular in plan, located toward the center of the site. Here bricks were stacked and covered to dry, the ditches serving to carry rain water away from the bottom brick courses. The kiln or clamp is located at the end of the site, opposite the well and mixing area.

As mentioned above, parallel ditches were found associated lines of postholes with Clamp 3 and Clamp 5. Also, lines of postholes FIGURE 20 — 18th-century French brickyard (see Footnote 26 for source). 61. between and along the outside of the ditches suggest that a crude gabled or A-frame structure stood over the areas between the ditches, perhaps more permanent form of shelter than the mats shown in Figure 21. However, the presence of the charred central posts and the burnt clay between the ditches at Carter's Grove shows that the bricks were also fired in the ditch area, which does not parallel the eighteenth-century plan. Perhaps both the drying, then later the firing process took place on the same spot at Carter's Grove. This would have been economical in that only one set of ditches and covers would have been necessary for both drying racks and clamps. Or perhaps the drying racks were located on the elevated area between Clamp 2 and Clamp 3 where the natural crown of the field and shallow ditches(?) would have carried off rain water and where plowing and erosion would have had a better chance to obliterate all the archaeological evidence.

Eighteenth-century clamps, built directly on the ground surface, are described in Neve's Builder's Dictionary (1736 edition) as follows:

They build their Clamps of the Bricks that are to be burnt something like the Method of Building the Arches in Kilns, viz. with a Vacancy betwixt each Brick's Breadth, &c. for the Fire to ascend by; but with this Difference, that instead of Arching, they truss, or span it over, by making the Bricks project one beyond the other, on both sides of the Place, FIGURE 21 — 18th-century sketch of a brickyard showing the drying rack section (see Footnote 26 for source).62. for the Wood and Coal to lie in, till they meet, and are bonded by the Bricks at the Top, which closes up the Arch: This Place for the Fuel, they carry up straight at both Sides, or, which is the same Thing, upright at both sides, 'till it is about 3 Feet high, and then they begin to lay the Bricks, projecting over inwards, till they meet in the middle, which they will do in about 3 or 4 Course of Bricks in Height, the Width of the Mouth being about two Feet and a Half. Above this Arch they lay the Bricks in the Order they do in a Kiln, to 8 or 10 Feet in Height, according as the Clamp is to be in Bigness; for they usually burn a great many thousands in a clamp at a Time…²⁸

The clamps at Carter's Grove, as mentioned above, must have had covers for protection from the rain. Moreover, the archaeological evidence suggests that the central poles supporting the roof ridge must have been left in place during firing, thus producing a central line of charred posts in the process. The side posts supporting the shelter eaves(?) would have been easily removed and salvaged before firing (thus leaving only empty postholes). But after the clamp bricks had been stacked around the central supporting posts, it probably was not worth the effort to remove them before the bricks were burned. It is possible that the tapered poles found in the fill of Well B had been used in the construction of the clamp shelters (See page 59 ).

The Burwell Account book shows that bricks were being "burned" at Carter's Grove as early as 1744 by a Negro Lot and again in 1749 by David Minetree.²⁹ However, in 1750, 63. Burwell purchased some 460,000 bricks, 100,000 of these from Mann Page.³⁰ Thus, it appears that the bricks used in the buildings at Carter's Grove came from two sources; i.e., those made on the site and those made elsewhere. The bricks used in the mansion at Carter's Grove (at least on the exterior) are distinctly purplish-pink in color, contain a considerable amount of quartz pebbles and none of the headers are glazed. On the other hand, the bricks found in the area of the brickyard site and in the filling of Well B (CG 1111) are generally orange or red in color, some were glazed and they contain little or no quartz pebbles. This brick type essentially matches the bricks used in the kitchen and office wings flanking the mansion. Therefore, based on brick types, there is reason to believe that the bricks used to construct the earlier kitchen and office wings of the mansion were those burned at the Carter's Grove brickyard, perhaps by Lot in 1744 or by Minetree in 1749, while the bricks used on the facade of the mansion were those purchased in 1750.

a. CG 940, The Square-Ditched Feature

The largest eighteenth-century feature, CG 940, (See Figure 25 and Plate XLVII) at first had the appearance of a backfilled cellar hole (13'0" x 13'0") Subsequent excavation, however, showed the feature to be only 1' 0", deep, probably never having been any deeper than 2' 6" below the surface. A ditch was found running along the west edge of the feature (See Figure 22). It continued 15' 0" north, made a 90° turn to the west and ran 75' 0" to the edge of a slope in Field 13.

The stratigraphy indicates that the square feature and the ditch were open, then backfilled at the same time. Both were filled with a layer of mixed grey and brown clay (CG 940B) and a layer of brown loam containing brickbats and brick dust (CG 940A&C). Apparently the bottom of the square feature had been dug so that it would drain into the ditch, and lenses of what appeared to be rust (CG 940J) were found on the bottom of both the square feature and the ditch. In the south end of the ditch, a concentration of brickbats and brick dust was found immediately below the plow zone (CG 940C).

The artifacts found in the brick layer (CG 940C), consisting of fragmented wine bottles of 1740-60 date, a foot ring from a plain white saltglazed stoneware saucer probably of mid eighteenth-century date, a barrel band and FIGURE 22 — Plan of Feature CG 940, Field 14. (See also Figure 25 and Plate XLVII) Plate XLVII. — View from the east of feature CG 940 field 14 showing square pit (with balk retained through center) and right angled drainage ditch (center and background) (See also Fig. 25 for location and Fig. 22 for photo angle). (72 INH 923) 65. a mortising axe (See Plate XLVIII) date the backfilling of the square feature and the ditch sometime after c. 1750. No artifacts were recovered in the lower level or on the clay floor surface. Occasional pieces of what appeared to be decayed bark were found in the gray clay fill.

The shape and uniform sides of the square feature suggest that it could have been inside a structure. If so, the absence of postholes around the edges suggests a structure built on sills laid directly on the ground surface. What such a square structure (if indeed it was a structure) with its drainage ditch could have been used for is uncertain. It is possible that the feature was used as part of leather curing or tanning operations with the ditch serving as its drain. The "rust" stains could actually be tannin deposits. Whatever its function, however, there is a good possibility that the backfilling layer containing the brick could have come from the construction of the mansion. It is almost certain that similar fill found nearby in another eighteenth-century feature (CG 980) came from the mansion construction.

b. Feature CG 980

A 7' 0" x 4' 0" rectangular feature (CG 980), also with a small ditch at one end, was found 50' 0" southeast of CG 940 (See Figure 25). In profile, the feature looked like an irregular Plate XLVIII. — Mortising axe (CG 940C) found in the fill of feature CG 940. Note: "T" shaped maker's mark (letters did not survive). (72 NJI 278) 66. pit or ditch with sloping sides. The feature had been dug 2' 6" below the surface of the subsoil in the center. Weather conditions and modern intrusions negated tracing the feature any further to the south.

Like CG 940, CG 980 had been backfilled with brown loam containing brick and brick dust. Fragmented wine bottles were found in the fill suggesting a post 1750 deposition date. Therefore, it is probable that CG 940 and CG 980 had a related function (stood open at the same time) and/or they at least had been backfilled at the same time with material coming from the same place.

The bricks in the fill in CG 980 indicate that they had come from the construction of the mansion. Water table bricks (CG 980A) (See Plate XLIX), identical in shape to those used in constructing the water table courses in the mansion, were found along with other rubbed bricks, presumably for use in window arches. These bricks were the same color as the brickbats and dust found in the fill of CG 940. However, since brick was found only in the backfilling material of each feature, it is probable that they have no relationship to the function of the features. At any rate, rubbed bricks, a mortising axe (found in CG 940) and wine bottles all could be part of debris left over from construction, and their deposition date in the 1750's corresponds to the construction of the mansion.³¹

Plate XLIX. — Examples of molded and/or rubbed water table (1-4) and (5) arch bricks (CG 980A) matching those used in the mansion found in the fill of feature CG 980, field 14. Note: brick with corner sawed away (6). (72 NJI 1330-1331)

c. CG 960, Eight Posthole Structure(?)

Eight square postholes were found on the central eastern edge of Field 14, ranging in depth from 1'6" to 2'2". (See Figures 23 and 25 and Plate L). They were arranged symmetrically and had apparently supported round posts 4" in diameter. One hand wrought nail (CG 960) was recovered in the posthole fill, suggesting the possibility that the feature dates in the colonial period; but all other evidence lacking, it is impossible to determine form or function.