Excavation in 1984-85 of the 1192 square foot Building t site exposed features associated with two phases of slave quarter construction on Mulberry Row. As expected, mechanical removal of early 20th-century roadwork revealed evidence (F05), albeit scant, of the easternmost of three log dwellings built in the mid-1790s and recorded by Jefferson on his 1796 Mutual Assurance Declaration as Building t. The unanticipated discovery of a cluster of four sub-floor pits (F01-04) representing an earlier barracks-style quarter, named the Negro Quarter, offers a rare insight into the conditions of enslaved laborers living on Mulberry Row between the 1770s and 1790.
In January of 1773, Jefferson purchased three slaves, Ursula and her sons, Bagwell and George. He probably also bought her husband Great George about this same time, but no record of the transaction survives (Bear and Stanton 1995:334). Their third son Isaac Jefferson, born at Monticello in 1775, recalled in 1847 that “the feedin’ place [for Jefferson’s deer park] was right by the house whar Isaac stayed” (Bear, ed. 1967:21-22). This house could well have been the Negro Quarter, located 200’ from the Park gate of the 1770s and 1780s (Jefferson: N131, N132, N221, N130, N522-6). If Isaac did live there, archaeological evidence tells us that his family shared these quarters with as many as three other families, each with access to their own sub-floor pit (F01-04).
On his 1796 Mutual Assurance Declaration, Jefferson described three buildings on the south side of Mulberry Row between the new log stables and the extant 1770s stone Workmen’s House (now called the Weaver’s Cottage):
r. which as well as s. and t. are servants houses of wood with wooden chimnies, & earth floors, 12. by 14. feet, each and 27. feet apart from one another. from t. it is 85 feet to F. the stable [subsequently replaced by the stone stables, which still stands at the eastern end of Mulberry Row]
These three log cabins were constructed no earlier than the winter of 1793-94 and formed part of Jefferson’s preparations to rebuild his house after his first retirement in 1794 (Hill 2002a and b).
In September of 1792, Jefferson wrote from Washington to his overseer Minoah Clarkson:
Five log houses are to be built at the places I have marked out, of chesnut logs, hewed on two sides and split with the saw, and dove tailed...They are to be covered [i.e., roofed] and lofted with slabs...Racks and mangers in three of them for stables. (Boyd 1950-, vol. 24:412-414)
In August of 1793, his son-in-law Thomas Mann Randolph wrote Jefferson that work on the “two houses for the servants are not yet built” but declared his intention to begin construction “as soon as the fall of the leaves commences” (Boyd 1950-, vol. 26:667). By 1796, Buildings r, s, and t were evidently in place: three dwellings, rather than the two originally planned.
Excavation history, procedure, and methods
Exploration of the presumed location of Buildings r, s, and t began in the spring of 1983 with the mechanical removal of the modern overburden: a 1934 paved parking lot; a layer of furnace detritus; and rubble from a 1925 parking lot (Sanford 1995:196). Archaeologists plotted the probable locations of the three buildings using measurements provided by Jefferson’s 1796 Mutual Assurance Declaration. Kelso treated each projected house site and its surrounding yard as a separate excavation. He undertook work on the Building t site last, in 1984-1985.
Four features (F01-04) discovered at the Building t site provided a surprising revelation. These shallow, partially graded, depressions contained fragments of ceramics dating to c. 1770-90, wood, ash, charcoal, and brick. Based on his experience at sites in the Chesapeake Tidewater, Kelso determined that these were the remains of sub-floor pits (F01-04) and evidence of an early, previously unknown slave quarter, which he named the Negro Quarter based on a notation in a Jefferson document (Jefferson: N87, N88).
Exposure of the Building t site revealed that 20th-century roadwork had graded away most of the architectural evidence of Building t. Only the lower portion of its sub-floor pit (F05) remained intact, along with four post holes (F13-16) of the 1809 garden fence. The dating of this fence is well documented, as is the activity of enlarging and leveling the vegetable garden which immediately preceded its construction (Betts 1976: 359-382).
In 1984, excavation proceeded within a 10' X 10' grid oriented on axis with the Monticello mansion. Quadrats of 8' X 8' were opened initially, leaving 2' balks on two sides. Subsequent removal of the balks resulted in the excavation of units of varying sizes, from 2’ X 2’ to 2’ X 10’. Within units, excavation proceeded in natural levels. Opening and closing elevations were recorded, but measurements were not related to a fixed datum point. Excavators used a method of careful troweling to recover artifacts, but they did not use screens.
Summary of research and analysis
Excavation of the Building t site recovered not only the anticipated remains of the documented mid-1790s single-family slave dwelling—Building t—but also features associated with an earlier, undocumented multi-family quarter, which Kelso named the Negro Quarter. DAACS completed recataloguing of the Building t site assemblage in early 2003, but reanalysis of the data has not yet been undertaken.
In addition to the sub-floor pit (F05) associated with Building t, excavation revealed a cluster of four sub-floor pits (F01-F04), which Kelso recognized as the remains of an earlier, barracks-style slave quarter. Ceramics recovered from these sealed contexts bracket the occupation of the Negro Quarter between the early 1770s and c. 1790 (cf. 1776/78-1790 acc. Kelso et al. 1985:37; Sanford 1995:180). Numerous lumps of fired clay and fragments of burnt wood indicate that, by intention or accident, flames destroyed this log dwelling. These artifacts also provide clues about the manner of its construction. Burning preserved pieces of chinking that bear the impressions of the debarked trees used to construct the house’s log cribbing. Once in place, builders apparently crammed clay from the inside against riven clapboards applied to the exterior. Marks of the fingers of enslaved workmen and the reverse mold of overlapping sheathing boards are evident on several large fragments of chinking. Many pieces of burnt clapboard were recovered from the sub-floor pits, including a recognizable portion of a tapered or ‘feathered’ end (Hill 2002a and b).
Kelso presented his conclusions about the appearance of Building s and, by extension, Building t, in an isometric drawing published in 1997 (p. 60, fig. 21). It illustrates a 12’ x 14’, one-room cabin crafted of logs squared off on four sides. The windows of s are depicted with wooden shutters. However, window glass from the sealed context of the sub-floor pit of Building t (F05) and the 1809 postholes (F13-16) indicate that the structure had glazed fenestration during the first ten years of its occupation (Hill 2002a and b). In Kelso’s rendering of Building s, a single door stands in the northern gable end opposite an exterior wattle-and-daub chimney. The sub-floor pit (Building s site, F01) is centered in front of the hearth. In contrast, the sub-floor pit of Building t appears to have been placed off-center within the building. These deviations suggest that, although Buildings r, s, and t may have begun life together, they had distinct occupational histories.
Twentieth-century road work graded away nearly all the structural evidence of Building t. Only the lower few inches of a 3’ X 3’ 6” sub-floor pit (F05) survived. Consequently, conclusions about the appearance of Building t rely on inferences based on the better preserved features of Building s. The Jefferson documents noted above strongly link the timing and manner of the construction of these two dwellings. Sanford (Kelso et al. 1985) initially dated the destruction of Building t to c.1810, twenty years earlier than Building s. He based his estimate on ceramics that had been recovered from the adjacent garden bed excavation prior to the work at the Building t site (p. 24). He later revised his estimate to 1820 (Sanford 1995:181). Preliminary DAACS analysis extends the date of occupation forward another decade to the early 1830s.