western end of the excavation, but towards the middle a large and irregular deposit of oystershells was sealed within the sand at a depth of 4 feet 6 inches sloping upward to 3 feet 6 inches towards the south wall. The shell layer averaged from 6 to 9 inches in thickness and was found to contain many of the more important artifacts.

Figure 6.—Remains of wooden floor boards in the cellar. (Photo courtesy of E. DeHardit.)

On the wooden floor of the cellar lay a thin ½-to 1-inch layer of wood ash, mortar, and occasional brickbats. Had this accumulation been considerably thicker it might have suggested that the building above had been destroyed by fire. But although the presence of this skin of debris could not be explained, it was far from sufficient to support such a conclusion.

The topsoil over the entire area had been disturbed to a depth of at least 1 foot, presumably by deep plowing. Over the cellar fill, humus and a sandy loam extended to a depth of 1 foot 8 inches at the south edge and to 2 feet 1 inch in the middle. The bottom of this stratum contained nothing but late 17th-or early 18th-century artifacts, including an important and well-preserved latten spoon.[16] A small 19th-century disturbance cut into the south cellar edge towards the west end of the excavation, but caused little disturbance to the main fill. Another, much larger, late 19th-century trash deposit had been dug into the fill to the northwest of the chimney and this had reached to a depth of 3 feet 6 inches below the modern grade. The removal of the walnut tree had created a similar disturbance immediately south of the refuse deposit, while a trench for a 20th-century water pipe had cut yet another slice through the same area. None of these disturbances had caused any damage to the lower filling of the cellar.

DATING EVIDENCE FOR THE CELLAR

The majority of the excavated artifacts were scattered throughout the cellar fill and were of similar types from top to bottom of the deposit. These objects included wine-bottle and drinking-glass fragments, potsherds of English and perhaps Portuguese tin-enamelled earthenware, and more that 600 tobacco-pipe fragments, all of them indicating a terminal date of about 1700. A quantitative analysis of the tobacco-pipe stem fragments using the Binford formula[17] provided a mean date of 1698.

Method of Excavation

Digging was initially confined to the immediate vicinity of the chimney foundation (Area B on fig. 3) and to the previously described test trench (A). An east/west trench (D) was next dug to link the two and to isolate the disturbed areas of the tree hole and 19th-century pit in Areas C and G.

Owing to a shortage of labor and the rigors of the weather, it was necessary to confine the digging to small areas which could be completed in a single day's work. Consequently, it was not possible to clear the whole area, as one part would be back-filled during the digging of the next. Mr. and Mrs. Jenkins, the owners of the property, were extremely tolerant of the damage that was done to their gardens, but after the clearance of the large area E, they indicated that the project had gone far enough. Nevertheless, they were persuaded to permit the cutting of another smaller test area to the west (F), but when this, too, failed to find the westerly extremity of the cellar, the project was abandoned. Subsequently, relatives of the owners cut into the exposed north face of area E and extracted a number of potsherds and other fragmentary objects from the sand filling.[18] The undercutting of the bank extended to a distance of 1 foot 6 inches without encountering the north edge of the cellar, thus showing that the total width was in excess of 14 feet.

Extensive probing all around the total area of excavation failed to produce any further traces of the building, though the 1 foot 8 inches of topsoil and sandy loam was found to be bedded on numerous small deposits of oystershells and scattered brickbats. Test holes found that all the located deposits north and west of the existing house had been laid down or disturbed in the 19th century. Five test traverses with a soil resistivity meter west and south of the excavation area produced numerous anomalies which, when checked out, all failed to be associated with the 17th-century cellar. It seemed that the misleading readings were caused by variations in the density and moisture-retaining qualities of the natural sandy clay subsoil.

Early in 1963, while planting a small tree to the south of the existing house, Mr. Jenkins encountered a stratum of oystershells at approximately 8 inches below the present grade. (Fig. 2, Area K.) A series of small test holes was subsequently dug to the south and southeast of the house, and showed that the layer of shells (average thickness 4 inches) overlay the subsoil and was spread over an area at least 15 by 10 feet. A small number of 19th-century pottery fragments were found mixed into the stratum, but the vast majority of the artifacts comprised bottle glass and earthenwares of similar types to those encountered in the cellar hole excavation.[19] The most important item was a pewter spoon handle of late 17th-century character (fig. 15, no. 27) stamped with the initial "M." The presence of this obvious domestic refuse was not satisfactorily explained, but it is concluded that it was originally deposited on the land surface and later disturbed by cultivation.

Figure 7.—Remains of decayed board on floor in front of underhearth. (Photo courtesy of E. DeHardit.)

Landscaping work towards the York River west of the house had yielded a few widely scattered fragments of colonial and Indian pottery as well as numerous 19th-century sherds. The colonial material was predominantly of late 17th-or early 18th-century date, but two sherds of Staffordshire combed dishes were of a type unlikely to date before about 1720. No archeological digging was undertaken in these areas.

Archeological Stratigraphy

Each excavated area was given an identifying letter (fig. 3) and each stratum a number. Thus an artifact marked "B2" was found in the archeological area that contained the chimney and was recovered from the top stratum of sandy loam and clay. It should be noted that not all layers and deposits tabled below were encountered in any one excavation area, while some were confined to single locations.