Colonial Williamsburg Digital Library

The Native American Sites

The methods and techniques used to carry out Phase III excavations on the three Hampton University sites were structured to obtain the types of data pertinent to addressing several research questions discussed in Chapter 1. The two Native American sites—44HT36 and 44HT37—provided a unique opportunity to broaden the current knowledge of the material and culture history of aboriginal populations in the Tidewater region during the latter part of the Middle Woodland Period.

Phase I and II test excavations performed by the Tidewater Cultural Resource Center at the College of William and Mary revealed subsurface features with few or no cultural remains in the topsoil overburden. When Colonial Williamsburg began the Phase III study, this overburden was therefore stripped by machine until features were evident. A grid was established over the sites, and the features were subsequently mapped and given appropriate numerical designations. All features were sectioned, either in halves or in quarters, depending upon size. A known percentage of the fill from each feature in excess of 2000 ml was retained for wet-screen, flotation sampling, and chemical analysis. The remainder of the fill was screened through one-quarter inch or one-eighth inch mesh hardware cloth. The carbon samples collected were sent to Beta Analytic Laboratories for radiocarbon dating. Artifacts recovered from the sites were generally transferred to the on-campus laboratory for processing at the end of each work day, but items requiring special treatment could be taken to the lab within minutes of their discovery.

The recording method used on the aboriginal sites was the same as that used on other Colonial Williamsburg sites. Features were given consecutive numerical designations as they were mapped. Layers within the features were given context numbers as they were excavated, and all pertinent information relating to the feature or context was recorded on forms designed for computer entry (see Figure 1). This information was later transferred to computer using dBase III-compatible database management software. Major features were photographed in plan and section.

The Historic Site

Excavation and recording techniques employed at HT55 were essentially identical to those used at the Native American sites. Previous archaeological investigations, years of surface collection, and other considerations prompted the stripping of topsoil and plowzone from the site, exposing the subsurface features. The area was gridded into the customary ten-foot squares in order to facilitate mapping of the large site. Again, features were given sequential numbers as they were mapped. Each unit was mapped on a 1:24 scale, producing a composite site plan. All features were sectioned and a portion of the fill retained for chemical, flotation, and wet screen sampling. Most features were photographed in plan and section.

Flotation
(Charles Thomas and Lucie Vinciguerra)

Soil samples recovered from the Hampton University sites were transported to the Department of Archaeological Research facilities in Williamsburg for flotation. The flotation device used in processing soil samples consisted of two main sections. The base was a 15-gallon cylindrical NALGENE tank with two holes drilled 8 Figure 1. Context record. 9 through the wall near the base. One hole was fitted with a plug to serve as a drain for emptying the tank between samples, and the other with an L-shaped length of copper pipe running upward through the bottom of the tank. An adjustable spray nozzle was attached to this pipe.

The upper portion of the device was constructed of a 20-quart utility tub, with a shallow, rectangular hole cut near the upper edge, and a broad, flat spout attached. The bottom of the tub was removed and replaced with a 24-by-24 brass mesh cloth (wide diameter: 0.014"), which retained all particles greater than 0.0277" (0.70 mm). The tub was selected to create a snug, water- tight fit when pressed down into the NALGENE tank.

Aluminum baking pans with bottoms removed and replaced with 40-mesh brass wire strainer cloth were used for collection of the floated material. These would collect particles of at least 0.0175" in diameter.

Water was run into the base tank from the bottom until it filled the NALGENE tank and spilled over the spout. Then, at high pressure, a stream of water was directed through the screen to create turbulence. The soil was sprinkled into the sample bucket, the lighter fraction of which was carried away through the spout and into the collecting tray. The sample could be agitated slightly by hand to break up any highly coherent soil, if necessary. The process continued until there was no further suspended material visible in the flow to the collecting tray. The tray was then set aside for air drying while the heavy fraction was retained in the screen fitted in the NALGENE tank. The water was drained and the sample extracted between each process.

This method of collecting flotation samples minimizes the handling of delicate material, which is especially fragile when wet. The standard poppy seed test, i.e., placing 100 poppy seeds in a sample before it is processed and counting them after flotation, produced an average recovery rate of 95.5%. This test was run periodically in order to monitor the conditions of the screens, adhesives, and fittings.

Chemical Analysis

Soil samples taken from the three sites were analyzed by the Virginia Department of Agriculture facility at Blacksburg for the following elements: potassium (K), phosphorus (P), calcium (Ca), and magnesium (Mg). Soil pH and the amount of humus were also measured.

Although the constituent element analysis was very revealing, the latter two tests revealed little of major archaeological interest. All of the samples taken were slightly acidic (Figure 2), both the control samples and those taken from specific features. However, the pH value of the soils is significantly affected by rainwater and modern farming practices, and it is likely that these are the principal factors being measured. The pH value of the soil simply gives an archaeologist an idea of how well some items, such as bone, shell, and metals, will be preserved in the ground. The slightly acidic conditions suggest that all of these materials were adversely affected, but probably not drastically so.

The presence of high levels of potassium in the soils of features is generally an indicator of the presence of wood ash. Phosphorus, on the other hand, seems to indicate the presence of human waste material and is usually extremely high in historic- era privies. High calcium levels are generally indicative of the presence of decaying shell or bone within a feature or layer (Pogue 1989). Large amounts of calcium will remain in the soil after the source has seemingly disappeared. Although the samplers were routinely tested for magnesium, which may be important for the health of some crops, its archaeological significance is yet unknown.

Description of Features

Feature 4/5/6 (see Figure 5) was a roughly circular pit measuring 3.1' by 3.3'in plan and extending 1.7' below subsoil level. The walls of the feature were relatively straight sided and the base gently rounded. Three layers of fill could be discerned within the feature.

Layer 4, the uppermost deposit, consisted of brown sandy loam with some charcoal inclusions and was about 1.0' thick. This layer contained the vast majority of artifacts recovered from the feature. Included were 42 shell tempered sherds: 22 cord marked, 18 knotted net impressed, 1 plain surfaced, and 1 unidentified. A minimum of 21 Figure 5. Feature section drawings—44HT36. 22 five vessels are represented in the assemblage: 3 cord marked and 2 knotted net impressed. Four highly-weathered sherds were also recovered. The layer also contained a fired coil of a fine, sandy paste ceramic, 1 quartzite flake, 1 quartzite mano, 4 fire-cracked rocks, and a quartzite cobble. One seed identified to Gramineae (Grass family) was also recovered.

Layer 5 was the second layer of fill in the pit, consisting of about 0.4' of light brown sandy silt mounded slightly higher in the center of the feature. The layer contained 4 sherds, all shell tempered: 3 knotted net impressed and one cord marked.

The bottom layer of the pit, Layer 6, consisted of tan sandy clay about 0.3' thick. Included among the fill were 3 shell tempered sherds: 2 knotted net impressed and 1 plain. One small fragment of quartzite with water-worn surfaces was also recovered.

Feature 7 was an oval pit, 2.0' by 1.5'. The feature had steeply sloping sides extending to a nearly pointed bottom 1.8' below subsoil level. Fill consisted of greyish-brown sandy loam which contained no artifacts.

Feature 13/14 (see Figure 5) was a circular pit 3.0' in diameter and 1.3' deep with sloping sides and a flat bottom. The upper layer of the feature, Layer 13, extended to 0.7' below subsoil and was filled with brown sandy loam. The layer yielded 9 shell tempered sherds: 6 cord marked, 2 knotted net impressed, and 1 unidentified, representing a minimum of three vessels. Three quartzite flakes were also recovered. The bottom of the feature was filled with a pale brown sandy clay which contained no artifacts. Layer 14 was 0.6' thick.

Feature 19/20 (see Figure 5) was a roughly ovate pit measuring 2.6' by 1.1'. The walls sloped abruptly to a flat bottom 1.2' below subsoil. The main feature, represented by Layer 20, was filled with mottled orange clay containing charcoal flecks but no artifacts. Artifacts were recovered from a pocket of fill, Layer 19, lying at the east end of the pit. Approximately 1.0' in diameter and 1.0' deep, Layer 19 consisted of dark grey-brown loam. Three sherds of shell tempered, cord marked ceramic representing one vessel were contained in the fill.

Feature 22 was a circular pit with three layers of fill which may represent a tree hole. The pit was 3.4' in diameter and extended to an overall depth of 1.5' below subsoil. Root disturbances were noted at the base of the feature on its northern side. In profile, the pit had rounded sloping walls and a rounded base. The central fill of the feature consisted of a pocket of dark brown sandy soil 1.4' in diameter and 1.0' deep. A layer of grey-white, dry, compact soil extended under this fill on one side of the feature. The bottoms and edges of the pit were filled with an orange-grey sandy clay with some charcoal flecks. No artifacts were recovered from the feature.

Feature 25 was a small circular pit, 1.5' in diameter and 0.9' deep. The walls of the feature sloped inward to a rounded base. Fill consisted of a grey-brown sandy loam which contained no artifacts.

Feature 26/30 was a circular pit, 2.5' in diameter containing a smaller pocket of fill. The main feature, Layer 30, had roughly straight walls and a flat bottom lying 0.55' below subsoil level. Fill consisted of brown sandy loam mottled with orange clay. Layer 26 was a pocket of dark brown mottled loam with flecks of charcoal situated roughly in the center of the main pit. The deposit was 1.3' in diameter and extended to the bottom of the main pit. Neither layer of fill contained any artifacts.

Feature 27 was a circular feature, about 2.0' in diameter and 0.55' deep. The pit contained an upper layer of dark greyish brown sandy loam 1.2' in diameter and 0.25' deep. Below this layer was mottled tan and brown sandy loam. No artifacts were recovered from the feature.

Feature 33/36/37 was a rectangular pit measuring 2.3' by 3.9' and containing three layers of fill. The walls of the pit were straight-sided, terminating at a flat base 1.35' below subsoil level. The upper layer of the pit, Layer 33, consisted of 0.2' brown sandy loam with some charcoal flecking. Below this was Layer 36, consisting of 0.6' 23 of light sandy silt with some pebbles. The bottom layer, Layer 37, consisted of 0.55' of tan sandy clay. One fire-cracked rock was recovered from the uppermost layer of the pit.

Feature 34 was a circular pit, 2.5' in diameter, consisting of 0.7' of greyish sandy loam fill containing only 1 shell tempered, cord marked sherd. The walls of the feature sloped gradually from south to north and abruptly from north to south.

Feature 38/40/41 (see Figure 5) was a circular pit 2.2' in diameter and 1.45' deep with three layers of fill. The walls of the pit were nearly vertical and the bottom was flat. The uppermost fill, Layer 38, consisted of 0.5' of dark brown sandy loam with some charcoal flecking. Next was Layer 40 which consisted of 0.65' of tan sandy loam with shell and charcoal inclusions. The bottom layer, Layer 41, was 0.3' thick and composed of orange sand clay. Artifacts were found only in the upper two layers of fill. Layer 38 contained 2 shell tempered sherds, 1 cord marked and 1 knotted net impressed, and one quartzite flake. Layer 38 yielded a single quartzite flake.

Feature 42/47/48/122 (see Figure 5) was originally identified during Phase II test trenching, at which time approximately one-quarter of the feature was excavated. The remainder of the feature was removed during the Phase III excavation. The feature was circular in plan, about 4.7' in diameter at the surface of subsoil. The walls of the feature sloped inward from subsoil level to a depth of 1.3', after which they descended vertically to a depth below subsoil of 3.8'. The lower portion of the feature was roughly 2.5' in diameter with a flat bottom.

Four layers of fill were discerned within the pit. Layer 42 varied in depth from 0.4' to 1.2' below subsoil level, sloping from the sides to the center. The fill consisted of grey-brown sandy loam with some charcoal flecking. The layer contained 6 shell tempered sherds: 3 knotted net impressed, 1 cord-marked, 1 plain, and 1 with an unidentified surface treatment. Another sherd with a fine sandy paste and knotted net impressed surface was also recovered. Two seeds from Nyssa spp. (Black Gum) were recovered in the flotation sample from the layer.

The next layer in the pit, Layer 47, existed only on the eastern side where it was about a foot deep, sloping towards the middle. The fill consisted of dark brown sandy loam with orange sandy clay mottling. This layer contained 3 sherds, all of which were shell tempered. Two were cord-marked and the other unidentified. Layer 122 sloped from the west, under Layer 47, almost extending to the eastern side. It was about 0.7' deep, consisting of dark brown sandy loam.

At the bottom of the pit was a layer of brown sandy loam with charcoal flecks approximately 1.6' deep. Layer 48 yielded 27 shell tempered sherds, all cord marked, which represent a minimum of six vessels. A large, flat, quartzite cobble 0.8' long was found lying horizontally about 0.4' from the base of the pit (Photo 3). The cobble had two small flakescars but otherwise exhibited no evidence of use. A radiocarbon date of A.D. 20 ± 70 years (1,930 B.P., Beta-25913) was derived from charcoal recovered from this layer.

Feature 43/45 was a circular pit approximately 2.5' in diameter and 1.2' deep with two layers of fill. In profile, the pit had steep sloping sides and a rounded bottom. Layer 45 was a 1.3' diameter circular concentration of charcoal in a dark brown sandy loam matrix which extended from the top of the pit at subsoil level to a depth of 1.0'. Layer 43, which surrounded Layer 45 and underlay it, consisted of orange sandy fill with Photo 3. Section/excavated plan of Feature 42. 24 some charcoal flecking. No artifacts were recovered from the charcoal concentration, but Layer 43 contained 10 shell tempered ceramic sherds, 9 cord-marked and 1 unidentified. A minimum of two vessels are represented in the collection. A fragment of fire-cracked rock was also recovered.

Feature 46 was a circular pit, 2.8' in diameter. The fill consisted of grey/tan sandy loam with some charcoal flecking. The sides were steeply sloping to a narrow, but flat bottom, 3.2' below grade. It is possible the feature is a tree hole. One shell tempered, knotted net impressed sherd was recovered from the fill.

Feature 49 was a small, 1.8' diameter pit some 0.6' deep. Its fill consisted of dark brown sandy loam with some unidentified pottery recovered near the surface.

Feature 51 was a 3.0' diameter circular pit about 0.7' deep. The fill consisted of dark brown sandy loam. The sides of the pit sloped gently to a rounded bottom. The pit contained 16 shell tempered sherds: 12 cord marked, 3 knotted net impressed, and 1 unidentified. One vessel is represented by the net impressed sherds, with a minimum of three vessels represented by those which are cord marked. The sherds from one cord marked vessel and the net impressed vessel are burned, as is an additional unidentified sherd recovered from the pit.

Feature 52 was a circular pit with overall dimensions at subsoil level of 2.0' in diameter. Only one layer of dark brown loam fill could be discerned in the pit, although the feature narrows to 1.1' in diameter at 0.9' below subsoil level. No artifacts were recovered.

Feature 60 was a small, roughly circular pit, 1.3' in diameter and consisting of 0.5' of dark brown sandy loam. The sides of the pit were straight and its bottom flat. Some charcoal inclusions were noted. The pit contained 6 shell tempered sherds, all cord marked, representing a maximum of 3 vessels.

Feature 62 was an irregular, somewhat circular pit, 2.2' in diameter and 0.5' deep, filled with dark brown sandy loam. The sides of the feature were vertical and the bottom flat. One shell tempered, cord marked sherd was recovered from the fill.

Feature 69 was a circular pit, approximately 2.0' in diameter and 0.5' deep, filled with dark brown sandy loam. The sides of the pit were nearly vertical and the bottom was flat. The fill contained 9 shell tempered sherds, 8 of which were knotted net impressed and 1 plain. One recently shattered fragment of quartzite was also recovered. All of the sherds from this pit could have originated from one vessel, with the plain sherd coming from near the base.

Feature 71 (see Figure 5) was a large circular prehistoric pit with a diameter of 2.75'. The feature was 2.7' deep, consisting of dark brown sandy loam fill. The sides sloped steeply to a rounded bottom. No stratigraphy could be detected in the pit. The fill contained 43 shell tempered sherds: 31 knotted net impressed and 12 cord-marked. Analysis of the ceramics suggested that a minimum of 7 net impressed and 4 cord marked vessels are represented in the collection. Lithic materials recovered include a quartzite shatter fragment, 4 fragments of fire-cracked rock, and a burned cobble. One Galium (Bedstraw) seed was also recovered.

Feature 73/75 was an oval pit about 4.0' long and 2.1' wide. The northern end of the feature lay outside of the excavation unit, so exact dimensions are not known. The sides of the pit were vertical with a very flat bottom about 0.6' below subsoil level. The upper layer of the feature, Layer 73, consisted of dark brown loam mottled with tan sand and silt. Some charcoal inclusions were noted. The layer was about 0.3' thick and contained 22 sherds of shell tempered prehistoric ceramics, 21 knotted net impressed and 1 cord marked, and 1 small quartzite biface (a side-notched point preform). All the net impressed sherds recovered from this feature probably originated from a single vessel. Historic artifacts recovered from Layer 73 included 1 fragment each of coarseware and pearlware, 1 nail, bricketage, slag, and coal. The lower 0.5' of the pit was filled with Layer 75, a tan sandy clay with 25

Feature 76 was a circular pit, 1.9' in diameter, consisting of brown sandy loam with charcoal flecking. It was about 1.1' deep, with almost vertical sides and a flat bottom. Several large, about 1", charcoal inclusions were noted in the fill, but no artifacts were found.

Feature 77 was a small circular feature about 1.8' in diameter, with a small pocket of oyster shell in the center at subsoil level. The pit was 2.8' deep and had irregularly vertical sides sloping to a nearly pointed bottom. The pit may have originated as a tree hole. There were no finds.

The plan of Feature 79/81 was obscured by a natural disturbance, but an oval feature approximately 3.0' by 1.5' was suggested. The feature was filled in two levels. The main fill, Level 81, was a mixture of very dark grey/brown sandy loam, orange clay, and tan sandy loam extending to a depth of 1.6'. The walls of the feature were irregular; the bottom basically flat. The fill contained 13 shell tempered, cord marked sherds, representing a minimum of 2 vessels, and 1 quartzite flake. Layer 79 was a small (1.4' by 1.0'), irregularly ovate deposit of dark grey fill, containing 4 shell tempered, cord-marked sherds, all from one vessel, and 1 quartzite flake. The deposit was situated roughly in the center of Layer 81 and extended from the surface of subsoil to a depth of 0.6'. The soil disturbance noted around the periphery of Layer 81 suggests that Layers 77, 79, and 81 may represent separate layers of fill within a large tree hole which once existed in grid unit S120 E70.

Feature 86/90 was a 3.1' by 3.7' oval, basin-shaped pit 0.8' deep. The main body of fill, Layer 86, consisted of dark brown sandy loam containing many charcoal fragments, but no artifacts. The east side of the feature sloped irregularly, whereas the west side was steep and regular. The bottom was irregular as well, with numerous tree root holes extending from the base of the main pit. Layer 90 was a small (0.75' diameter), conical deposit contained within Layer 86. The fill consisted of brown sandy loam and extended from the surface of subsoil to a depth of 1.9'. Small flecks of charcoal and ash were noted in the fill, but no artifacts were recovered. Feature 90 probably represents the remains of a tap root associated with the tree hole represented by Feature 86.

Feature 88 was an irregular pit, probably a tree hole, measuring 4.4' by 4.1' in plan and 1.1' deep and consisting of grey/tan sandy fill. The sides sloped gently to an irregular bottom. The pit contained 7 shell tempered sherds, 6 knotted net impressed and 1 unidentified, which were probably recovered from a shallow lense of darker fill in the southern half of the feature at subsoil level.

Feature 109 probably represents another tree hole. The feature consisted of large soil stain, irregular, but approximately 7.0' in diameter. A portion of the feature was excavated, revealing 27 about 2.0' of brown sandy loam on the east side. No artifacts were recovered.

Feature 116 was an ovate feature approximately 2.2' by 2.8' in plan, filled with dark brown sandy loam. The feature was about a foot deep. The north side of the pit sloped very gently to a rounded bottom, whereas the south side was steeper. Some charcoal inclusions were noted. The fill contained only 1 shell tempered, knotted net impressed sherd.

Feature 118 is likely the remains of a tree hole. The feature consists of a deposit of dark brown, sandy loam 6.8' by 4.7' in plan and 1.2' deep. The bottom of the feature is highly irregular, shallow at one end and deep at the other. The fill yielded one shell tempered, cord marked sherd, one knotted net impressed sherd with a fine sandy paste, and a cobble core of quartzite.

Feature 120/123 was a large pit, probably a tree hole, with two layers of fill. The core of the feature, Layer 120, was a roughly circular deposit of grey-brown sandy loam 3.0' in diameter and extending to a depth of 2.6'. The sides of the deposit sloped steeply to a flat bottom. This fill contained 1 shell tempered, cord marked sherd and four quartzite cobbles, three of which showed signs of having been burned. Two of the cobbles had flakes removed, probably a result of testing the quality of the lithic material. Layer 123 of the feature was an outer ring, about 4.8' in diameter, of brownish orange sandy clay containing some charcoal flecking. The walls of this deposit also sloped steeply and did not extend beyond the depth of Layer 120. No artifacts were recovered from this fill.

Feature 124 was a very irregular feature consisting of two layers. Part of the feature extended beyond the excavation area, so only the north-south dimensions (2.0' to 4.4') are known. Layer 124 consisted of 0.3' of light brown sandy loam and Layer 126 of 0.4' of light brown loam. There were no finds.

Soil Chemistry (Figure 6)

At HT36, 10 samples showed significant amounts of potassium well above the control sample amounts. Layer 47, part of a large prehistoric pit also containing Layers 42, 122, and 48, tested highest in potassium content, with 48 ppm (parts per million) more than the control sample, which had an average of 22 ppm. All of the layers contained prehistoric ceramics and charcoal flecking. There was no visible evidence of ash, but the primary components may have been leached out of the soil over the past 2000 years. Feature 20, a small pit also containing charcoal and pottery, tested high for potassium as well.

Only one feature, Feature 71, contained significantly high amounts of phosphorus, measuring 80 ppm above the control (6 ppm). The pit was about 2.5' in diameter and over 2.5' deep. Pottery and cobbles were found near the surface.

Again, only one feature tested high for calcium and was the only feature containing more calcium than the control samples. The feature, Feature 120, contained only a few fire-cracked rocks, ceramics, and a few charcoal flecks, but no shell or bone. The feature, which was probably a tree hole, may have contained a large amount of animal bone or shell, which may account for the high reading.

Description of Features

Feature 1001 was a roughly circular, 2.0' diameter prehistoric pit containing human bone in its upper level (Photo 5; Figure 8). The fill consisted of brown sandy loam about 0.5' deep with inclusions of shell. The feature was bowl-shaped, with rounded sides and bottom. The human remains contained within the feature represent a secondary burial involving at least two individuals. Recovered from the pit were two right femurs, one from a juvenile and one from a mature individual, Photo 5. Detail of human bone in Feature 1001. and a right tibia which probably belongs to the individual represented by the larger femur. A few other unidentifiable fragments of bone were also recovered. The feature fill contained no other artifacts.

Feature 1002/1005 was a small circular prehistoric pit, about 1.8' in diameter. No depth was recorded. Fill consisted of dark grey sandy loam containing snail shells, oyster shell, animal bone, 1 shell tempered, cord-marked sherd, and a small amount of carbonized wood. The feature appeared to have been filled in at least two levels. Layer 1002, which comprised a core area about 0.5' in diameter, appeared to have been deposited after the fill of Layer 1005, which contained the majority of shell, had settled.

Feature 1003/1018 (see Figure 8) was a small, circular prehistoric pit roughly 2.0' in diameter and 1.2' deep with sloping sides and a rounded base. The inner fill of the feature, Layer 1003, was a pocket-shaped deposit, 1.3' in diameter and 0.7' deep. This dark, greyish brown sandy loam contained many oyster shells as well as 13 sherds of shell tempered, knotted net impressed ceramic representing a minimum of two vessels. A radiocarbon date of A.D. 40 ± 80 years (1,910 B.P., Beta-25914) was derived from charcoal in Layer 1003. One Galium (Bedstraw) seed, one Quercus spp. (Oak) acorn cap, and 1 Carya spp. (Thickshell Hickory) nut shell were identified in the analysis of a flotation sample from the feature. The lower fill of the pit, Layer 1018, 32

Figure 8. Feature section drawings—44HT37.

Feature 1006 was a small, circular pit, roughly 2.0' in diameter consisting of medium brown sandy loam. It was about 0.5' deep, with irregular sides and bottom. One fragment of fire-cracked rock was recovered.

Feature 1007/1015 was a large circular pit 5.6' in diameter and 3.4' deep, which likely represents a tree hole with various fill deposits. The central part of the feature (Layer 1007) was filled with a dark grey-brown sandy loam which graded to a medium grey-tan sandy loam. This deposit was bowl-shaped with steeply angled sides and a rounded bottom. It contained both prehistoric and historic artifacts. Prehistoric artifacts included 35 shell tempered sherds (19 cord marked, 12 knotted net impressed, 3 plain, and 1 unidentified), four low-fired, untempered lumps of clay, and one quartzite flake. One fragment each of colorless non-leaded glass and brick were also recovered. Also present in the fill were 1 cf. Cyperaceae (Sedge family) seed, 1 Galium (Bedstraw) seed, and some carbonized wood. Layer 1015 was a small, semi-circular deposit of darker fill and oyster shell contained within Layer 1007. The deposit was 1.6' long and extended from subsoil level to a depth of 0.6'.

Feature 1021 was a small, circular deposit of dark grey sandy loam about 1.1' in diameter which appeared to be contained within a larger 34 tree hole disturbance. Within the center of Layer 1012 was a pocket of charcoal (Layer 1022). Neither deposit was excavated.

Feature 1023/1035 was a very large irregularly-shaped feature measuring 8.2' north-south and 5.3' east-west, and consisting of dark brown sandy loam. The feature appears to be a tree hole. Two layers of fill were excavated. The uppermost layer, Layer 1012, was 0.5' deep on its north end, dropping to 2.8' just south of its center, then back up to 0.5' at the south end. One fire-cracked rock spall was recovered. Layer 1035 was the second layer in the large pit, consisting of dark brown sand with many pebbles and some charcoal flecking. At about 0.8' below subsoil this deposit extended 6.0' in diameter. The north wall of the feature sloped to a flat bottom, whereas the south wall was almost vertical. The deposit extended to a total depth of 5.8' below subsoil level.

Feature 1024 (see Figure 8) was a circular prehistoric pit, 2.8' in diameter, consisting of dark brown sandy loam and containing flecks of charcoal, shell, and pebbles. The east side of the pit sloped at about 40°, whereas the west side was nearly vertical. A small, flat bottom was found at a depth of 2.8' below the surface. The feature contained 7 shell tempered sherds: 6 knotted net impressed and one cord marked. A radiocarbon date of A.D. 300 ± 70 years (1,650 B.P., Beta25915) was derived from charcoal samples obtained from this feature. Paleobotanical remains from the feature were quite significant and included the following: a pine cone fragment (Pinus spp.), carbonized wood, 1 thickshell hickory nut fragment (Carya spp.), 1 nut fragment of the Walnut family (Juglandaceae spp.), 1 bedstraw seed (Galium spp.), 1 seed of the Smartweed family (Polygonaceae), and 1 possible maize kernel fragment (Zea mays).

Feature 1026/1087/1089 appears to have been a large tree hole with several layers of fill intruded by a recent drainage ditch. As represented by Layer 1089, consisting of light to medium brown sandy loam, the main feature was 7.0' in diameter and 3.5' deep with a deeply rounded profile. Within this depression were two smaller deposits. Layer 1087, a fill deposit of dark grey and medium brown sandy loam, was 2.0' in diameter and 3.4' deep with steeply sloping sides. A smaller deposit (Layer 1026) of very dark grey sandy loam with oyster shell was contained within the upper 0.7' of Layer 1087. Layer 1026 was 1.0' in diameter. The only deposit within the feature to yield artifacts, Layer 1026 contained 1 shell tempered, knotted net impressed sherd, 1 quartz flake, and two fragments of polished animal bone representing either tools or ornaments.

Feature 1030 (see Figure 8) was roughly circular prehistoric pit, 2.7' in diameter, containing brown sandy loam fill with charcoal flecking. The sides of the pit sloped steeply from the east and gradually from the west to a depth of 2.9'. Artifacts recovered include 1 shell tempered, knotted net impressed sherd and 3 chunks of fired, unshaped sandy paste ceramic.

Feature 1036 was a small, 1.8' diameter, circular pit consisting of brown sandy loam. It was irregularly bowl-shaped with a depth of 0.9'. There were no finds.

Feature 1039 was probably roughly circular; however, the recent drainage ditch appears to have removed half the feature. The original feature was probably about 3.2' in diameter and may represent a tree hole. The profile indicates it was bowl-shaped and 1.1' deep. The greyish-brown sandy loam fill contained no artifacts.

Feature 1042 was a large pit uncovered along the northern edge of the excavation area. Fill consisted of light brown gravelly sand. The portion of the feature excavated revealed a profile approximately 4 feet deep with steep sides and a rounded bottom. There were no finds.

Feature 1043 was an ovate pit measuring 2.0' by 1.5' and consisting of brown sandy loam containing some charcoal flecking. Sides were bowl-shaped with a rounded bottom 0.8' below subsoil level. The feature contained 5 shell tempered sherds, 4 of which are knotted net impressed. The remaining shell tempered sherd showed an unfinished exterior surface with scrape marks over unwelded joins between coils. The 35 feature also yielded one sherd of a fine, sandy paste ceramic with knotted net impressed surface; one fired, unshaped lump of sandy clay; and 2 fragments of fire-cracked rock.

Feature 1054 was a 2.0' diameter, circular pit consisting of medium brown sandy loam with oyster shell inclusions. The walls of the feature sloped slightly to a rounded base 0.7' below the surface. Ceramics recovered from the feature included 15 shell tempered sherds, 14 knotted net impressed and one cord marked, likely representing a maximum of three vessels. An additional shell tempered, cord marked vessel is also represented (the vessel collapsed into numerous tiny fragments in excavation). Two fragments of fire-cracked rock and some carbonized wood was also recovered from the feature.

Feature 1059 was partially obscured by the adjacent baulking. The feature was probably circular, approximately 2.3' in diameter. The sides sloped steeply to a narrow, flat bottom about 2.0' below grade. The fill consisted of grey-brown sandy loam. Inclusions of charcoal flecks were recorded. The pit contained 1 shell tempered, knotted net impressed sherd.

Feature 1062 was a circular pit about 2.2' in diameter. The fill consisted of medium brown sandy loam. The feature was bowl-shaped, with sloping sides and a rounded bottom some 1.2' below grade. Two shell tempered, knotted net impressed sherds were recovered, each representing a single vessel.

Feature 1080 was roughly circular, 1.4' by 1.5' consisting of brown sandy loam with concentrated oyster shell. The pit was 0.7' deep.

Feature 1093 was roughly circular stain, 3.0' in diameter consisting of brown sandy loam. It was not excavated.

Soil Chemistry (Figure 9)

Only one feature, Layer 1002 in Feature 1002/ 1005, had a notably high potassium content. The layer was a small pocket of fill, ca. 1.0' in diameter, containing some oyster shell and charcoal. The same feature also contained significant amounts of phosphorus, indicating the possibility that this was a place where human wastes were deposited.

Several features at HT37 contained significant calcium levels: Feature/Layers 1002, 1003, 1015, 1054, and 1087. Calcium can indicate the presence of decaying shell or bone, and all of the features or layers contained oyster shell in abundance, probably leading to the high readings.

Ceramic Artifacts

Ceramics were overwhelmingly the most abundant type of prehistoric artifact recovered in the excavations, although the entire collection from all three sites totaled only 571 sherds. Apart from radiocarbon dates derived on charcoal from three features at HT36 and HT37, ceramics proved the most informative of the recovered artifacts in determining the cultural affiliation or date of occupation of the three sites. In general, the ceramics indicated a late Woodland I (2000 B.C.-A.D. 1000) date for sites HT36 and HT37. The excavated area designated as site HT55 was largely dominated by historic features relating to a colonial-period occupation. The ceramics associated with the few prehistoric features identified within this area, as well as those incorporated into colonial contexts, are evidence of earlier prehistoric occupations dating from the late Woodland I Period and the Woodland II Period (A.D. 1000-1600).

Study of the ceramic collections from the three sites began by sorting the sherds by size. Only those sherds measuring greater than 2 cm in any direction across the interior or exterior surface were examined further. Analysis proceeded by sorting sherds into groups defined by attributes of temper, paste, and surface treatment. Other attributes relating to manufacturing techniques, such as rim profile and preparation and vessel shape, were also noted. Vessel size was assessed through an estimate of rim diameters obtained by 36 Figure 9. Soil chemistry—44HT37. comparing the curvature of rim sherds against a template on which concentric circles in increments of 2 cm in diameter had been drawn. The collection included very few rim sherds, however, and those fragments which could be measured commonly represented no more than five to ten percent of the total vessel rim. Only one of the sherds examined bore any decoration other than cord or net marking on the lip, an attribute which is probably more a by-product of shaping the rim with a wrapped paddle than a result of intentional decoration.

Frequencies of wares and types discussed below are based on sherd counts made prior to mending (Table 3). Mends within features were noted in the artifact catalog, but no attempt was made to mend between features. Estimates of the number of vessels represented within each fill layer are provided in the descriptions of features for HT36 and HT37. It should be noted that examination of the sherds suggested, in general, that very few vessels were represented in individual features.

Three ceramic wares, comparable to those previously defined for the Virginia Coastal Plain, were identified in the collections. These included Mockley Ware, a shell tempered late Woodland I Period ceramic characterized by cord marked, net impressed, and plain surfaces (Stephenson and Ferguson 1963:105-109); Townsend ware, dating from the Woodland II Period and characterized by a shell tempered paste and fabric impressed surfaces (Blaker 1963:14-22); and Roanoke Simple Stamped, another shell tempered ware, dating from the late prehistoric and early Contact period (Blaker 1952; Harrington 1948). 37 Photo 6. Site 44HT36 during excavation.

The vast majority of sherds recovered from the three sites relate to the Mockley series. At HT36, where a total of 245 sherds were recovered in the excavations, 238 were classified as Mockley ware. Of these, surface treatments on 52.9% were cord marked, with 42.9% knotted net impressed, and 1.7% plain, smoothed surfaced (it is likely that the plain surfaced sherds from both HT36 and HT37 represent vessel sections from just below the rim or near the base). Surface treatment on 2.5% of the Mockley sherds was unidentifiable. The remainder of the collection from HT36 consisted of two sherds with a fine sandy paste and knotted net impressed surfaces, and five sherds whose poor condition prevented classification.

The assemblage recovered from HT37 was also dominated by Mockley ceramics. Of a total of 107 sherds, 104 were classified as Mockley ware (Figure 10). Three sherds with a fine sandy paste and knotted net impressed surfaces were also recovered. Among the Mockley ware, 22.1% were cord marked, 69.2% knotted net impressed, 6.7% plain, and 1.9% unidentified.

The collection from HT55 is more diverse (see Figure 10), although Mockley ware again predominated. Of a total of 205 shell tempered sherds, 158 (77.0%) were classified as Mockley ware. Of these, 63.3% were cord marked and 36.7% were knotted net impressed. Seven sherds recovered were assignable to the Townsend series based on their shell tempered paste and fabric impressed surfaces. Seventeen sherds were classified as Roanoke Simple Stamped. Of the remaining shell tempered sherds, five were plain surfaced and eighteen had badly eroded surfaces and could not be definitely assigned to a ware group. The remainder of the ceramic collection from HT55 was comprised of five sherds with a sandy paste, four of which were cord marked and one unidentifiable. Two additional sherds were not assignable to any ware group.

The Mockley ceramics from the three sites conformed fairly well with Stephenson and Ferguson's (1963:103-109) original definition of the ware based on an assemblage from the Accokeek Creek site on the Potomac River in Maryland, as well as with subsequent descriptions of occurrences within the Coastal Plain of Virginia (Egloff and Potter 1982:104; Photo 7). The sherds commonly fractured along joins between coils. Coil manufacture was also indicated by one sherd recovered from HT37 (Feature 1043), with a plain surface and scrape marks which had not completely obliterated the joins between coils. Rim profiles and rim diameters suggested that both jar and bowl forms were present, although only jars were found in the assemblage from HT37 (Table 4). Of four basal sherds recovered, two were subconoidal, one conoidal, and one rounded with a slight flattening. The presence of flat bottomed vessels similar to the "beakers" described by Painter (1977; see also Egloff et al. 1988:28-32) was possibly suggested by one knotted net impressed sherd from HT55 which appeared to originate from the base of the vessel wall, a position where these types of vessels commonly have a thickened, ring-like bulge. Rims on the Mockley ceramics from Hampton University were sometimes thinned or beveled, with the lips either smoothed or marked by cord or net impressions. Vessel wall thickness measured on rim sherds ranged from 0.5 to 0.95 cm. Interior scoring was noted on 13% of the sherds from HT36 and HT37 and less than 1% (one sherd) of those from HT55.

The group of sherds which was identified as Mockley ware within the three assemblages from Hampton was quite variable in paste composition. A small number, about five percent, of the Mockley ware sherds from HT36 and HT37 were characterized by shell temper in a silty paste with no sand admixture. The remainder of Mockley 39 Figure 10. Prehistoric pottery—44HT36, 44HT37, and 44HT55.

Photo 7. Examples of Mockley ware.

Mockley ware has been described elsewhere as containing, in addition to shell temper, varying amounts of sand, limonite, and organic matter in the paste (Stephenson and Ferguson 1963:104; Egloff and Potter 1982:103). Sherds of mixed sand and shell were noted in assemblages from the Croaker Landing site (44JC70, 44JC71) on the York River (Egloff et al. 1988:26), and sand and organic matter were described as common inclusions within the Mockley ware recovered from James Madison University's (JMU) 41 excavations at the Addington (44VB9) and Sherwood's Forest (44VB92) sites at Great Neck in Virginia Beach (Smith and Andrews 1986). It was also noted by JMU archaeologists that shell was a common inclusion in the paste of those ceramics from Great Neck identified as Mount Pleasant ware, a sand tempered, late Woodland I Period ceramic with cord marked, net impressed, fabric impressed, and plain surfaces defined by Phelps (1983:32).

The meaning of this type of variability in paste is unclear at present. Egloff (Egloff et al. 1988:26,75, 78) concluded that the sherds from Croaker Landing with both shell and sand in the paste most likely represent local variation in the clay source used to produce Mockley ceramics, although he suggested that the paste variety may represent an early temporal phase within Mockley ware. In the Outer Coastal Plain of southeast Virginia, assessment of the meaning of variability in paste is complicated by the fact that shell temper and sand temper traditions have been suggested to have had distinct distributions during the late Woodland I Period. Phelps (1983:32) has noted that Mount Pleasant ware is distributed on inland and coastal sites in the northern half of the North Carolina Coastal Plain. Mockley ware has been found in this region on only a few sites along the Chowan River, and shell temper appears as a common trait only within the Woodland II Period.

Because of variability in paste, differentiating between Mount Pleasant and Mockley wares in southeast Virginia is complicated. The two wares are apparently contemporaneous. Radiocarbon dates on Mount Pleasant components in North Carolina range from around A.D. 200 to 800 (Phelps 1983:32). In Virginia, Mockley ceramics have been dated from A.D. 460 to 875 at the Maycocks site in Prince George County on the James River (Barka and McCary 1977) and from A.D. 260 to 460 at the Great Neck site in Virginia Beach (Egloff et al. 1988:26). Apart from the relative proportions of sand versus shell, vessels identified as Mount Pleasant and Mockley ware at the Addington and Sherwood's Forest sites are virtually identical in terms of manufacturing methods, vessel thickness and form, surface treatment, frequency of interior scoring, and decoration (Smith and Andrews 1986:30-48, 80-98). As they are currently defined, the most striking difference between the two wares is the addition in the Mount Pleasant series of fabric impressed surface treatment (Phelps 1983:32). The Mockley series includes only cord marked and net impressed types. It may be significant that only one out of sixty-eight Mount Pleasant vessels identified at the Addington and Sherwood Forest sites was fabric marked (Smith and Andrews 1986: 30-48, 80-98).

The possibility that mixed sand and shell in the paste of Mockley ceramics represents an early variation in the ware cannot be eliminated on the basis of evidence from the Hampton University assemblages. Two of the radiocarbon dates for the occurrence of Mockley ware at Hampton University are rather early considering other dates which have obtained for the ware. Charcoal recovered from Feature 42 at HT36 yielded a date of A.D. 20 ± 70 (1,930 B.P., Beta 25913). The feature contained thirty-six shell tempered sherds with a sandy paste: thirty cord marked, three knotted net impressed, one plain surfaced, and two unidentified. One net impressed sherd with a sandy paste was also included. At site HT37, a date of A.D. 40 ± 80 (1,910 B.P., Beta-25914) was obtained on charcoal from Feature 1003, which contained all sandy paste, shell tempered sherds: nine knotted net impressed and four plain surfaced. The plain surfaced sherds probably represent a single vessel base. A third radiocarbon sample on Mockley ware yielded a later date, however. The date was obtained on Feature 1024 at HT37, which yielded ceramics similar in paste to those from the other two dated features. Charcoal from Feature 1024 was dated to A.D. 300 ± 70 (1,650 B.P., Beta 25915). The feature contained five shell tempered sherds with a sandy paste, four knotted net impressed and one cord marked, and two knotted net impressed sherds with a fine sandy paste.

The two earlier radiocarbon dates obtained from the prehistoric sites at Hampton University 42 contribute to a growing body of evidence that a shell tempered ceramic tradition appeared in the Coastal Plain of Virginia during the middle portion of the Woodland I Period (Egloff et al. 1988:30). Painter (1977) has identified an early shell tempered ceramic in the outer Coastal Plain of southeastern Virginia and northeastern North Carolina which, because of its similarity to Marcey Creek ware, is believed to date ca. 1200 B.C. At the Currituck site in North Carolina, Painter has obtained radiocarbon dates ranging from 810660 B.C. on sherd, sand, or shell tempered, flat-bottomed "beaker" vessels with either cord, net, or fabric impressed surfaces (Painter 1978). The early dates derived from the Hampton University sites also lend support to Custer's (1989:276277) suggestion that the Mockley ceramic tradition spread from south to north along the Atlantic coast, with its origin in the outer Coastal Plain of southeastern Virginia.

Very little can be said about the Late Woodland Period ceramics recovered from site HT55. No rim or basal sherds or decorated fragments were included among the seven fabric impressed and seventeen simple stamped sherds in the collection. These shell tempered ceramics were generally characterized by a silty paste. Simple stamping was accomplished with both flat, thong-like and more fibrous materials.

The small number of Townsend and Roanoke Simple Stamped ceramics at HT55 does suggest rather limited use of the site during the Late Woodland Period. In fact, very few, if any, features associated with this period were identified at the site. Townsend ware has been radiocarbon dated in Virginia from A.D. 945 to A.D. 1590 (MacCord 1965; Outlaw 1978; Potter 1982:239-242; Waselkov 1982:286-287). Roanoke ware is suggested to have appeared very late within the prehistoric period. The ware has been found in early colonial contexts at Fort Raleigh on Roanoke Island, North Carolina (Harrington 1948) and at what is believed to be the site of a colonial trading post at Kecoughtan in Hampton. A radiocarbon date of A.D. 1515 was obtained by Painter from a pit at Great Neck in Virginia Beach containing mainly Roanoke ware (Fleming 1981).

In addition to vessel fragments, the ceramic collection from HT37 contained several fired, unshaped clumps of clay which may be waste from ceramic manufacturing activities. The clay is generally of a sandy texture. None of the clumps was observed to contain any shell inclusions. The debris was recovered from Features 1007, 1030, and 1043. The latter feature also yielded the sherd, described above, on which the coil joins had not been obliterated.

The only other prehistoric ceramic artifact recovered in the excavations at Hampton University is a shell tempered, Townsend ware rim sherd which was pressed into the face of a historic ceramic floor tile before the tile was fired. The exterior face of the sherd is visible and is decorated just below the lip with two bands of vertically-oriented notches which seem to have been formed by pressing a short length of dowel into the clay.

Lithics

The sites at Hampton University are remarkable for the very small number of lithic artifacts contained in the assemblages. The collection from HT36 contained only six flakes, three each of quartz and quartzite. Two of these bear cortical surfaces. One quartz biface was recovered from Feature 73, considered to be an historic feature. The artifact was a preform to a small side-notched point. Feature 4 yielded a quartzite mano. The remaining lithic artifacts from the site were a quartzite cobble core and a few fragments of fire-cracked rock.

Site HT37 yielded only two flakes, one each of quartz and quartzite. Aside from a few fragments of fire-cracked rock, the only other lithic artifact recovered was a fragment of a small quartzite projectile point with a contracted stem, which may be of the Rossville type (Ritchie 1971; Stephenson and Ferguson 1963). Both the stem and blade are broken. Width at the shoulder is 43 2.43 cm. Radiocarbon dates from the northeastern United States compiled by Gleach (1985) suggest the Rossville point was in use from about 500 B. C. to A.D. 400.

A total of 340 flakes was recovered from HT55. Of these 267 are quartz (78.5%), 63 quartzite (18.5%), and two jasper (0.6%). The remaining flakes included one each of green slate and sandstone. Also recovered were six flakes of a distinctive material referred to as oolitic quartzite (Bottoms 1968), characterized by quartz grains embedded in a chalcedony-like matrix. Two hundred sixty of the flakes in the assemblage (254 quartz, six quartzite) were recovered from Feature 407, a large shallow basin which also yielded the remnant of a quartzite core and the tip of a quartz projectile point. Three Mockley cord marked sherds were also recovered from the feature.

Five additional bifaces were recovered at HT55. These, which were in an early stage of reduction, include three of quartz and one of quartzite. Feature 33 yielded a fragment of a small, unidentifiable quartzite projectile point, the base of which had been broken. Two modified flake tools were identified in the assemblage, one each of quartz and quartzite.

The excavated assemblage also included four ground stone tools. One is a possible metate fragment of quartzite sandstone which has been fire-cracked. Two other artifacts, which were recovered from the same historic context, appear to have been used to abrade other tools. One was a rectangular block of sandstone, approximately 3 cm long and 2 cm wide, with two parallel linear grooves incised on a face and an edge. The other artifact was a cobble of fine-grained sandstone about 9 cm long, 4 cm wide, and 1 cm thick. One face of the cobble had been abraded to a very smooth surface, and one edge had been squared off from abrading. The remaining ground stone artifact was a schist cobble, roughly 7 cm in diameter with a maximum thickness of 2.3 cm, recovered from Feature 203. The cobble has shallow pecked depressions on each face, suggesting its use as an anvil stone. The feature also yielded 19 Mockley sherds, 17 cord marked and 2 plain surfaced.

Three notable artifacts were recovered at HT55 while cleaning the surface of the subsoil after removal of the plowzone. One was a tear-drop-shaped, bifacial endscraper of quartz, 4.85 cm long with a maximum width and thickness, respectively, of 2.37 cm and 0.75 cm. A small, stemmed, quartzite projectile point of the Savannah River type (Coe 1964) was also recovered. Small variants of the Savannah River type have been radiocarbon dated to about 2000 B.C. at the Plum Nelley site (44NB128) in Northumberland County, Virginia (Potter 1982:276-329). The dimensions of the point from HT55 were: total length 4.13 cm; shoulder width 2.91 cm; stem length 1.07 cm; stem width at base 1.50 cm; and maximum thickness 0.74 cm. The remaining artifact was a triangular-shaped biface of quartzite, probably a projectile point preform. The tip has been chipped slightly, but the existing length was 4.35 cm. Basal width and maximum thickness were, respectively, 2.50 cm and 1.11 cm.

Bone and Shell Artifacts

Fragments of two unidentifiable polished bone tools were recovered from Feature 1026 at HT37, which also yielded one Mockley knotted net impressed sherd. At HT55, one tubular shell bead was recovered.

Structure A

The remains of Structure A were composed of a brick-lined cellar and numerous post holes (see Figures 11 and 12). The post holes represent the placement of three transverses, an east, central, and west, around which the two-bay building was constructed. The post holes will be described in pairs, beginning with the eastern transverse.

Photo 8. Aerial view of 44HT55.

Figure 11. Detail of Structure A.

Feature 5 was the northeast post hole of Structure A, measuring 2.65' in length, 2.41' in width, and 1.25' in depth. It was roughly rectangular, containing brown sandy loam fill and post mold 76. The fill also contained a small amount of brick, a speck of mortar, and a little charcoal flecking. The post mold measured 0.85' by 0.60' by 1.17' and was rectangular as well. Its fill was a dark brown sandy loam containing some bricketage, mortar, and charcoal.

Feature 8 once contained a post which was a repair/replacement for the one previously standing in Feature 5, so indicated by its intruding the latter feature. The roughly rectangular post hole measured 2.4' by 1.8' x 0.72' and contained a sandy clay fill with some brick, mortar and charcoal. The post mold, Feature 9, measured 0.84' by 0.60' by 0.72', was also rectangular, and contained dark brown sandy loam fill.

The southeastern post hole group is, of course, the corresponding set for the east transverse. Post hole 363, with post mold 364, are the original southeast structural posts for Structure A. The post hole was large, measuring 3.0' by 2.5' by 1.56', rectangular, filled with mottled brown sandy loam, and containing some shell and charcoal. The post mold, Feature 364, was 0.7' square and 1.6' deep. It contained only dark brown sandy loam with no apparent inclusions.

Post hole combination 132/133 was the repair post for 363/364, but was not intrusive. Feature 132 was a square hole with straight sides to a depth of one foot, then sloping towards the center, below the post mold. It measured 1.7' by 1.7' by 1.25', containing mottled fill with shell, brick, mortar, and charcoal inclusions. The post mold (133) was also rectangular (.85' by 0.55') with a flat base resting on a clay cap of redeposited subsoil. The fill was a clay and sandy silt combination with brick and charcoal inclusions.

The original post hole for the north central traverse was Feature 347. No mold was visible, 47 Figure 12. Posthole sections—Structure A. 48 its having been obliterated by a later repair. The post hole was large, measuring approximately 3 feet by 2 feet, but extending only 0.7' below subsoil.

Replacement post hole/ mold 344/345 intruded 347 and was probably a later repair. Post hole 347 was roughly 2.0' square and 0.75' in depth. It was intruded by post hole 341, a yet later repair. Post hole 344 contained brown sandy loam fill with heavy orange clay mottling and some shell bits. Post mold 345 was about 0.8' by 0.6', rectangular with a depth of 0.7'. It was filled with dark brown sandy loam.

Post hole/post mold 341/342 appears to be a subsequent repair to the north central traverse of Structure A. The post hole consisted of brown sandy loam mottled with clay and loam. It measured 2.1' by 1.8' and was about 1.0' in depth, containing some shell and brick inclusions. Post mold 342 was roughly 0.7' square and also about 1.0' deep. This post mold contained a Harrington farthing minted in 1613 or 1614.

The southern corresponding post for the central traverse on the south side of Structure A was represented by Feature 316, which was about 1.8' square and 1.1' deep. Its fill contained a brown sandy loam mottled with clay and some brick bits and charcoal flecks. Post mold 317 was large, about 1.0' square, though partially obliterated by later repair post hole 30/330. Post hole 30 consisted of grey-brown sandy loam and measured 1.7' by 1.7' by 1.3'.

The north original post hole/post mold for the western traverse was composed of Features 131/ 129. The post hole (131) was rectangular, measuring 1.7' by ca. 2.0', the western edge having been destroyed when repair post hole/post mold 126/128 was placed. Post mold 129 consisted of dark brown loam and measured 0.5' square. Repair post hole 126 measured 1.9' by 1.5' and contained a 0.5' square post mold (128).

The southern post hole of the same traverse consisted of original post hole 77 with post mold 79, and repair post hole 74 with its post mold 76. Post hole 77 measured 2.7 by 2.4 by 0.9', was roughly square, and had vertical sides and a flat bottom. The fill was mottled light and dark grey sandy loam containing a significant number of seventeenth-century artifacts, shell, bricketage, mortar fragments, and charcoal. It was the only original post hole in Structure A which contained artifacts in large numbers. Its post mold, Feature 79, consisted of dark grey sandy silt, was roughly 0.65' square and 0.75' deep. It, too, contained numerous artifacts and other material. The repair post hole, 74, had sloping sides and a flat bottom under the post mold (76). The hole measured 2.2' by 1.6' by 0.7' and contained grey sandy silt/loam mottled with orange clay, some artifacts (primarily iron), shell, bricketage, mortar, and charcoal inclusions. The post mold (76) was round, with a diameter of 0.8' and 0.9' in depth, consisted of dark grey sandy silt and contained some artifacts as well as shell, brick, and charcoal in small amounts.

Post hole/post mold 353/354 was located midway between the post holes making up the eastern traverse. It may have held a gable post next to the cellar, but since no corresponding post hole was found on the western end of the structure it is more likely that it served as a fireplace support. A passage was added when the cellar was constructed. Post hole 353 was roughly square, ca. 2.0' by 2.0' by 0.95', and filled with brown sandy loam. Its post mold (354) was rectangular (0.5' by 0.6') and about 0.85' deep, filled with dark brown sandy loam. No inclusions were noted.

Structure B

Structure B was also a traverse-raised, two-bay dwelling located only a few feet to the west of Structure A (Figures 13 and 14). It exhibited, as did Structure A, signs indicating several repair episodes. A description of the post holes and mold associated with Structure B will begin with the southernmost traverse.

Post hole/post mold complex 320/321 was located at the southeastern corner of Structure B. The post hole was rectangular, 2.5' by 2.0' and 1.8' in depth. The fill consisted primarily of 49 Figure 13. Detail of Structure B. orange clay with tan, grey and brown sandy loam mottling. There were no artifacts or inclusions in the fill. The post mold was about 0.6' square, filled with dark grey sandy loam. No repairs were evident.

The corresponding post hole/post mold on the southwestern corner of the structure was designated Feature 306/307. The post hole measured 3.2' in length and at least 2.7' in width, with a depth of 2.1' It was, however, intruded on the east side by a repair post and a twentieth-century cow burial. The fill was primarily orange clay (redeposited subsoil) mottled with brown sandy loam. Some bricketage was found in the fill, but little other cultural material. Its post mold (307) was roughly 0.7' square, 1.8' deep, and contained dark brown sandy loam with some brick bits and charcoal. Repair post hole/ post mold complex 303/304 intruded the northeast corner of 306 and was in turn intruded by the modern cow burial. Feature 303 was 1.7' in length and probably about 1.5' wide, with a depth of about 2.0'. The post mold, 304, was 1.8' deep, roughly 0.8' square and filled with dark brown sandy loam with some brick and charcoal bits.

East-central post hole 276 measured 2.6' by 3.0' and was 1.9' deep. It was intruded on the east side by a slot fence trench connecting Structures A and B. The fill was a mixture of loam and clay with shell, brick, and charcoal inclusions. The post mold (277) was about 0.8' square and 1.8' deep. Its dark sandy loam fill contained only nails.

The west-central framing post hole/mold, 291/ 292, was intruded by Features 289/288 on the north side. Post hole 292 was rectangular, measuring 3.3' by 3.0', had steeply sloping sides and a flat bottom, two feet in depth. The fill consisted of alternating bands of clayey redeposited subsoil and sandy loam, containing aboriginal pottery, tin-enamelled earthenware, case bottle glass, lead shot, and 7 nail fragments. A nearly whole brick was embedded in the upper part of the fill. 50 Figure 14. Posthole sections—Structure B. The post mold, 291, was circular, 0.8' in diameter, containing brown sandy loam fill and bricketage, along with a nail and a domestic pipe bowl fragment. The post mold extended to the bottom of the post hole.

Repair post hole 289 was smaller than 291, roughly 1.6' square and 1.4' deep. The south edge was steeply sloping, whereas the north sloped more gently to a flat bottom. The fill was hard-packed clayey redeposited subsoil with tan sandy loam mottling. Two nails, one piece of lead shot and a fragment of daub were recovered from the section. Post mold 288 consisted of dark brown loamy fill, was circular, about 0.7' in diameter, 1.3' deep, with a flat bottom extending about 0.15' 51 below the bottom of the post hole. The mold contained two nails.

The north-east post hole, 163, measured 3.0' by 2.6' was rectangular and intruded by repair post hole 162 on its south end. The fill was 1.6' deep, mixed orange and yellow clay mottled with light brown sandy loam, containing aboriginal pottery, a nail, and two brick fragments, shell, and charcoal. The post mold, 165, was roughly 0.8' square, also 1.6' deep, consisted of dark brown sandy loam containing no artifacts, but some brick and charcoal flecks.

The post hole/mold complex intruding the southern part of the original post hole were Features 160 and 162. Feature 160 was roughly 2.0' by 1.6' by 0.9', rectangular, consisting of mixed clay fill, and containing aboriginal pottery and two roofing tile fragments. The post mold (162) was also roughly rectangular, 1.0' by 0.8', consisted of dark grey sandy loam fill and contained a fragment each of aboriginal pottery, tin-enamelled earthenware, two nails, and a domestic pipe stem fragment.

The northwest post hole complex (472/473) was intruded by repair post hole 470. Feature 472 was roughly 2.8' square, 1.6' deep, consisting of brown sandy clay and containing one nail, and some brick and charcoal fragments. The post mold (473) was about 0.8' square, 1.7' deep, flat-bottomed and contained fill consisting of dark brown sandy loam containing two nails and a faunal bone fragment. The intrusive repair post hole complex 470/471 obscured the southeast corner of 472/473.

Post hole 470 was irregularly-shaped, measuring 2.4' by 1.7' and 1.7' in depth. The fill consisted of brown sandy loam mottled with orange clay containing one fragment of container glass and some brick and charcoal fragments. The post mold (471) was roughly 0.4' square, also 1.7' deep and consisted of brown sandy loam fill containing one nail fragment. Both the hole and the mold had relatively flat bottoms.

Post hole/mold 115/116 was located between the two southern framing posts and probably served as a fireplace support as did Features 353/ 354 in Structure A. Post hole 115 measured 2.1' by 1.7'and was 2.02' deep. Both sides were steeply sloping and the bottom flat. The post mold (116) was rectangular, 0.6' by 0.7', resting on the bottom of the hole cut.

Structure C

Structure C (Figures 15 and 16) was probably the first building erected at HT55. It was rectangular, approximately 20' by 40', oriented east-west. The building consisted of eight post hole/ post mold complexes forming three long bays.

The northernmost post hole of the eastern traverse was designated Features 271/270. The post hole was rectangular, measuring 2.75' by 2.1' and 1.05' in depth.

It had steeply sloping sides and a flat bottom. Fill was re-deposited subsoil mottled with topsoil mixed when the post hole was dug. No artifacts or inclusions were evident. The post mold (270) was nearly square, 0.8' by 0.7', and also 1.05' in depth. Its fill was dark brown sandy loam and was sterile except for one brick fragment.

Post hole/post mold 283/284, located at the southeastern corner of Structure C, were severely truncated by the later east-central framing post for Structure B. The north-south measurement for 283 was 2.2', but the east-west length was unattainable. The post mold, 284 was 0.7' north-south and was probably square.

Post hole/mold 138/139 was roughly rectangular, 2.45' by 2.18', and was 1.0' deep. The fill was re-deposited orange clay subsoil with brown sandy loam lenses. Artifacts recovered were one fragment of aboriginal pottery, a flake, and a fragment of English flint. The post mold (139) was 0.8' in diameter, circular, and consisted of tan sandy clay fill mottled with brown sandy loam. One nail, one straight pin, and a glass linen smoother handle were recovered from the 1.0 feet of fill.

Feature 295 was probably roughly circular, measuring 2.4' in diameter, but had been subsequently intruded on the west side by the west-central framing posts from Structure B. The post 52 Figure 15. Detail of Structure C. mold (294) was also roughly circular, measuring about 0.9' in diameter and filled with brown sandy loam. Both the hole and mold were about 2.3' deep.

Post hole 172 was roughly rectangular, measuring 2.7' by 2.3' with a depth of 1.6'. The sides of the post hole were basically straight, sloping toward the base to a flat bottom. A lens of light grey sand was found beneath the post hole. The fill was otherwise typical and contained no artifacts or inclusions. The post mold, 173, was roughly 0.6' square with very dark grey sandy loam fill. There was a clay deposit below the post hole, raising it 0.3' above the bottom of the post hole. No artifacts were recovered, but inclusions of brick bits were noted.

Feature 368, the eastern post hole, was roughly 2.3' square and contained brown sandy loam mottled with orange clay fill. Shell and charcoal inclusions were recorded, but no artifacts were recovered. The post mold, 369, was 0.6' square and contained dark brown sandy loam fill. No artifacts were found.

The northwest corner post hole, Feature 179, was somewhat smaller than the others, measuring about 2.0' square. The bottom of the post hole, about 1.4' below the surface was flat and rounded in plan. The fill was re-deposited subsoil 53 Figure 16. Posthole sections—Structure C. mottled with brown sandy loam, containing no artifacts or inclusions. The post mold, 180, was rectangular, 0.65' by 0.50' and extended to a depth of 1.06'. Its dark brown sandy loam fill contained one nail and a roofing tile.

The southeastern post hole for Structure C was Feature 315. Unfortunately, it was almost entirely obliterated by the later southeast corner post of Structure D.

Structure D

This southeastern corner post hole, Feature 279, was part of Structure D (Figures 17 and 18), a building which obviously postdated the abandonment of Structure C. The post hole was nearly square, 2.2' by 2.1', and intruded by Structure E post hole 265 on its southern edge. It was 1.85' deep consisting of brown sandy loam mottled with orange clay and containing three nails, a domestic pipe stem, faunal bone and some bricketage inclusions. The post mold (280) was 0.7' square, contained dark brown sandy loam with orange clay inclusions, and was 1.85' in depth. One wrought nail and a sliver of wood were recovered from the fill.

Post hole/mold 169/170 once represented the northeast corner post for the structure. It was irregularly shaped, measuring 2.5' by 2.0' with a depth of 2.0'. Its fill was the typical orange sandy clay re-deposited subsoil with brown sandy loam mottling, containing only one aboriginal pottery sherd and a few brick and charcoal bits. The post mold, 170, was ovate, 1.0' by 0.6', and also attained a depth of 2.0'. Its fill consisted of dark brown sandy loam and contained, in addition to brick, shell, and charcoal inclusions, nails, case bottle glass, English flint, a straight pin, and unidentified iron and lead objects. The presence of this material in the post mold fill may indicate that the post was pulled up, allowing surrounding debris to enter the resulting hole.

The northern post hole/mold in the middle traverse, Feature 149/150, was a large, nearly square post hole, measuring 3.0' by 3.08', attaining a depth of 1.8'. The post hole's sides were straight and its base flat. The fill was re-deposited subsoil mottled with brown sandy loam containing, in addition to brick and charcoal inclusions, two sherds of aboriginal pottery, a fragment of container glass, two nails, a copper alloy thimble, a bit of bog iron, and a faunal bone. The post mold, 150, was rectangular, 1.2' by 0.9', 2.1' in depth, consisting of light brown and grey loam mottled with tan sandy clay, and containing three aboriginal pottery sherds, two nails, a domestic pipe bowl and stem, and six fragments of unidentified iron.

The corresponding post hole/mold in the central traverse, Feature 337/336, was approximately the same size and shape as 149/150, measuring 3.0' by 2.9' and 2.0' deep. The fill was also redeposited subsoil mottled with brown sandy loam and contained a great number of historic artifacts: 54 Figure 17. Detail of Structure D. Figure 18. Posthole sections—Structure D. 55 three aboriginal pottery sherds, two blue coarseware pieces, one faience sherd, one Westerwald fragment, twelve nails, one straight pin, two fragments of daub, faunal bone, brick, shell, charcoal, and an assortment of domestic and imported tobacco pipe stems and bowls. The post mold, 336, was surprisingly small, measuring only 0.6' by 0.4', filled with dark grey sandy loam and containing a case bottle fragment, a nail, and two faunal bones. The large amount of cultural debris found in the post hole was somewhat unexpected. There seem to be two explanations for the material, the structure was originally constructed a long enough time after Structures C that trash had built up around the site and was subsequently deposited in the hole, or the post hole was disturbed by later post hole 299.

The northern post hole (Feature 223) in the western traverse was a large, 2.8' square feature with a post mold, 224, in its southwest corner. The 1.45' deep post hole contained re-deposited subsoil mottled with brown sandy loam fill and numerous seventeenth century artifacts, similar to 149. The post mold was irregularly shaped, measuring 1.1' by 1.0' with a depth of 0.8', substantially less than that of the post hole. The fill was grey sandy loam mottled with orange clay containing four nails, an iron buckle, daub, and bone. Shell, brick and charcoal inclusions were also noted.

The southwest corner post hole in Structure D, Feature 312, was a large, 2.5' roughly square post hole almost entirely obscuring the southwest corner post hole for Structure C. The post hole was 1.7' deep consisting of orange, tan, grey, and dark brown loam mottled with light tan silt, and containing such historic artifacts as coarseware fragments, Westerwald, nails, window lead, flint, pins, bog iron, pipe bowls, pipe stems, iron, bone and oyster shell. The post mold, 313, was relatively small, measuring 0.7' by 0.6', and almost square. Its dark brown loamy fill contained nails, copper alloy tacks, bone, shell, chalk, pipe stems and bowls, and a possible iron knife blade. It, too, was approximately 1.7' deep.

Feature 349 was a post hole located on the middle east end of the structure, probably serving as fireplace support post. It was irregularly shaped, approximately 2.2' by 1.9', consisting of 0.8' of brown sandy loam fill containing two nails, one case bottle fragment and two faunal bones. The post mold, 350, was circular, 0.6' in diameter and consisted of dark brown sandy loam fill containing two nails, one faunal bone, and oyster shell.

Structure E

The final building, Structure E, was the last to be built at HT55, replacing Structure D when it apparently fell into disrepair (Figures 19 and 20). It shares no walls or post holes with the previous buildings, was ca. 20' wide, 24' long, and made up of six framing post holes. This building appears to have been side-raised, rather than traverse-raised like the majority of the others. Two small off-set post holes on the western end of the building may indicate a fireplace.

Feature 219 was the northeastern corner post hole for the structure. It was almost square, measuring 2.5' by 2.2' and containing 1.5' of orange clay mottled with brown sandy loam fill. Two aboriginal pottery sherds, two nails, a flake, and a large amount of bricketage was recovered from the fill. The post mold, 221, was 0.8' circular, containing brown sandy loam, but no artifacts.

The southeast corner post hole, Feature 265, was rectangular, measuring 2.2' by 2.4'. The bottom of the hole was square and flat some 1.5' below the surface. The brown loam mottled with orange clay contained shell, brick, and charcoal inclusions along with a considerable number of seventeenth-century artifacts (see Artifact Catalog, Volume II). The post mold, 266, was circular, 0.7' in diameter, consisted of dark brown silt, and contained a case bottle fragment, nails, pins, pipe stems, a pipe bowl fragment, and bone. It, too, was approximately 1.5' deep, with a flat bottom.

North-central framing post hole 157 measured 2.1' by 2.6', was rectangular and 1.65' in depth. 56 Figure 19. Detail of Structure E. Figure 20. Posthole sections—Structure E. 57 The sides were steep and straight, the bottom flat. The fill consisted of light yellowish brown clay mottled with brown, grey and black sandy loam. Artifacts recovered were numerous, including aboriginal pottery, 21 nails, daub, brick fragments, imported and domestic pipe stems and bowls, and an unidentified iron object. The mold, 158, was 0.9' circular, with slightly irregular sides and a flat base resting on about 0.2' of packed clay. Artifacts recovered included widow glass, 6 nails, brick, and bone.

South-central post hole 299 was 3.0' by 2.4' rectangular, consisting of light brown sandy loam mottled with clumps of yellow clay. The sides were straight and the bottom flat. Artifacts recovered included 15 nails, one window came, an English flint, one clothing pin, mortar, brick fragments, shell, and domestic and imported smoking pipe fragments. The post mold, 298, was 1.0' by 0.9' by 1.6', consisting of loose, soft, dark brown sandy loam. A depression in the bottom of the post hole suggests that the post was driven past the bottom of the hole or sank under the weight of the building. Artifacts recovered included aboriginal pottery, clothing pins, 2 domestic pipe bowl fragments, and 2 domestic pipe stem fragments.

The northeast corner post hole, Feature 176, was nearly square, measuring 2.4' by 2.3' with a depth of 1.6'. Its sides were straight and its bottom flat; the fill consisted of orange sandy clay mottled with tan and grey sandy loam. Artifacts recovered from the feature included two aboriginal pottery sherds, redware, case bottle glass, nails, flint, an iron clothing hook, imported and domestic pipe stems, with brick, shell, and mortar inclusions. The post mold, 177, was about a foot square, contained dark brown sandy loam fill, but only one nail and another clothing fastener. Brick, shell, and charcoal inclusions were also recorded.

Two small post holes, Features 373 and 374, were located between the two western support posts. Both of these post holes were small and did not contain post molds. They probably functioned as supports for an entrance or possibly a chimney. Post hole 373 was about 0.9' square and only 0.7' deep. The fill was grey/tan sandy loam with orange clay mottling and contained a nail, brick fragments, and oyster shell. The bottom appeared to be slightly rounded. Post hole 374 was slightly rectangular, 0.7' by 0.9', and only 0.35' in depth. The fill was identical to that of 373 and its bottom was also slightly rounded. Artifacts recovered included six fragments of tin-enamelled earthenware, a nail, and an imported pipe stem.

Non-Structural Features

In addition to the five buildings erected at 44HT55 during the first half of the seventeenth century, several other non-structural features that require detailed descriptions. These include the two large trash pit just north of Structure A (Trash Pits A and B), the well, Trash Pits C, D, and E, the slot fence trenches, and the boundary/drainage ditch in the north central area of the site. Features or layers making up a larger entity were given so-called "macro-feature" designations and were considered related in the artifact analysis.

Trash Pit D

Trash Pit D was roughly circular, about seven feet in diameter and 1.75' in depth (Figure 25). The top layer, Layer 49, consisted of dark brown sandy loam with a very heavy concentration of oyster shell, very similar to Trash Pit C. About 1.2' in depth, Layer 49 contained about 600 seventeenth-century artifacts. Layer 64 was the second deposit in Trash Pit D, consisting of a 0.15'thick lens of light tan sandy loam, mottled with grey sandy loam containing inclusions of shell, brick, mortar, and charcoal, in addition to a good assortment of seventeenth-century rubbish. Below 64, another lens (Layer 67), about 0.2' deep consisting of grey ash was located. Inclusions of shell, brick, mortar, and charcoal were recovered, in addition to artifacts such as coarseware, aboriginal pottery, case bottle glass, nails, pins, aiglets, coal, and domestic and imported smoking pipe fragments.

The bottom layer in Trash Pit D was Layer 68, a re-deposited subsoil layer about 0.2' deep containing only charcoal inclusions and aboriginal pottery and flakes. This layer may represent soil washed-in between the time the pit was dug and its initial use as a trash receptacle.

Figure 25. Section—Trash Pit D.

64

Trash Pit E

Pit E was an irregularly-shaped pit located about 15' west of Trash Pit D in the northwestern section of the site. On the surface the feature was about 8.5' long, 6.0' wide on the eastern side and 4.0' wide on the western. About 0.7' below the surface, however, the pit became rectangular, approximately 7.0' long and 2.4' wide, very much like a grave feature. The top fill (Layer 87) appeared to have been disturbed somewhat by the growth of a later tree, but consisted of dark brown sandy loam with inclusions of shell and brick (Figure 26). Although it appeared to be a trash pit, relatively few artifacts in comparison with the other trash pits, were recovered from the fill. The bottom surface of 87, and therefore the top surface of Layer 97, was undulating, about 0.7' deep at the deepest point. Layer 97 completely filled the grave-like part of the feature also resting on the bottom of the upper part of the pit. The fill was a fine tan sandy soil containing no inclusions and only five fragments of aboriginal pottery and two sherds of English coarseware. Although the feature appeared to be a human grave shaft, virtually identical to those excavated at contemporary sites, no trace of human remains could be found. Perhaps the intended occupant recovered from his illness or the soil in 350 years completely absorbed the bones. Since most people in that time period were buried without the benefit of clothing, no aiglets, buckles or buttons would have been interred with the body and thus none subsequently survived.

Figure 26. Section—Trash Pit E.

Prehistoric Features

Four features of probable prehistoric origin were identified at HT55. All contained Mockley ware ceramics as found at sites HT36 and HT37, suggesting that they date from the Late Woodland I Period.

Three of the features are relatively small, shallow pits, all bowl-shaped in profile. Feature 155 (S260 E280) was 2.0' in diameter and 0.85' deep. The orange, brown sandy loam fill contained 4 sherds of shell-tempered, cord-marked ceramic and 2 flakes, one each of quartzite and green slate. A slightly larger pit, Feature 203 (S250 E340) was filled with grey-brown sandy loam with flecks of shell. Extending 0.9' in depth and 2.3' in diameter, the feature yielded 19 shell-tempered sherds (17 cord-marked, 2 plain) and a small schist cobble possibly used as an anvil stone. Feature 70 Figure 30. Section drawing of well. 71 194 (S160 E370) was 2.2' in diameter and 0.4' deep. Eight shell-tempered cord-marked sherds and 6 quartz flakes were recovered from the fill.

Feature 407 (S150 E380) was a large oval pit, very shallow in depth. The feature measured 4.5' by 3.5' in plan and extended only 0.35' below subsoil level. The bottom was irregular. Fill consisted of light tan-grey sandy loam containing 3 shell-tempered, cord-marked sherds. In light of the very small number of lithic artifacts recovered during the Hampton University excavations, Feature 407 was remarkable in that it also contained 254 quartz flakes (most smaller than 20 sq cm in plan), 6 quartzite flakes, the remnant of a quartzite core, and a fragment of the blade tip to a quartz projectile point.

Chemical Analysis

Although chemical analysis was carried out at HT55, few of the sampled features contained significant amounts of potassium or phosphorus (see Figure 31). Those containing high levels of calcium (Features 77 and 79) both contained oyster shell, probably leading to the high readings.

Introduction

Animal bone is among the most important material on an archaeological site, providing not only evidence of diet but also some indication of food procurement strategies, marketing and redistribution, seasonality and activity scheduling, environmental conditions, and even trade relations and the effective use of the surrounding resources. The study of archaeological faunal remains—called "zooarchaeology"—investigates these and other topics, using data and techniques from archaeology, anthropology, history, zoology, and animal ethology.

Much of the faunal data gathered in the Chesapeake is summarized in a major work titled "Colonization and Subsistence Change on the 17th Century Chesapeake Frontier," by Henry Miller of the St. Mary's City Commission. Miller's work provides by far the best synthesis of our present state of knowledge about the foodways of early colonists, and it will be referred to often in the discussion below. Other important sources include studies of the food remains from Clifts Plantation (Bowen 1979) and the Maine (Barber 1978), as well as contemporary natural histories (Beverley 1947; Hemor 1957), traveller's accounts (Durand 1934), and studies of seventeenth-century food and foodways (Spencer 1982).

Despite this work, however, there are still significant questions in seventeenth-century Chesapeake zooarchaeology. Many were raised during a 1987 planning conference on Chesapeake foodways (Gaynor et al. 1985), which concluded that a comprehensive and multi-disciplinary research program is needed in order to understand both seventeenth- and eighteenth-century developments in the field. On the archaeological end, this involves continued research on a variety of 79 sites such as those at Hampton University, which will provide a fuller and much richer database to draw upon.

The Hampton University historic site (44HT55) contained some 13,884 individual elements, 3584 identifiable and 10,300 unidentifiable. This is a much larger assemblage than most of those cited in Miller's 1984 work, and indeed is one of the largest bone assemblages from any historic site in the area. The percentage of identifiable elements—26%—is roughly similar to that seen on other local sites, reflecting relatively good preservation conditions and relatively little post- depositional disturbance.

Methods

All animal bone from the Hampton University site was bagged by context and washed. The bone was then sorted into "identifiable" and "unidentifiable" components—identifiable bone being defined as bone which could be classified at least to the taxonomic level of Order. Unidentifiable bone was subdivided into several major taxonomic groupings: fish, reptile or amphibian, bird, bird or small mammal, mammal, small mammal, medium mammal, large mammal, or undetermined. The bones were also subdivided on the basis of major element categories—long bones, flat bones, ribs, vertebrae, crania, mandibles, teeth, spines, and scales. Each particular taxon/element grouping from each particular context was then counted and weighed, receiving a so-called "unique bone number" for use in the computerized database. Once entered, the unidentified bone was put aside.

Each identifiable bone was labelled with its context number and its own unique bone number, in order to facilitate future tracking. These bones were identified by reference to comparative skeletal collections, in most cases using Joanne Bowen Gaynor's collection in the Colonial Williamsburg Department of Archaeological Research. Some of the birds, and all of the reptiles and amphibians, were identified using the collections of the Departments of Birds and Herpetology of the Smithsonian Institution's Museum of Natural History. In all cases, bones were identified on the basis of species-specific morphological characteristics; when such characteristics were virtually identical in several species, the bone was identified to a higher taxonomic level such as genus or family.

Due to the lack of specific morphological indicators in some elements (particularly ribs and, to a lesser extent, vertebrae), many of these elements were grouped together at the level of Family or Order. It should be noted, however, that much of the "Family Phasianidae" material is probably chicken, and much of the "Order Artiodactyla" material is probably pig. When an element was judged to be very likely (but not undoubtedly) associated with a particular taxonomic category, it was labelled "cf." (literally, "compare"). These elements were generally lumped along with the definitely-identifiable elements in most of the analysis below.

Two categories—Artiodactyla I and Artiodactyla II—were used to group certain elements. Loosely based on terminology used by Crader (1984), Artiodactyla I is defined as members of the Order Artiodactyla (even-toed ungulates) which include pig, deer, sheep, and goat. Because pig bones can often be separated from those of the latter three taxa, Artiodactyla II is defined as members of the Order Artiodactyla which include only deer, sheep, and goat. These latter three share many morphological characteristics, but once again deer bone can often be distinguished from that of sheep or goat. The latter two, however, are extremely similar skeletally, so much so that they are most often lumped together as sheep/goats (sometimes called "ovicaprids") by modern analysts.

The identifiable bones were entered into a database created in Ashton-Tate's dBASE III Plus. Recorded information included unique bone number, context number, taxon, element, side, portion that was present, epiphyseal fusion state, relative size, tooth type and wear, weight, and presence or absence of modifications such as burning, weathering, carnivore or rodent chewing, and butchering.

Collection Methods

All soil from the site was screened through one-quarter inch mesh. Portions of selected contexts such as the well were also wet-screened through one-eighth inch mesh. Flotation samples were taken, but little if any identifiable bone was recovered.

It has been argued that screening is extremely important in order to provide a representative sample of all faunal remains (Payne 1972). Many of the sites described by Miller (1984), however, were excavated using the "hand-picking" method; it appears, based on a gross comparison of the percentages of small fragile bones in the various sites, that similar recovery was obtained, and thus these sites can profitably be compared with the Hampton University assemblage (and the other "screened" sites in Miller's sample).

Methods of Analysis and Aggregation

Artifactual and architectural data suggests that the site was occupied for a reasonably short period, circa 1630 to 1660. Aside from a few modern contexts, therefore, all faunal material was aggregated into a single assemblage. The data will be treated as such in the work below, except for certain specific instances where the analysis of seasonal indicators or preservational bias requires that particular features be analyzed independently.

Four different methods of quantification have been used: number of identified specimens (NISP), minimum number of individuals (MNI), usable meat weight, and biomass based on skeletal allometry. Each has its staunch adherents, but it is generally recognized that there are faults in each method. Because they have individual strengths and weaknesses, they will be briefly discussed here.

The NISP method involves simply counting and comparing the numbers of identified elements for each taxon. It is certainly the most basic measure of relative importance, and many analysts have suggested recently that it has been criticized unfairly (Grayson 1984). It is clear, however, that there are problems with this approach. It is skewed toward animals with fragile, easily broken bones; it does not account well for situations where a single animal is disproportionately represented by many bones (as is the case with the cattle found in the Hampton University well); and it obviously must be transformed in some way to indicate that, for instance, the bones of a cow represent far more meat that those of a rabbit.

The MNI method is used to mitigate problems of element interdependence. First proposed for archaeology by paleontologist Theodore White in 1953, it is based on the determination of the smallest number of animals that could have accounted for the faunal remains. In practice this is accomplished by pairing lefts and rights, and counting the largest number of unique elements (all paired bones plus unpaired elements). Following current practice, in this study size variation and morphological differences will be used as factors when pairing elements, thus producing a larger and more accurate MNI count.

Although this seems a conservative and safe approach, the MNI method has been accused of having certain important biases. The counts vary significantly with changes in sample size and differences in the way that the units are aggregated. One approach is to use the entire site as a single assemblage (which is done here), and to determine MNIs for the site as a whole. It is also possible, however, to determine MNIs individually for each feature or group of features, under the assumption that the bones deposited represent different animals than those deposited in other features on the site. The cumulative MNI count using this latter technique will be generally higher, but lacking good archaeological evidence for the function and contemporeity of features it is difficult to determine which count is more accurate.

Nonetheless, MNI counts are the basis of most modern zooarchaeology. They are used in the third major method of quantification, which is simply a transformation technique resulting in a comparison of actual usable meat. The "pounds of usable meat" method involves multiplying the MNI count for each taxon by the average amount 81

The final method of quantification is based on skeletal allometry: the principle that any two dimensions of an animal's body (for instance, bone weight and total live weight) will vary in a relatively predictable fashion. Used most innovatively by Elizabeth Wing of the Florida State Museum and Elizabeth Reitz of the University of Georgia, this study of "biomass" involves determining the bone weight for a particular taxon and inserting it into a regression equation based on empirical studies at the University of Georgia (for the allometric constants, see Table 6). The principal problems with this approach involve the accuracy of these modern estimates to the allometric relationship of past animals, as well as the significant changes in actual bone weight that can be caused by burning and "post-depositional" decalcification.

Several other attributes of the assemblage were studied as well. The relative proportions of elements were determined for the major taxa, in order to test preservational indicators as well as to determine possible butchery and disposal methods. A feature containing an abundance of hard, dense bones or teeth and very few softer, less dense bones, for instance, should be viewed with caution, since preservational conditions such as soil acidity may have destroyed a significant portion of the original "death assemblage." Such a feature suggests, in fact, that many taxa such as fish and certain birds may no longer be represented at all, and this must be borne in mind for all other related analyses.

Aside from preservational bias, however, element distributions are also useful for indicating butchery and disposal practices. It is often thought, for example, that the remnants of meals should be distinguished by a relative abundance of long bones and vertebrae, while butchery waste dumps should contain mostly non-meaty elements such as phalanges and sesamoids, along with cranial parts. In the seventeenth- and eighteenth-century Chesapeake, however, no such distinction is visible on most sites.

Animal husbandry is another important subject that can be studied. The age structure of the animal populations buried on the site can be determined by evaluating epiphyseal fusion and/or 82 tooth eruption and wear. Epiphyseal fusion distributions have been used in this study, following methods described by Chaplin (1971), and will be discussed shortly.

Tooth eruption is potentially even more discriminating, since the times of eruption have been carefully studied in livestock management. It has been suggested, however, that eruption stages were different in the past, and it is the values for historic populations given by Silver (1969), Payne (1973), and Bull and Payne (1982) that are now generally used. Tooth wear study is becoming increasing popular as well, although the stages are less readily accepted (except Payne's [1973] study of sheep and goats). Nonetheless, an accurate age distribution is only possible with a fairly large sample; none of the taxa from the Hampton University site contained enough teeth to make such a study profitable.

Finally, osteological measurements can be used to help establish the probable size of the animals themselves. The study of so-called "dimensional allometry," however, is not far advanced, and it is generally difficult to determine even the average size (much less the range of variation) for animals in the historic past. Although not used for specific analysis, therefore, the osteological measurements of the Hampton University material were taken in order to contribute to a database being compiled at the University of Georgia, so that this significant avenue of inquiry can be better followed in the future. All measurements taken of this assemblage followed the techniques described by von den Dreisch (1976). The data is on file in the Department of Archaeological Research of the Colonial Williamsburg Foundation.

Taxa Represented in the Assemblage

The Hampton University assemblage contained approximately 14,000 bone elements representing forty-eight species: one crustacean, ten fish, two amphibians, seven reptiles, fourteen birds, and fourteen mammals (see Table 7). Roughly twenty-six percent by total number, and seventy-nine percent by weight, were identifiable to at least the level of Order. A brief description of the species identified will be presented below.

Domestic Mammals

The major domestic mammals—cow, pig, and sheep or goat—were by far the most important contributors to the meat diet of the occupants, supplying some 75 percent of the meat as measured using the biomass method (see below). Cow, in particular, was extremely important, supplying both the greatest number of elements and the greatest amount of meat in the assemblage.

The domestic cow (Bos taurus) was brought over from the Old World almost immediately after settlement began. Cattle prospered in the new environment, although it has been suggested that they were not exceptionally well-cared-for. They were mostly allowed to roam free, and by 1633 the predations of carnivores (and local Native Americans) upon cattle herds was one reason for the erection of a six-mile-long palisade across the James-York peninsula.

Of course, the typical seventeenth-century cow was much smaller than the modern breeds, apparently providing only some 400 pounds of usable meat per adult animal (Miller 1984). There is considerable interest in determining the physical characteristics of these early cows, and thus it was remarkable that the well fill contained, at a depth of about six feet, a virtually-complete adult cow skull (Photo 14). The skull, which most resembles a modern Devon in morphology, provides a clear and important picture of the general appearance of these animals in the early colonial period.

87

Photo 13. Cut deer antlers.

The domestic pig (Sus scrofa) was the second most important source of meat for the colonists throughout the seventeenth and eighteenth centuries. Pigs were allowed to roam free in the woods, subsisting on what food they could scavenge, and thus required little care. Miller (1984), in fact, has suggested that it was this foraging ability, along with a prolific reproductive rate and fast growth rate, which made them an ideal source of food on the Chesapeake frontier. Probate inventories indicate that, before 1660, almost everyone had one or more pigs; in contrast, in the earlier periods only one-third of inventoried households owned cattle (Miller 1984:297).

Sheep (Ovis aries) or goat (Capra hircus) was not nearly as important as cow or pig. Sheep, being particularly vulnerable to natural predators, were hard to raise successfully in the Tidewater, and they were found in consistently low numbers in the twenty-one sites or phases described by Miller.

Photo 14. Reconstructed cow skull.88

Other domestic mammals recovered in relatively low numbers include horse (Equus cabellus) and domestic cat (Felis domesticus). Three bones—a mandible, humerus, and tooth— may represent the remains of either domestic dog (Canis familiaris) or wolf (Canis lupus).

Dietary Contribution

Relative dietary contributions were assessed in two ways—using pounds of usable meat and biomass values (see Table 8). Both methods suggested that cattle were far and away the most significant food resource, accounting for 57.6 percent of the pounds of usable meat and 62.0 percent of the biomass. Pig was second in importance, representing 26.4 percent of the pounds of usable meat and 13.8 percent of the biomass. Next in significance were wild mammals—principally white-tailed deer—which accounted for 7.7 percent of the pounds of usable meat and 3.6 percent of the biomass. Fish were also a major food source, with 4.4 percent of the pounds of usable meat and 2.1 percent of the biomass.

Despite the presence of a variety of wild game and fish, it is clear that the occupants relied quite heavily on domestic livestock. The major domesticates—cow, pig, sheep, goat, and chicken— accounted for 87.2 percent of the pounds of usable meat and 76.1 percent of the biomass, as opposed to 8.4 percent and 3.6 percent, respectively, for wild mammals and birds.

This reliance on domestic livestock, particularly cattle, is mirrored at the other sites described by Miller, where, except for the Maine Site, cow accounted for anywhere between 42.19 and 72.16 percent of the total amount of usable meat. Likewise, pig was consistently the second most important food source, ranging between 11.22 and 36.64 percent. At the Maine Site (the earliest

These similarities can be seen most easily in a chart produced by Miller (Table 9), which breaks down the mean percentages of usable meat by three temporal periods: 1620 to 1660, 1660 to 1700, and 1700 to 1740. In the earliest period cow accounted for some 44.26 percent of the total usable meat, followed by pig at 24.65 percent, deer at 16.83 percent, and fish at 10.66 percent. As one can see from a comparison with Table 8, cow was slightly more important at the Hampton University site, with deer and fish correspondingly less important.

Mean percentages for the period 1660 to 1700, however, are much more skewed in favor of domestic mammals, and the Hampton University data thus clearly suggests a temporal placement in the latter part of the period 1620 to 1660.

Livestock Management and Husbandry Patterns

Given the extraordinary contribution of domestic mammals in the diet of the colonists, even during this early period, it is important to understand some of the details of animal husbandry. Most of these patterns, of course, were imported with the colonists; Trow-Smith (1957) and Miller (1984) give useful summaries of livestock management in England during the sixteenth and early seventeenth centuries. Many patterns, however, had to change to accommodate differences in climate, predator pressure, reproductive rate, and animal use between the Old World and the New.

Livestock husbandry can be investigated archaeologically using so-called "kill-off patterns" derived from the bones themselves. As an animal matures, the bone grows outward from various centers of ossification (in long bones, the diaphysis or shaft and two or more epiphyses—at least one on either end). When growth stops, the cartilaginous plates separating the diaphysis and epiphysis themselves ossify, and the epiphysis is said to have "fused." As this stage of skeletal maturity varies among different elements for any individual species, it is possible to separate fused and unfused bones systematically and to construct what amounts to an age distribution for that species (see Figure 36).

These age distributions provide important clues about the ways in which the animals were used. For example, the pig distribution from the Hampton University site indicates that some 70% of the animals were apparently killed in their first 90 Figure 36. Kill-off patterns—pig, cow, and sheep/goat. year, with a lesser number killed before 30 months and a few low percentage killed after this time. Pigs obtain their adult weight fairly early, by the age of about twelve months. It appears that most were slaughtered soon after, possibly so that they did not require further care. Given their freedom to roam in the woods, however, another cause may have something to do with the relative ease with which younger animals could be caught, in contrast to older, presumably wiser, animals.

The distribution for cow suggests that the great majority were slaughtered after two years of age, with roughly equal percentages killed between two and three years and between three and four years, and a slightly higher percentage killed after four years of age. This would be expected in a so-called "subsistence-based" agricultural economy, where cattle were raised for a variety of purposes. Although a small percentage were slaughtered as veal calves, most were raised to maturity and only slaughtered when they had obtained their optimum weight at about two years of age. As shown by the high percentage of older cattle, however, many were retained as breeders, milk producers, or draft animals.

The distribution for sheep or goat shows that all animals were apparently killed after the age of forty-two months. This suggests that sheep were 91 being raised for their wool, and only killed and eaten after the quality of wool began to decline at six or seven years of age. Relatively few animals were kept, judging by the total number of elements—probably due to the difficulty of protecting them from disease and natural predators.

Site Catchment and the Use of the Environment

In a 1979 review article, Donna Roper defined a site catchment as "that area for which a site (or more properly, the inhabitants of a site) derived its resources" (1979:120). First used in archaeology by Claudio Vita-Finzi and Eric Higgs (1970), it is a useful conceptual tool for explaining resource use and exploitation. Given the lack of a comprehensive environmental and ecological reconstruction for the Hampton University site, however, it will be treated here on a fairly gross level.

Generally speaking, the faunal remains from the site suggest that the occupants were exploiting a variety of local habitats, including woodland, brushy scrub, marshlands, inland swamps, shoreline tidal flats and channels, and even perhaps open Bay waters. Tables 10-13, taken largely from Miller (1984:123-136), shows the normal habitat preferences of the species identified on the site.

It is likely, in fact, that the abundance of small local "micro-environments" in the region was a major reason for its settlement in the first place. It has long been recognized that archaeological sites, both prehistoric and historic, tend to cluster at the junction of environmental zones, where there is access not only to the main resources of those zones, but also to the abundance of wildlife that inhabits the transitions. Bay and stream shorelines are among the most prolific of all transition environments—a fact recognized not only by the earliest Virginia settlers, but also by settlers during later periods.

In these highly-diverse "ecotones," the historic-period settlements were often built atop or near the remains of earlier prehistoric occupations. This is of course the case at the Hampton University site, where the historic occupation was located only some 1000 feet south of a cluster of Middle Woodland-period prehistoric sites, and in fact, it is likely that HT55 itself has probably obliterated most traces of an earlier prehistoric site in the same location.

The Hampton University assemblage does suggest that the enormous species diversity in the area of the site was being exploited. At least forty-two wild taxa were used in some way. Nonetheless, analysis of the data leads inescapably to the conclusion that it was domestic livestock, and not the natural products of the local environment, that was the subsistence base of the occupants. Despite the faunal diversity of the area, the settlers relied heavily on cattle, and to a lesser extent pigs, and almost undoubtedly shaped their own environment accordingly—clearing fields for grazing, protecting them from predators, and foregoing a certain amount of hunting and fishing.

The heavy reliance on domestic livestock—a trend which would become more and more pervasive in the eighteenth century—may have several causes. Miller (1984) suggested three: that livestock was a form of economic security in a largely exchange-based economy; that they provided a means of passing on an inheritance to one's children; and that they provided the settlers with a subsistence system which was very much like the one that they or their parents grew up with in England. In the end it is this last reason which may be most important of all; by raising livestock in a traditional British manner, the occupants of the site were able to subsist without relying solely on a rich, but still probably unfamiliar, local environment—one which in addition was subject to periodic seasonal fluctuations in both quantity and diversity.

Nonetheless, the rearing of domestic livestock did not totally obviate the effect of the environment on the site's occupants. It appears that there were plagues throughout the seventeenth century, among them a major cattle plague in 1672-73 that "killed many thousands of animals in Virginia and Maryland"(Miller 1984:297) . At such times 91

Seasonality and Scheduling

Nonetheless, the presence of migratory waterfowl and other transients in the assemblage suggests that some clues can be discovered regarding seasonality and scheduling. Seasonality refers to the periodic fluctuations in the variety and nature of the resources themselves, as conditioned by migrations, reproduction, and variation in the animal's yearly cycle. Scheduling, on the other hand, is defined by Kent Flannery (1968:227) as "a cultural activity which resolved conflicts between procurement systems"—in other words, a mechanism by which humans were able to effectively make selective use of seasonally-varying resources in a relatively efficient manner.

A brief look at the assemblage indicates clearly that, like most in the Chesapeake, the site was occupied year-round. It is likely, of course, that wild resources such as migratory waterfowl were exploited most heavily at particular times of the year, and that at these times other resources were exploited a little less heavily (a scheduling decision). This is even true in a subsistence economy based on domestic livestock; since meat preservation was difficult, most animals were probably slaughtered in the late fall or early winter when the colder temperatures would serve to keep the meat somewhat fresher.

It appears that this slaughtered livestock was the majority of the diet throughout the winter and early spring, and that there was a shift toward a greater reliance on wild mammals, turtles, and fish in the summer and early fall. Migratory waterfowl were probably taken at their periods of greatest abundance, in the spring and fall (Miller 1984:272273).

This seasonal round can be used to investigate larger questions as well. Miller (1984) was able to define the depositional period of a variety of features using the faunal remains which they contained. Such analyses of course presume that the feature was filled relatively quickly, and that the fill was not redeposited soil from elsewhere in the yard (which could have been originally deposited during another season). Thus a seasonal determination is possible only for features such as the trash pits, and not the well, which was probably filled using nearby sheet refuse and trash deposits that had built up over a period of years.

Several of the trash pits from the Hampton University site were analyzed in this way. Trash Pit A, for example, contained the remains of several fish (striped bass, sheepshead, black drum, Atlantic croaker, and seatrout), as well as turtles including stinkpot and diamondback terrapin. A single bone from a mallard or domestic duck (Anas platyrhynchos) represents the only possible example of migratory waterfowl. In composition this assemblage appears to most resemble Feature 28A from Bennett Farm Phase I, which Miller (1984:439) suggests was deposited in the summer.

Trash Pit B, on the other hand, contained an even wider variety of fish (sturgeon, catfish, striped bass, sheepshead, black drum, Atlantic croaker, and seatrout), along with stinkpot, Eastern mud turtle, and several snakes. Migratory waterfowl include Canada goose and mallard or domestic duck. Pig bones were relatively common (51 out of 502 identified to Class), but only six cow bones were recovered. This relative scarcity of domestic livestock, and presence of migratory waterfowl, turtles, and fish, suggests a spring to summer deposition.

Other trash pits suggest a fall to winter deposition. Trash Pits C, E, G, and H, for example, include almost all domestic livestock. Although relatively few bones were recovered in some of these features, it appears that they were deposited during a period when few waterfowl or fish were being taken.

Finally, Trash Pit D suggests deposition over a relatively protracted period. A variety of fish were found, along with two types of turtles—suggesting deposition in the spring or summer. However, domestic livestock, particularly cow and pig, were fairly abundant as well, suggesting some deposition in the winter.

Conclusions

The faunal remains of the Hampton University site reveal important information about diet, environmental utilization, and animal husbandry in the period 1620-1660. They do not, however, provide the complete picture regarding food and foodways; additional data from oyster shell analysis, paleobotanical research, and documentary studies must be factored in as well.

The information gathered at the Hampton University site suggests that, as on most contemporary archaeological sites, domestic livestock was by far the most significant food source, accounting for the vast majority of the meat consumed. Significant contributions, however, were made by wild local resources, including deer, small mammals, wild terrestrial game birds, migratory waterfowl, turtles, and a variety of near-shore fishes. The occupants apparently utilized most of the wide range of ecological communities in the area, taking advantage of the diversity of resources that had naturally been presented to them.

Analysis

A number of attributes were examined and recorded for the samples. First, it was noted whether a shell was an upper (right) or a lower (left) valve. Uppers are distinguishable by a ridge in the center of the hinge which fits into a corresponding groove in the lower hinge. Uppers are also typically flatter than the more cup-shaped lowers and they also usually have a smoother, whiter surface which has been less pockmarked by other marine organisms.

Only valves which had a complete hinge were examined. First it was noted whether the valve itself was complete. If a valve was complete its height-length ratio (HLR) was generally calculated to determine the environment in which the oyster grew, since an oyster's shape is influenced by its environment. Oysters have been categorized into four types determined by their HLR and other physical characteristics (Kent n.d.:30).

As described by Brett Kent, an oyster which grows on a firm sandy substrate will develop a more or less round profile with a HLR of less than 1.3, the height being measured from the hinge down and the length being measured perpendicular to the height at the widest point across (Figure 37). This type of shell is referred to as a "sand oyster." Sand oysters usually grow individually, unattached to other oysters. "Channel oysters" have a HLR of greater than 2.0, which means that they are very long and narrow. They are long because they grow in deep water on a muddy bottom and so they have to reach up through the mud to get clear water. Another type of oyster which has a HLR of greater than 2.0 is a "reef oyster." It is easy to distinguish from a channel oyster because it grows in clusters instead of individually and it is proportionally much smaller than the large channel oysters. The last type of oysters is the type which was most commonly Figure 37. Oyster shell anatomy. found at HT55. This is the "bed oyster," which has a HLR of between 1.3 and 2.0; it grows on a mixture of sand and mud and it may or may not be clustered to other oysters.

Other attributes were examined as well. Clustering was noted by the presence or absence of other oyster valves or attachment scars on the lower valves for over half their length (Kent n.d.:33). Ribbing and coloring was also examined. These two characteristics develop on valves as a result of how much sunlight an oyster receives. Therefore, sand oysters, which typically develop in intertidal areas, are much more likely to exhibit deep purple coloring on their exterior surface and strong radial ribs which originate at the shell's hinge than are channel oysters, which are found in deeper water (Kent n.d.:30). This ribbing and coloring also is more apparent on lower valves than on the uppers, which are typically smoother.

Oysters can tolerate a wide range of salinity so, just as there are four general types of environment in which an oyster grows, oysters have also been categorized into four regimes based on the relative salinity of the water in which they grew. Basically, according to Kent, these are indicated by the presence or absence of boreholes produced by small (Cliona trutti-type) sponges and large (Cliona celata) sponges, both of which attack the oyster valves. If there are no boreholes in the shell produced by these sponges the water in which the oyster lived was not very saline. This is because, in low-salinity environments, the sponges could not live; even the oysters would 97 find this a particularly stressful environment (Doms and Custer 1983:8), since the variety of plant and animal life in areas of low salinity is less than in adjacent areas of higher salinity (Galtsoff, 1964:398). These low-salinity zones are defined as Salinity Regime I, where the water would have a salinity of less than 10 parts per thousand (ppt) for about half the year, and seldom becomes greater than 20 ppt. If the oyster valves exhibit boreholes from only the smaller sponges, they probably came from Salinity Regime II, where the salinity would be less than 10 ppt approximately one-quarter of the year, less than 15 ppt half of the year, and greater than 20 ppt at some times. Oysters from Salinity Regime III will show boreholes mostly from the small sponges, but will also have some large boreholes. They would have been found in water which had salinity less than 15 ppt occasionally and greater than 20 ppt for one-quarter to one-half of the year. Finally, oysters which grew in very saline water, where the salinity seldom fell below 15 ppt and usually was above 20 ppt, are defined as coming from Salinity Regime IV (Kent n.d.:41).

The presence or absence of other boreholes caused by polydora worms and boring clams was also recorded. Polydora worms, which live in subtidal mud, enter the oyster at the margin between the two valves and leave distinctive double holes around the edge of the oyster. Boring clams, which are found in water with low salinity, leave a hole which is relatively small at its mouth but which opens up to a large cavity within the shell. However, unless there is some erosion of the oyster shell which makes this cavity visible, it tends to be rather hard to determine whether a hole was caused by a boring clam or by a C. celata sponge.

It was also noted whether or not there were any barnacles or net-like bryozoan sponges attached to the oysters, although it is unclear as to whether these are very useful as indicators of environment or salinity.

Results

From this data several patterns emerged (Figure 38). First, 57% of the shells were classified as bed oysters, which are found in a mixture of sand and mud. The fact that 86% of the sample had polydora worm holes also suggests that the shells grew in a muddy subtidal environment. Less than 20% of the shells exhibited moderate to heavy ribbing, and less than 5% exhibited moderate to heavy coloring which also indicates a subtidal environment. Only 12% of the sample were classified as sand oysters, and less than 6% were classified as channel oysters. Thus, most of the oysters did not come from either intertidal zones right along the shore or from very deep waters. Over 70% of the sample exhibited either no or only small cliona holes, indicating water of low salinity. However, boring clams, which are found in water Figure 38. Oyster types and salinity regimes. 98 of low salinity, were definitely noted on only 4% of the sample, although, as was noted, boring clam holes were difficult to identify.

Site HT55 is located between two small creeks or inlets off the Hampton River, which would provide the best conditions for the growth and development of an oyster population (Goode 1887:731; Galtsoff 1964:400). Kendall and Haven's 1982 report on oyster leases along the Hampton River supports basically what the oysters themselves indicated—that, although the main channel of the river varies in depth from approximately nine to fourteen feet, most of the bottom is shallower than five feet. Hence, historically, river oysters must have been gathered in fairly shallow waters.

Kendall and Haven also stated that "the bottom was mud except for some sand near the shore." This should come as no surprise, since most of the oysters found at HT55 were bed oysters, which grow in a mixture of sand and mud. It is interesting to note that of the four oyster leases which were examined in 1982 by Kendall and Haven, two of which are on the same side of the river as HT55 and two of which are on the opposite shore, the one which has the greatest density of oysters per acre (67%) is the one located adjacent to the site.

Although it may be tempting to project backward in time and say that this may have also been the case when the site was occupied, there are a number of dynamic factors which influence the growth and development of an oyster community. Factors such as the character of the bottom, water movements, salinity and temperature of the water, and food supplies all affect the oysters, while sedimentation, pollution, competition, disease and predation will all work against the population (Galtsoff 1964:398-399). In addition, humans, while essentially predators, may also aid oyster beds through seeding to extend their natural environment.

Historical Research

From historical research it seems that oysters were extremely plentiful in Tidewater Virginia. Records from as early as 1607 comment on how the "oysters lay on the ground as thick as stones" (Wharton 1957:5). In fact, during the seventeenth century, there are numerous accounts of settlers and "Indians" surviving on little more than oysters. Apparently oysters were generally easy to obtain even without a boat or canoe, and oysters in the lower James River could even be picked over by hand at low tide (Wharton 1957:26-27). With a boat even more oysters could be gathered with ease, as it was recorded in 1687 that "almost every Saturday my host…had only to send one of his servants in one of the small boats and two hours after ebb tide he brought it back full. These boats…can hold as many as fourteen people and twenty-five hundredweight of merchandise" (Wharton 1957:35).

It was also recorded that oysters were usually caught on Saturday at ebb tide with tongs (Wharton 1957:37-38). The practice of tonging for oysters goes back to the Native Americans, who taught the colonists in Tidewater Virginia how to use tongs not unlike those still in use today (Jackson 1988:66; see Figure 39). This was probably a method which was in use near 44HT55 to gather oysters. By 1701 travellers to the area still were amazed at the incredible abundance of oysters and told of banks of oysters so large that ships must avoid them (Pearson 1942:218; Jackson 1988:63). However, there was little organized industry dealing with oysters until the nineteenth century (Wharton 1957:41) which made unlimited opportunities for exploitation during the time site HT55 was occupied.

Oysters played an important part in the diet of Virginians and are mentioned again and again in the historical literature. Generally, oysters were consumed during the fall and winter months and were avoided during the summer (or, more colloquially, in any month without an "R" in its name because it was considered to be unhealthy to eat them then Pearson 1942:217; Jackson 1988:62) and they were said to cause fevers (Wharton 1957: 38). However, historic references were found which refer to eating oysters in every month except May, April and August. It seems that in the early seventeenth century oysters were eaten in 99 the summer just to keep from starving, and that by the late eighteenth century some people came to disregard the widespread superstition about not eating oysters in the warm months.

Actually, there are valid reasons for not consuming oysters during this period of the year, because the oysters are spawning or have just spawned and thus are played out and less tasty (Jackson 1988:62). Oysters are at their best as energy-producing food between August and November because of the glycogen which they have stored during the summer (Russell 1923:9). In addition, the oyster responds to the salinity of the water; low salinity makes the oyster swell up by absorption of water, and the opposite effect occurs when the salinity increases, generally from June or July until October, apparently as a result of altered osmotic pressure. The percentage of proteins in oysters is highest in the spring and summer and lowest during the rest of the year, while the percentage of carbohydrates varies in an inverse manner (Claassen 1986:34). The "fattening" of oysters in the autumn is as a result of the accumulation of these carbohydrates (Russell 1923:4).

Oysters were, and still are, prepared in a wide variety of ways in Tidewater Virginia. Historically, oysters were smoked and dried, served raw, stewed caked in fritters, made into dressing, fried, pickled, fricasseed, scalloped, curried, deviled, broiled and brochetted, in salads and soups and so on (Wharton 1957:46; Jackson 1988:68). They were also roasted, and oyster roasts have been a Virginia institution which was inherited from the Native Americans. Four percent of the oysters sampled at 44HT55 did show signs of being roasted, but it is unknown how the other oysters were served. The contexts which contained the most roasted oysters were Context 39, with a total of seventy-eight oyster valves (38% of the shells from that feature); Context 3, with a total of seventeen valves (7% of that feature); and Context 237, with fifteen shells (13% of that layer). By contrast, Context 26 contained thirty-five roasted oysters, only 2% of the 1708 valves which were analyzed from that feature.

Figure 39. Oyster tongs.

Conclusions

The sample of oyster valves from HT55 seems to comprise a predictably homogeneous collection as analyzed so far. The majority of shells from all eleven contexts were categorized as bed oysters, which are found in a mixture of sand and mud. Over 86% of the total number of shells analyzed contained polydora worm holes; since these worms live in mud they indicate a muddy environment for the oysters. In every analyzed context, the majority of the shells contained no cliona holes, which indicates that they grew in water of relatively low salinity. There is also consistency in the amount of ribbing and coloring; most have little or none, which means the oysters were found in deep enough water so as not to receive much sunlight. It would be interesting to conduct seasonality studies on the shells to observe what other patterns, if any, emerge. Unfortunately, time and 100 fiscal constraints precluded this exhaustive an analysis.

This study of oyster shells from HT55 is part of a continuing attempt by Colonial Williamsburg's Department of Archaeological Research to analyze oyster shells from historic sites; two previous studies, yet unpublished, have been conducted on shells from the Peyton Randolph and Shields Tavern properties within the Historic Area. As we realize the potential of what it is oysters can tell us about how, from where and when they were gathered originally, it is hoped that with more refined collection strategies further research we can glean even more information from shells collected from future excavations.

Methods

All 114 paleobotanical samples were scanned for carbonized seed, wood, cultigen and plant remains. An American Optical Microstar Stereoscope with a magnification range of 7-45X was used in the identification of all carbonized plant remains. Identification of paleoethnobotanical material was made through comparison of archaeological material to known modern material from the University of Missouri-Columbia/American Archaeology Division Paleoethnobotany

In general, identification of seeds is accomplished by comparing morphological characteristics of the archaeological sample to those of known, modern taxa. Species differences are not always apparent and critical diagnostic characters (e.g., awns, outer seed coat, and color) often are lost in carbonization. Wood charcoal is identified by examination of the internal anatomy of a specimen and the use of a key developed for carbonized-wood identification. The arrangement of vessel elements, rays, and parenchyma tissue are used in the identification of archaeological specimens.

The Paleoethnobotanical Record

Plant remains from archaeological sites, whether from middens, hearths or features, represent only a fraction of the possible ranges of plants used by a particular human group. It is important to remember that the paleoethnobotanical record, i.e., plant remains and their depositional contexts at archaeological sites, is a product of the differential occurrence of—and the interaction among— various environmental and behavioral site-formation processes (Voigt 1989). Some of these formational processes include differential preservation among plant taxa in their carbonized and uncarbonized states, chances of the remains entering the archaeological record through intentional discard or accidental carbonization, and pedological factors (Asch and Asch 1981; Johannessen 1984; Lopinot 1988; Wagner 1989).

Assumptions about the relative economic importance of plant taxa to human groups are based on: (a) the frequency of co-occurrence of the same taxa at different sites, (b) the relative abundance of different plant taxa at a site (i.e., their ubiquity), (c) spatial and temporal patterning of the paleoethnobotanical data, (d) ethnohistorical, historical, and ethnographic data, and (e) taxon-specific indicators of cultivation or domestication (Voigt 1989). However, such things as "the number of seeds recovered" can serve neither as an adequate nor as an accurate measure of the relative importance of a taxon in the diet of a particular human group.

Taxa Identified—Wood Charcoal and Carbonized Nuts

Much of the information from each taxon is from Harow and Harrar (1969). When possible, wood charcoal and nuts were identified to the species level. However, the carbonization process often leads to the loss of important anatomical characters and this results in specimens being classified only to the family or genus level.

Summary of Results

Feature Correlations

The crossmending analysis of the ceramic remains revealed several contemporaneous relationships among major features on the site. The well and the large trash pits (Trash Pits A and B), located just north of Structure A, were proven contemporaneous by crossmends in a locally-made coarse earthenware vessel believed to be a strainer (Catalog Number 1143; see Photo 15) and by crossmends in four Portuguese tin-enamelled ware plates (Catalog Numbers 1037, 1038, 1039, and 1040). Trash Pits D and H were proven contemporaneous by a crossmend in a Westerwald stoneware jug (Catalog Number 1071). Trash Pit A and the cellar fill of Structure A were proven contemporaneous by a crossmend in a vessel believed to be an English stoneware butterpot (Catalog Number 1085) and another in a coarse earthenware storage jar (Catalog Number 1140). Three small trash pits and the western fence slot complex (Trash Pits F, G, H and Slot Fence A) were proven related by a crossmend in a large coarse earthenware milk or cream pan (Catalog Number 1094). Trash Pit F and Trash Pit A are related through a crossmend in a coarse earthenware porringer (Catalog Number 1153). Trash Pits B and F were proven contemporaneous by a crossmend in a coarse earthenware unglazed flask or bottle (Catalog Number 1155).

Through this analysis, it was established that the well to the southwest of the site, the cellar fill of Structure A, the large trash pits just north of Structure A, and the smaller trash pits and fence 117 Photo 15. Red earthenware strainer, interior and side views with profile. complex to the west of the site were contemporaneous features.

The white clay tobacco pipes from HT55 were also crossmended and yielded further evidence of contemporaniety among features on the site. The presence of a marked tobacco pipe bowl (Catalog Number 6123), which may be attributable to the London pipemaker William Collins, who died in 1686 (Oswald 1970:138), provides a terminus ante quem for the construction of Structure A (for details, see the discussion of pipe stem crossmending below). The ceramic artifacts retrieved from Structure A and the major features of the site suggest a date of construction some time during the first half of the seventeenth century.

Ware Type Profile

Site HT55 produced 128 identifiable ceramic vessels (Table 21). The majority of those vessels (67, or 52.3%) were coarsewares. These wheel-thrown vessels were composed of an earthenware body which was often coated with a clear lead-fluxed glaze. Because they were easy to produce and relatively inexpensive to purchase (I. Noël Hume 1969:102), coarsewares often formed the core of a colonial-period vessel assemblage, especially of utilitarian wares. It is not coincidental, then, that the overwhelming majority of vessels associated with food preparation, processing, and storage in the HT55 assemblage were coarsewares. Locally-made red coarsewares comprised 38 of the 67 coarseware vessels. The remaining 29 vessels were imported from England, Italy, Spain, and possibly from Central America (Figure 41).

One small, thinly-potted, unglazed earthenware vessel (Catalog Number 1155; see Photo 16) was retrieved from Trash Pits B and F. This vessel is unique among the ceramic forms not only in its form, a short, bulbous flask shape, but also in the clay body from which it was made. This red earthenware contains a large proportion of micaeous material in the clay body, which is easily visible to the unaided eye. It is possible that this vessel may be of Central American origin, possibly from Mexico (I. Noël Hume, personal communication).

Note: Chinese porcelain was also recovered from HT55, but the fragments were too small to make a positive vessel form identification. Accordingly, they have been excluded from this listing.

Figure 41. Percentages of ware types represented.

While coarseware vessels were the most numerous, tin-enamelled wares were also well represented within the ceramic vessel assemblage. Delftware, faience, and majolica were among the several varieties of tin-enamelled ware that were encountered. Made from an earthenware clay body, tin-enamelled wares are distinguished by their glaze, which is obtained by opacifying a clear lead glaze with tin oxide. Compounds of cobalt, manganese, iron, copper, or antimony were commonly painted on such wares to produce colorful decorations. In the HT55 assemblage, these wares numbered 50 vessels, comprising 39.1% of the total.

Stonewares, including vessels produced in the Rhineland and decorated with either cobalt blue on a gray body or iron oxide brown, were not well represented in the vessel assemblage. Similarly, only one stoneware vessel identified as a Staffordshire butterpot was identified. These wares have a high-fired ceramic body that is heated to the point of vitrification. Stonewares are generally glazed with common salt to give their exteriors a characteristic orange-peel texture. Stonewares, primarily German in origin, have been found on many seventeenth-century sites in the Chesapeake. These durable European wares were imported to the colonies through England in considerable quantities throughout this period (I. Noël Hume 1978:276). Therefore, it is surprising that there were only 11 stoneware vessels in the assemblage, 8.6% of the total.

Form and Function Analysis

To facilitate the analysis of the ceramic assemblage, vessels were further categorized according to their form and possible function (Tables 22 and 23). Of the 128 vessels identified initially, the five unidentified forms and the crucible were omitted from this section of the analysis. Each of the remaining 122 vessels was examined individually to judge its exact shape and size based on diagnostic characteristics of rim, base, and body. 120 Photo 16. Unglazed earthenware flask, Mexican (?), with profile. These evaluations were used, along with the ceramic type, to determine a specific form designation for each identifiable vessel. Vessels of like physical appearance were then grouped together according to their functional role within the domestic setting, and six functional categories were created. The largest group consisted of those vessels used for food serving and/or consumption. It included such seventeenth-century tablewares as bowls, dishes, plates, porringers, tygs, mugs, and assorted hollowwares that were used for the presentation or consumption of food or beverages. Storage vessels were those forms such as jugs, olive jars, butter pots, storage jars and flasks, which were used for the containment or transportation of food or liquid awaiting consumption. Several costrels and unidentified hollowwares were placed in a separate category of serving/storage containers to reflect their dual functions. Vessel forms used for domestic food processing, preparation, and cooking included pipkins, pans, and colanders/strainers.

While the functional nature of most of the ceramic vessels involved the preparation, storage, or consumption of food or drink, a number of non-food related vessels attest to several of the many other domestic needs on the site. Numerous drug pots used for the storage of dry medicinal preparations were placed in a separate category of pharmaceutical vessels. The two chamber pots were likewise categorized separately as toiletry vessels.

According to these functional categories, the food serving vessels constituted the largest percentage (39.3%) of the total ceramic vessel assemblage (Figure 42). Of the 122 vessels studied and regrouped by their function, forty-eight were probably used for serving food. Vessel forms in this category included: 2 dishes (1.6%), 3 pans/ dishes (2.5%), 10 bowls (8.2%), 28 plates (23%), 1 plate/shallow bowl (0.8%), 2 porringers (1.6%), 1 mug (0.8%), and 1 tyg (0.8%). Within this category, it is interesting to note that all of the 28 identified plates are tin-enamelled wares. When the 28 plates are added to the remaining 18 food serving vessels, these 46 food serving and consumption vessels contrast dramatically with the two ceramic drinking vessels which were identified on the site.

Such an uneven ratio of food to beverage consumption vessels might be explained by the occupants' use of leather, wood, or pewter drinking vessels. Such materials were commonly used for drinking vessels in the seventeenth century; 121

Figure 42. Vessel form/function.

Vessels used for food storage made up the next largest group with 23.7% of the total. Of the 29 vessels in this category, there were 12 storage jars (9.8%), 10 jugs (8.2%), 3 bellarmine jugs (2.5%), 2 olive jars (1.6%), one flask (0.8%), and one butter pot (0.8%). Coarsewares, tin-enamelled wares, and stonewares were all present in this functional category; they included both locally-made and imported varieties. A separate category, which contained 3 costrels (2.5%) and 3 miscellaneous hollowwares (2.5%), was created to reflect both the food serving and storage capacities of these vessels. This category contained solely imported wares and accounted for only 5% of the total vessel assemblage.

The 27 vessels associated with the processing and preparation of food represented 22.1% of the total. Eighteen pans (14.8%), 6 pipkins (4.9%), 1 pipkin lid (0.8%), 1 colander (0.8%), and 1 possible strainer (0.8%) made up the vessel forms in this category. It is noteworthy that of these 27 vessels, two-thirds were identified as pans. Such a large quantity of pans of the size and configuration traditionally used for processing dairy products suggests that dairy-related activities may have been an important part of the seventeenth-century domestic routine on this particular site. In addition, a large hollowware vessel with a single centrally-located drain hole in its base is believed to be a strainer used to separate whey from curds. This would have been an indispensable vessel in the production of cheese.

Local Coarsewares

The extensive use of locally-made earthenware vessels indicates the possible presence of itinerant, skilled potters who filled a need for utilitarian vessels not satisfied by the imported vessels. The production of locally-made pottery has been documented by Noël Hume at several of the major sites at Martin's Hundred, where spectrographic analysis of the clay bodies proved them to be of Virginia manufacture (I. Noël Hume 1982:105108). A wide range of ceramic forms and pottery wasters, believed to have been produced and discarded 125 Photo 17. Thumb-impressed earthenware storage jar, with profile. 126 Photo 18. Red earthenware pipkin handle, with profile. Photo 19 (below). Unglazed pipkin lid, with profile. 127 on or near Site "C" (1618-1622) and Site "B" (post-1631), have been retrieved from the Martin's Hundred site, but unfortunately, the kiln sites have not yet been discovered (I. Noël Hume 1982). Coarse earthenwares with the same visual characteristics as those from Martin's Hundred were represented among the HT55 ceramic assemblage. There are striking similarities in the appearance of the various locally-made ceramics, in terms of form and decorative characteristics on all of the major sites along the James River (see Photos 17 and 18). Additionally, coarse earthenwares with the same distinctive "Martin's Hundred" clay body (the terminology used by Bly Bogley Straub in finds inventories for the Causey's Care site), though with somewhat different potting styles, were excavated from the College Landing site further up the James River.

Bly Bogley Straub and other researchers have used the term "Jamestown" to refer to a particular form of domestically-produced coarse earthenware body characterized by a fine, sandy body, which appears as a pale, dusty brownish-grey to a rosy-peach color. The glaze most frequently associated with this body is a soft, dull caramel or greenish caramel-colored lead glaze which is usually applied sparingly to the interior surfaces of the utilitarian forms. This ware type has been documented on the Causey's Care and Jamestown sites, and appears at HT55 in a number of vessel forms. One of the most notable is an unglazed pipkin lid (Catalog Number 1181; Photo 19).

In the author's opinion, the so-called "Martin's Hundred" and Jamestown" wares are, in fact, the same. The differences in glaze color and clay body hardness and color can be attributed to minor variances in firing and chemical composition of the clay used in the manufacture. The control that could be exercised by an itinerant potter who fired his wares in temporary, surface-built kilns may have varied sufficiently to result in immature wares found at Jamestown, HT55, and Causey's Care, but not at Martin's Hundred. Spectrographic analysis of the clay bodies of these locally-made wares, in comparison with documented wares from abroad, is the only reliable method to determine their true origin.

The vessel forms which appear in the so-called "Jamestown ware" and the more common local red earthenwares are very similar in character. The rim configurations of these bowls, storage jars, and milk pan forms are also strikingly similar.

A nearly complete red earthenware storage jar (Catalog Number 1141), which is believed to be of local manufacture, was found in the well at HT55. This vessel (Photo 20) is unique in the configuration of the rim which is quite angular with an internal "step" for a lid.

Imported Earthenwares

Tin-Enamelled Wares

Stonewares

Site HT55 yielded relatively few stoneware vessels, all but one of which is believed to be of German origin. The "Bellarmine" or Bartmann jug found on other early seventeenth-century sites along the James River was primarily a food storage and serving vessel. The fragments of this form 136 Photo 33. English delftware polychrome plate, interior and exterior views with profile. Photo 34. Dutch delftware dish base, interior and exterior views with profile. 137 (Catalog Number 1075; Photo 35) found in the well at Hampton University bear the typical bearded face on the neck. According to Noël Hume, the characteristics of the face become stylized and grotesque by the second quarter of seventeenth century (I. Noël Hume 1969:57). The fragment retrieved from HT55 shows some of the grotesque, stylized facial features, but retains a fluidity of line and form in the beard. A similar Frechen brown stoneware jug fragment was retrieved from the College Landing site; it has the same stylized facial features, but with a flowing beard (Pittman, in Edwards 1987:56-57).

In addition to brown stonewares, HT55 contained some fragments of blue and grey Westerwald stoneware jugs of the type found by Noël Hume at Martin's Hundred. These impressively-decorated jugs (Catalog Number 1071; Photo 36) are illustrated and described in detail by Noël Hume (I. Noël Hume 1969, 1982:139). The fragments recovered from HT55 are decorated with incised lines, sprig-molded appliques, and carefully hand-painted areas. Fragments of these food storage and serving vessels have also been found at the Causey's Care site and at Jamestown.

Porcelain

Site HT55, like most other early seventeenth-century sites in the Chesapeake, yielded very few fragments of Chinese porcelain. This imported ware was initially marketed to the public by the Dutch as early as 1602 (van der Pijl-Ketel 1982:9) and was extremely popular among the wealthier purchasers of ceramic wares. As this ware did not reach Europe and consequently English North America, in large quantities until the mid-seventeenth century (van der Pijl-Ketel 1982:28), it is not surprising to find only three small fragments in discrete archaeological contexts on a site as early as HT55. The fragments were too small to make a vessel form determination, but all of the fragments are painted with the distinctive cobalt blue hand-painted decoration commonly found in the first twenty years of the seventeenth century (J. Curtis, personal communication). The presence of this expensive ware is yet another indication of the relative wealth of the early HT55 colonists.

Photo 35. Bellarmine jug neck, with profile.

Photo 36. Westerwald blue and grey stoneware jug neck fragments, with profile.

Glass Containers

Case bottles account for a significant number of storage vessels within the assemblage. For the 139 most part, case bottles were used to store and transport beverages. The seventeenth-century case bottle, was a "… common large bottle was blown into a square-sided mold and had a nearly flat base and a short neck with an everted lip, the latter feature frequently concealed beneath a threaded pewter collar and cap"(I. Noël Hume 1969:62 ). Based on stylistic attributes, these bottles were probably imported from England (I. Noël Hume 1969:70), but because of the relative thinness of their metal, they do not survive intact as well as the round or globular bottles of the second half of the century. Case bottles are helpful in establishing depositional dates because they were used prior to the round-bodied, green-glass bottles which appeared only after the beginning of the second half of the seventeenth century (I. Noël Hume 1969:62).

The artifact inventory from HT55 shows a total of 2703 fragments of case bottle composed mostly of a pale straw green metal. Of these fragments, 22 empontilled bases and 37 necks were identified. Three of these necks came from the uppermost layers of two trash pits (Trash Pits A and F), which contained intrusive material in their upper layers, probably generated by plow disturbance. These three necks were not excluded from the minimum vessel count because the nature of the features suggest that they were original components of those features. The work done by Riordan and Pogue in associating plowzone artifacts with discrete features provides the rationale for including the questionable necks (Riordan 1988; Pogue 1988). Case bottle fragments were recovered from most of the major features on the site which were proven contemporaneous by ceramic crossmends. As there were more necks than bases, the larger quantity was used to establish the minimum vessel count for the site.

Window Glass

Sixteen fragments of window glass were found at the Hampton University site. While window glass fragments were a significant portion of the glass assemblage, they were not used for interpreting occupational dates for the site. Since "the process [for making window glass] changed little over the centuries so it is not generally possible to come to conclusions about age…"(Frank 1982:142) . First imported to the colonies in the 1620s (Wilson 1976:152), the presence of window glass as an architectural element in early seventeenth-century colonial Virginia is a good indicator of wealth or status.

The use of casement windows in the seventeenth century had its origins in medieval architecture (Wilson 1976:156). Indeed, the concept of earthfast construction techniques which were used at the Hampton University site has its origin in Anglo-Saxon and Medieval England (Carson et al. 1981:136). As has been ably described by Carson et al. (1981), the earthfast structures that grew out of this European building tradition were not temporary structures, but were obviously intended for long-term habitation. According to Carson et al. (1981:139-140), "whether they came to farm, to trade, or to follow some other occupation, immigrants expected to make an investment in buildings."

The addition of imported window glass, lathe-marked plaster (Catalog Number 5001; see Photo 44), and roofing tiles (Catalog Number 1240 and 1241; see Photo 45) to these earthfast structures speaks of an uncommon affluence of the HT55 residents in the early years of colonization.

The window pane fragments excavated from the site may be described using terminology derived from Richard Neve's Builder's Dictionary of 1736. In this work, diamond-shaped panes with an overall length of 6 inches, maximum width of 4.8 inches, and having an acute angle of 77 degrees and an obtuse angle of 102 degrees are identified as a "square quarrel." The analysis of the HT55 window pane fragments yielded two fragments with sufficiently well-defined cut edges whose angles closely match the square quarrel angle measurements (see Figure 43). Additionally, one fragment, in the form of an isosceles triangle, perfectly fits within Neve's dimensions for 140 Figure 43. Window pane angles. a square quarrel with an obtuse angle of 104 degrees. This "half square quarrel"¹ may have been used along the edge of the casement window to square off the sides. Two other smaller fragments with relatively well-defined edges and measurable angles of 74 degrees each do not comfortably fit within Neve's prescribed measurements, but the deterioration of the metal along the edges may account for this discrepancy.

Although 16 fragments of window glass would not be sufficient to make even one small window, these fragments in association with 24 fragments of window leading confirm the presence of these architectural features. One of the fragments of window glass (Catalog Number 260), a small equilateral triangle of glass measuring approximately one and three-quarters inch on each side, was completely encircled by its original turned-lead trim. The recent discovery of dates and other inscriptions inside turned window leads by Noël Hume and subsequent work by such researchers as Susan Hanna et al. have proven a means of applying dates after which these architectural elements were assembled (I. Noël Hume 1982; Hanna 1986:303-309). Because the dates of the inscriptions found in drawn window leads were imparted by a removable element of the glazier's vice, which may have been used over a long period of time, archaeologists must use caution when applying these dates to the buildings that they excavate. All of the drawn lead fragments from HT55 which have undergone conservation have been clean and, unfortunately, show no inscriptions of any kind.

Mirror Glass

Seven colorless, thin, flat fragments of glass were retrieved from various layers within the well. Al 141 though they have not retained any evidence of silvering, these fragments have been identified as mirror glass. Mirror glass was "a highly prized luxury" in the late sixteenth century (Godfrey 1975:236), and a relatively expensive and valued object for an early seventeenth-century colonial site. Thus a mirror or mirror glass found at HT55 would be the type of artifact indicative of wealth or high status.

Luxury Table Glass

Ten pieces of container glass, six colored and four colorless non-leaded fragments, were found during the course of excavation. Of the colorless, non-leaded fragments, one piece of polychrome enamelled glass and one piece with trailed decoration were identified (Catalog Number 193; see Photo 37). Additionally, one notable piece of dark green container glass (Catalog Number 5000) from Trash Pit F is remarkably similar to a base fragment of a drinking glass similar to one illustrated by Robert Charleston in his book, English Glass (Charleston 1984:44; Figure 44). The fragment exhibits marks of abrasion on the standing surface of the folded foot, as well as surface dimpling in the metal, which is indicative of tool manipulation during manufacture. The presence of this somewhat specialized vessel form in the assemblage is in marked contrast to the more commonplace case bottles.

Thirty-five pieces of table glass, nine colored pieces and twenty-six pieces of colorless non-lead glass were identified. Three of the fragments of colored glass were decorated with applied horizontal strings of the same color metal and appear to be similar to a sixteenth-century green glass goblet illustrated in "A Restoration Glass Hoard From Gracechurch St., London" (Oswald and Phillips 1949:30). Three other fragments of colorless, non-lead glass also had applied decoration. In this case, strings of opaque white glass had been vertically applied to the metal. Three small handle fragments of a dark black-green metal from unidentified vessel forms were also associated with this group.

Photo 37. String-trailed glass container fragment, with profile on projected vessel form.

Stemmed glassware was also represented in the assemblage. Most notable among these was one complete free blown base and stem made of a pale greenish-blue metal. This base (Catalog Number 036) exhibits a rather clumsily folded foot, a welded stem with a single ball-shaped knop (see Photo 38). A striking parallel to this folded foot and stem, like the drinking glass base 142 Figure 44. Drinking beaker foot fragment on projected vessel form. fragment mentioned above, is illustrated by Robert Charleston and is attributed to Gracechurch Street, London and is dated about 1600 (Charleston 1984:Plate 19d).

One other fragment of stemmed glassware in a colorless, non-leaded metal was retrieved. This fragment bears all of the characteristic crizzling and decomposition typical of excavated European or early English "cristallo" and is a fragment of a funnel-shaped solid knop. This piece of "luxury glass" (Charleston 1984:43) is important because it speaks of a level of affluence unexpected in early colonial Virginia. Furthermore, the cristallo is deeply rooted in the medieval tradition of glassmaking (Charleston 1984:45).

A diminutive fragment of a folded foot made of colorless non-leaded metal was also identified. It is uncertain from which vessel form this fragment comes; however, the delicate turning of the foot, the thinness of the metal, and the relatively small size of the form were unique among the glass assemblage at HT55. Finally, three fragments of colorless, nonleaded metal were retrieved from the well. Two of these fragments mend, and the third is non-contiguous. They are thought to be fragments of a wine glass stem of the type known as a "cigar-stem." A remarkably similar wine glass, documented by Robert Charleston, was attributed to the first half of the seventeenth century (Charleston 1984:Plate 15a).

A total of five fragments of thin blue glass were retrieved from Trash Pit A. These fragments, which do not crossmend, are remarkable for their thinness and distinctive color. Ivor Noël Hume suggests that they appear to be fragments from a French drinking goblet of cylindrical form with ball feet (I. Noël Hume, personal communication).

Utilitarian Glass

The handle of a linen smoother (Catalog Number 035) was retrieved from a post mold associated with Structure B (Photo 39). This specialized textile-related tool is discussed by Charleston and is found on English and Scandinavian sites as early as the thirteenth century (Charleston 1984:3738).

Harrington/Binford Stem-Hole
Diameter Analysis

A sample of 649 tobacco pipe stems from HT55 were subjected to the standard analytical procedures devised by J.C. Harrington and Lewis Binford, in order to determine a suggested date range of occupation (Harrington 1954, 1978) and a calculated mean date of occupation (Binford 1978). The pioneering work done by these two scholars has proven useful to many historical archaeologists who have excavated seventeenth-and eighteenth-century sites throughout the eastern United States, particularly along the James River. James Deetz used the Harrington method to convincingly define the shifting settlement patterns which existed at Flowerdew Hundred during its long period of occupation (Deetz 1987). In the same manner, the combined Harrington and Binford dating techniques were utilized in the analysis of HT55 tobacco stems to suggest temporal co-existence between the major features on the site and to indicate the range of occupation of the site as a whole.

As can be seen in the calculations and graphs below, the results of these computations suggest a Harrington range of occupation of the entire site predominantly in the period 1620 to 1650 (Table 25; Figure 45), gradually tapering off in the period 1650-1680. As the significant concentration occurs in the earlier period, it may be deduced that the heaviest occupation of the site was between 1620 and 1650. The calculations based on the Binford straight line regression formula provide a mean date of 1639. In this case, the two techniques would seem to agree and support the dating of the site which is indicated by the ceramic assemblage.

The Harrington computations for the five major features of the site including the well, and four trash pits, follow a similar pattern indicating a predominant occupation range between 1620 and 1650, with the exception of Trash Pit F which falls later in the 1650 to 1680 period (Figure 46). The Binford mean date for Trash Pit F is 1651, which does not fall near the midpoint of the date range, 1665, provided by the Harrington method. In this case, the two techniques would seem to disagree. This late deposition suggests that HT55 may have been occupied or at least used for trash disposal as late as the third quarter of the seventeenth century.

The statistical date for Trash Pit A is 1637, which exactly coincides with the calculated date for the well. This is not surprising in light of the fact that these two features were proven to be temporally related by a ceramic crossmend (Catalog Number 1143).

The quantities of tobacco pipe stems from the trash pits designated as Trash Pits A and F (40 and 36 pipe stems, respectively), were initially considered too small to yield reliable results from Harrington calculations (see Figure 48). However, the results from this analysis yielded two entirely different mean dates of 1651 for Trash Pit F and 1630 for Trash Pit H. Because these features were in very close proximity, and during excavation had been perceived as being related, they provided an opportunity to indicate the terminal point of occupation or trash disposal on this site. The variance of 21.5 years between the two features suggests two separate periods of deposition and further suggests that the occupation of the site extended into the third quarter of the seventeenth century.

The Harrington histograms for Trash Pits F and H indicate opposing trends between the two features. The date range for Trash Pit H falls into the 1620 to 1650 period, while Trash Pit F falls even later into the 1650 to 1680 period. When the two are combined, a bi-modal frequency appears to span the period from 1620 to 1680. The fact that there were no ceramic crossmends between these two closely positioned features strongly suggests that they were filled at separate times.

Imported and Marked Pipe Bowls

Table 26 shows the marked imported bowls and their provenience and suggested origin. The reader will note that this space is inadequate to fully 146

Figure 45. Harrington histogram—entire site.

Figure 46. Harrington histogram—site by feature.

Of the 53 marked pipe bowls, 10 (18.8%) are thought to be of English manufacture; 30 (56.6%) are thought to be Dutch, and the remaining 13 (24.5%) are of unknown origin (Figure 49).

The highest concentration of marked pipe bowls occurred in Trash Pit B, which is a large trash pit located just north of Structure A (Figure 50). This feature held 16 marked bowls—the majority being of Dutch origin. The deposition of 149 Figure 47. Harrington histograms—Well, Trash Pit A, and Trash Pit B. tobacco pipe fragments and other domestic refuse in this large feature located so near Structure A suggests the nature of trash disposal practiced by the inhabitants of the property either during the course of occupation or at the termination of residence. It is not surprising to note that four pipes with the same mark of "W C" were found in this trash pit and in a posthole of Structure A (Catalog Numbers 6117, 6119, 6121, and 6123). This mark has been attributed to the English pipemaker William Collins (Oswald 1970:138). Even though the maker's marks found on tobacco pipes are not always the best indicators of date of manufacture, this mark can at least provide a date before which deposition of refuse in the pit and the construction of Structure A could have taken place. Numerous examples of this mark have appeared on sites throughout the James River Basin with depositional dates ranging from circa 1620 to 1660 and are discussed below in the descriptive catalog of tobacco pipes.

The well held 10 marked bowls, but unfortunately only two could be associated with Dutch manufacture, one with English, and the majority 150 Figure 48. Harrington histograms—Trash Pit F, Trash Pit H, and Trash Pits F and H combined. (7) of unknown origin. The well also contained large quantities of domestic refuse and may have been filled at or about the same time as the large Trash Pits A and B north of the major structures on the site. Unfortunately, no ceramic crossmends were found to prove a contemporaneous relationship between Trash Pit B and the well; however, a crossmend in a locally-made coarse earthenware vessel (Catalog Number 1143) was found to prove the temporal association of Trash Pit A with the well.

Trash Pits G and H each contained six marked bowls. Four in each pit were of Dutch origin, while the remainder were of unknown manufacture. Trash Pit F contained only four marked pipe bowls, all suspected of being Dutch in origin. The association of Trash Pits F and G is suggested but not proven by the presence of pipe bowls with the same maker's mark, "S H," in each and also by a non-contiguous ceramic fragment association in an English delftware jug (Catalog Number 1017). Dutch tobacco pipes with this mark have been found on other sites with depositional 151 Figure 49. Imported marked pipes—probable origin. Figure 50. Marked pipe concentrations. dates prior to 1650 (A. Noël Hume 1979:2526; Oswald 1969:140).

Capacity Study of Tobacco Pipe Bowls

The capacity of some of the imported and domestic tobacco pipe bowls from the HT55 was recorded using aluminum oxide crystals of the type used in airbrasive conservation equipment. This substance was chosen because the microscopic crystals are fairly regular in size, measuring no more than 50 microns in diameter, and because this powder has a low compression factor, thus insuring greater consistency in the measurements of weight. The stem holes at the bases of the bowls were plugged to prevent any of the powder from entering the stem and compromising the measurements. The capacities of the sample pipe bowls are calculated based on the assumption that a rigid container with the internal measurements of 2.32 cm by 2.32 cm by 0.48 cm deep contains 2.58 cubic centimeters. This container was completely filled with the same aluminum oxide powder that was used to measure the pipe bowls. The powder contents of this container weighted 3.7 grams. The calculations for cubic displacement of the bowls were then made using the following basic ratio formula:

A random sample of 20 white-clay pipe bowls was selected to be compared with a similar group of domestic pipe bowls in the hopes of finding some correlation between the sizes. It was impossible to perform this same comparison between pipes of English and Dutch origin because of the measurable and marked HT55 pipe bowls, only three were identified as possibly English in origin, as opposed to 18 Dutch bowls. The obvious imbalance in the quantities of examples would have biased the averages and made any other calculations meaningless.

The average bowl capacity of the white-clay sample was 2.64 cm3, while the average capacity of the domestic clay bowls was 4.30 cm3. This amounts to approximately 61.4% greater capacity for the domestic pipes over that of imported examples (Figure 51). This variance can possibly be explained in part by the methods of construction of the two types of pipes. The imported pipes were invariably molded, while a significant proportion of the domestic pipes were hand-formed 152

Contemporaneous Relationships Established Through Crossmending

The white clay tobacco pipes from HT55 were crossmended, yielding two contemporaneous relationships between macro-features. A small trash pit (Trash Pit E) and the well were proven contemporaneous by a single crossmend in a pipe bowl (Catalog Number 6102). A crossmend between two fragments of a plain tobacco pipe stem (Catalog Number 6106) proves a relationship between Trash Pits G and H.

Other Artifacts

The author regrets the lack of complete photographic illustration of the objects under discussion in this section of the report. The conservation treatment of the vast majority of the metal and organic artifacts had not been completed at the time of this writing. It was therefore impossible to provide the same level and quality of photographic documentation for these objects and at the same time meet the designated report completion date. Pre-treatment conservation photographs and drawings of the objects submitted for conservation treatment were made and are available for inspection and comparative study.

The non-ceramic and non-glass artifacts recovered from HT55 provided some unique insights into the quality of life enjoyed by early Virginia settlers (Table 27). The diversity and range of non-ceramic artifacts found on the site may be useful in suggesting activities areas and functions of the various major structures and related features. In light of the suspected, but not proven, ecclesiastic association with the Parish of Kicoughtan, this functional/locational analysis may prove useful in understanding the land-use practices of the inhabitants of the site. For ease in systematically addressing this body of material the finds were divided into ten functional categories, including: Non-Ceramic Food Preparation and Consumption, Cutlery, Tools and Equipment, Weaponry and Armor, Personal Items, Textile Related, Trade Related, Furniture Related, Architectural, and Nails. Even though cutlery items are technically associated with both food consumption and preparation, it was decided to separate these tableware items from the other food-related objects because of the large quantity of knives. It is felt that a distribution of these cutlery items in comparison to the other foodways objects may be revealing.

Operating on the long-established premise that concentrations of artifacts on a site may indicate areas of specialized activity, the finds were analyzed to suggest possible craft activities, trash disposal practices, and possible functions of the major structures (Beck et al. 1983; Stewart-Abernathy 1986; King 1988).

It is evident from the distribution, that the majority of food preparation and consumption objects were retrieved from the well, which is located on the west side of the site. These objects appear to be metal drinking cups and cooking pot fragments.

A similar pattern is evident in the distribution of cutlery objects as in the food preparation and consumption artifacts (Photo 40). The well to the west of the site held the majority of knife blade fragments, followed by Trash Pit D, also located on the west side of the site. The highest concentration of bone cutlery handles occurred in Trash Pit B just north of Structure A, on the east side of the site. There was also an even distribution of bone handles in all of the major features with the exception of Trash Pits A, C, and H. Pewter spoon handle fragments were found in the well to the west and in Trash Pit B to the east, but in none of the other major features. The concentrations of food-related artifacts on the eastern and western 156