The third site excavated in 1972 was Rush Point, named by Haag (1958: 40) for the local reference to the northwest point of Big Colington Island, rather than the official map name of "Eagleton Point". At the point is a U. S. Coast and Geodetic Survey bench mark, a square concrete post with the typical brass plaque (dated 1915) embedded in the top, now over-turned from wave action around its present position 2 meters from the shore (Figure 3). This marker is only one of the indicators of shoreline loss at 31DR15 and along this section of the island. (Page 37)

The Rush Point site extends from just south of the marker for a distance of 225 meters southward along the northwestern shore of the island (Figure 13a). Near the point, oyster shell from the midden is interbedded with peat which apparently developed over the midden but is now submerged and being eroded by the rise in sea level. The midden has been completely destroyed by erosion in the small embayment south of the point, but appears as lens-shaped deposits of shell on top of an old humus layer as it rises along the dune to the south (Figure 13b). At the north end, elevation of the midden is approximately 30 centimeters (1 foot) below sea level; on the south it attains a maximum height of 17 meters (56 feet) above sea level. Between these two points the old humus zone and its occasional overburden of midden deposits undulates to conform to the dune surface of that time. At two spots along the bank, an even older humus level was observed, but that stratum was devoid of cultural material in the exposed sections.

Along the beach and for some distance out into the sound is an extensive secondary deposit of oyster shells and sherds, all derived from erosion of the site. The shoreward slope of the dune from which the cultural material is eroding has an angle of descent ranging from 45 to 60 degrees, and is practically devoid of vegetation (Figure 13b, 14). On the inland (eastern) side of the dune, two areas of shell were observed on the surface where modification of the dune slope by construction had occurred. Figure 14a, a photograph taken in 1977, shows the cottage constructed over the area excavated in 1972, and steps leading to the beach. By 1979, the date of the aerial photo in Figure 13a, a bulkhead had been constructed at the foot of the steps to retard continuing erosion.

Most cultural material collected from the surface of the site was found along the beach, but a few specimens were retrieved from the exposed shell areas on the east side of the dune ridge. Elsewhere the site was buried beneath the modern dune. The 1972 collection included 288 ceramic sherds, 275 of the Mount Pleasant series and 13 of the Colington series, and a flake blade (Table 4). Haag's 1955 surface collection lists 326 sherds (Haag 1958: Table 1); 186 of these fall within the present classification of Mount Pleasant (Haag's sand tempered, grit tempered and clay/grit tempered), and 140 shell tempered specimens fit the Colington series typology. Also collected from the beach in 1955 were a flat slab mortar and a steatite sherd. The much higher number of shell tempered sherds in the earlier collection may result from combining beach collections from 31DR33 and 31DR15 and subsuming them under the latter site; an alternative possibility is that construction and erosion overwash toward the northern end of the site may have obliterated a small Colington component. Whatever the answer may be, the 13 Colington sherds from the 1972 beach collection are the only late material found. The intact portions of the site date to an earlier period.

A human patella, the only evidence of human bone from any of the sites on the island, may indicate occasional human burial at 31DR15. (Page 38)

(a) (b) Figure 13. Views of the Rush Point and Stick sites. Aerial view (southeast) of the Rush Point site, extending from the point in the left to the dune crest in the center of the photo, and the Stick site, to the right of the photo adjacent to the harbor outlet (a). Erosion exposed roots and peat at Rush Point and the intact section of 31DR15 in the background. Crew at left laying out Square C (b). (Page 39)

Figure 14. Rush Point site in 1977. A winter view of the site taken from Rush Point; cottage and steps built after 1972 work. Note erosion of beach to steps and dark line of the 3rd century humus angling up the dune face to the right.

Following the surface survey of Rush Point, temporary reference stakes were set along the beach on the intact sections of the site, and their locations mapped in reference to the U.S. Coast and Geodetic Survey marker at the point. From the reference stakes, site topography, was mapped, the dune profile over the beach was plotted. and excavation units were laid out. Exposure of a small section of shell midden in the beach profile and similar evidence of midden on the eastern side of the dune ridge dictated the placement of squares A and B (Figure 3) on the northern slope of the ridge. These 2 x 2 meter units were augmented by a third square, C (1 x 2 meters), located at a point in the embayment where it appeared that the old humus layer dipped below beach level (Figure 13b). Upon excavation, however, this area gave evidence of construction disturbance, wave erosion and over-wash, and no intact deposits were observed.

Squares A and B and the beach profile confirmed a stratification similar to the sequences in 31DR14 and 31DR33, with exactly the same five zones. Square A is typical of the excavated units: Zone I, (Page 40) the modern humus, extended to an average depth of 10 centimeters; Zone II, the modern dune sand (Fripp fine sand) was absent, however, and the Zone III shell and organic midden directly underlay the modern humus. The midden thickness ranged between 25 and 35 centimeters and was underlain by the older humus, 10 to 20 centimeters in thickness, of Zone IV. Below this the sterile yellow sand of Zone V continued downward to an unknown depth. In the beach profile Zone II was observed to have a thickness ranging from a few centimeters to 5 or more meters, depending upon modern dune development on the ridge. The possibility of remnants of a third, earlier humus zone within Zone V was observed along the profile, but its depth precluded investigation during the project and the exposed section was culturally sterile.

As in the other sites, Zone III was the only cultural deposit in the sequence. This zone was comprehensive in Square A, but in Square B only the edge of a shell deposit was exposed in the northeast quadrant. A sample of oyster shell from the lowest level of Zone III in Square A was radiocarbon dated to A.D. 265 (UGa-1088; B.P 1685 + 65). A second sample of oyster shell from Level 2, Zone III in Square (?) was dated at A.D. 890 + 80 (UGa-3849; 1060 + 80 B.P.). This date seems late for the depositional sequence in the square, and it is possible that its proximity to the surface may have resulted in contamination. Seriation of materials from arbitrary 10-centimeter levels excavated within Zone III indicated no temporal or typological differences within the stratum.

One feature (Feature 1) was observed in Square A. This was a shallow basin-shaped pit, measuring 70 centimeters in diameter and 17 centimeters in depth, which intruded from the bottom of Zone III into the upper part of Zone V. No artifacts were found in the mixed fill of sand, charcoal, ash, calcined shell and a few fish bones. The feature is presumed to have been a refilled hearth or cooking pit.

Of the 378 artifacts recovered in 1972, 377 are ceramic sherds of two series; the remaining artifact is the only lithic specimen (Table 4).

Mount Pleasant series ceramics account for 364 of the 377 total sherds from the site, and was the only ware found in primary context in the excavations. The types represented (Figure 15a-f) are cord marked (25), net impressed (53), fabric impressed (122), plain (11), and incised (6). Incised surface decoration has not previously been included in this series, but the specimens from (Page 41) 31DRIS have technological attributes comparable with series definition. The five incised sherds from Square A (Table 4) appear to belong to the same vessel which had a rectilinear incised design on its smoothed surface (Figure 15f). It is possible that the incised type is trade material, but future work will have to resolve the problem. Variation in the tempering material was noted at 31DR15; the majority of the Mount Pleasant sherds have liberal amounts of sand with frequent small pebble inclusions, but many of these also have clay lumps mixed into the paste. A few sherds have only clay temper, again reminiscent of the Hanover type, such as the fabric impressed specimen in Figure 15g. Residual Mourt Pleasant sherds, 147 in number, resulted from a high percentage of eroded and water-polished specimens (130) from the beach collection.

Colington series ceramics at this site were found only on the beach. The presence of the 13 specimen (Table 4) is thought to result from sporadic visits to the site from 31DR33 or other nearby Colington phase sites, rather than represent a major occupation during the late Woodland period.

One tan quartzite flake with an edge retouched for use as a cutting tool was collected from the beach. Haag (1958: 87; 95) reported a steatite sherd and a mortar from his collection along the beach.

The most obvious food remains are the masses of oyster shell which make up the mass of the midden lenses. The site location indicates availability of oysters in Albemarle Sound, in beds fairly close to the western shore of Big Colington Island. Mixed with the oyster shell in the midden were the shells of a few Busycon caricum (knobbed pear conch), whelks and moon shells. (Page 42)

Figure 15. Ceramics from 31DR15. Mount Pleasant net impressed (a), fabric impressed (b-c), cord marked (d) and plain (e). The three incised sherds (f) are from a single vessel and have typical Mount Pleasant series temper. A clay tempered fabric impressed sherd is shown in g. (Specimens are half actual size). (Page 43)

Quite surprising, was the recovery of appmximately 50 fresh water mussel shells in a concentration at the bottom of the shell midden in Square B. These shell fish cannot inhabit the saline environments which support oysters and the other species found in the midden, and their presence in the site suggests that they were brought in from the upper reaches of the Albemarle Sound or its tributaries where the species is still numerous in fresh water habitats.

Except for the reliance of oysters and presence of similar land faunal remains, the subsistence pattern at 31DR15 is strikingly different from that at the White Court and Kitty Hawk Bay sites. The only fish remains in the site were those of a few small sharks or rays (Appendix B), indicating that fishing was a very minor activity. It is possible that the sharks or rays were killed during the process of collecting oysters, rather than being intentionally sought. One deer and one raccoon were the only representatives of the mammalian class; an unidentified species of duck was the only bird; and diamond-back terrapin was the only reptilian food remain (Table 10). Approximately 155 bones (vertebrae and ribs) of a water snake were found concentrated in the lowest level of Zone III in Square A, and probably represent a "fright" kill rather than food.

There is little suggestion of seasonality from the faunal remains at 31DR15 (Appendix A), and oyster collection patterns indicate possible occupation anytime between late spring and autumn.

Although the 13 shell tempered sherds from the beach collection at Rush Point suggest occasional visits or brief use of the site during the Colington phase, the major occupation of the site was during the Middle Woodland Mount Pleasant phase (300 B.C. - 800 A.D.). 99.97 percent of the ceramic sample from the site and all ceramics from the midden excavations belong to the Mount Pleasant series. Radiocarbon dates from the lowest level of the midden in Square A indicates a beginning date of the Rush Point occupation around A.D. 265 and an ending date around A.D. 690. This temporal span appears to be excessive for the presumed depositional history, and the latter sample may have contained modern carbon from surface intrusions.

At the time of occupations the areas of shell midden were deposited over the humic surface then exposed and probably stabilized with a forest cover similar to the present mixed oak and pine community. The northwest point of the site, now 30 centimeters below sea level, was at that time 7 to 8 feet (2.1 to 2.4 meters) above water, if the minimum rate of rise in sea level (Page 44) (0.5 feet/century) (Bellis, et al, 1975: 13-14) is multiplied by the 17 centuries since initial occupation of the site. From the points the land surface rose gradually to the south, undulating with the uneven deposits of wind-blown dune sand below. Much of the western side of the site has been lost to erosion, probably most intense in the summer from prevailing southwesterly winds and storms. The rate of shoreline erosion in similar situations is 2 to 5 feet per year (Bellis, et al, 1975; Table 2) but this formula cannot accurately be applied over many centuries because of its direct relationship to current sea level. Other variables which affect the formula such as sea level rise, sedimentation, dune building, and other processes, are not sufficiently well understood in each particular situation. If the erosion formula is applied to the Rush Point site, using the minimum erosion rate of 2 feet per year, the site would have lost approximately 3400 feet since 265 A.D.; this would have placed the original site over one-half mile inland, an illogical location for a settlement established to exploit oyster beds in the sound.

The intrasite pattern at Rush Point is similar to the other sites on Big Colington; lens-shaped deposits of shells, bone, charcoal and other debris, each containing a few artifacts, are randomly scattered over the old land surface. These represent refuse piles remaining after food-gathering, preparation and consumption; few other cultural activities are present in the evidence. The picture is, again, one of temporary subsistence camps, set up by people from a mainland or interior area, located to exploit nearby oyster beds. However, the absence of fish (other than a few sharks and rays), and reptilian fauna from the fresh water ponds and uplands suggests groups of people not yet totally familiar with the estuarine environment, and exploiting only those resources of which they had some knowledge--shellfish and land fauna. A possible indicator of the group's insecurity is the concentration of mussel shells in Square A, which suggests a food supply brought with the inhabitants, perhaps from the upper reaches of Albemarle Sound, on their trid to the site. The resources exploited at Rush Point were available in close proximity to the site, and a catchment area of one-half mile is suggested for the site.

The artifact assemblage is almost totally ceramics, probably containers for cooking and storage, also brought with the people from their home base. The milling stone from the 1955 collection and the blade from the later work are minimal equipment. No evidence of lithic production was recovered, and no evidence that the people of the Mount Pleasant phase who occupied this site were aware of the jasper sources on the outer islands. This, like the mussel shells and absence of fishing, suggests a group from the interior of the estuarine zone rather than people totally familiar with the immediate coastal area. (Page 45)

The Stick site (Figure 3) was described by Haag (1958: 41) as occupying the dune ridge north of the outlet from a "nearby landlocked embayment", which is now called Colington Harbour. Haag observed the older humus layer on the eroded side of the dune ridge, and recorded pockets of oyster shell in the dips of the layer. He sifted "several hundred cubic feet" of the humus without recovering any artifacts, but did collect "sherds and flints" from the beach.

Since 1955, the south end of the dune ridge on which both 31DR15 and 31DR16 are located has been removed to accommodate the club house, swimming pool, tennis courts, parking lots, dock and picnic area which axe the recreation nucleus of the Colington Harbour development (Figure 13a). In 1975, only a few sherds were observed and collected from bare spots in the picnic area, but the beach was literally covered with oyster shells, among which sherds were fairly common.

This concentration along the beach is separate from the debris adjacent to 31DR15, and begins approximately 125 meters south of the latter. From there the distribution of shell and sherds is continuous for 125 meters, nearly to the mouth of the outlet. The eastern extent of the site in Figure 3 is estimated since disturbance of the area by construction has removed all primary context. The original site occupied the southern end of the dune ridge (Fripp fine sand) and was, like 31DR15, probably deposited in separate shell piles on the older humus which is buried to varying depths beneath the modern dune.

Haag's (1958: Table 1) collection from the Stick site was 40 sherds of sand, grit and clay tempered sherds; there is no further reference in his report on the "flints" he collected from the beach, and it is presumed that these were spalls. Haag's ceramic collection fits typologically into the definition of the Mount Pleasant series, and is classified by type in Table I along with the 1975 collection of 92 sherds. The 1975 collection included Mount Pleasant cord marked (10), net impressed (11), fabric impressed (25), plain (2) and residual (44), the higher nwnber of the latter resulting from water polishing and erosion in the surf. Most of the specimens are tempered with medium to coarse sand with frequent small pebbles, but a few also have fired clay lumps along with the clastic tempering materials. In all respects, the ceramic sample from the Stick site is a duplicate of that from 31DR15.

The Stick site may, in fact, have been simply an extension of the occupation area represented by intact deposits farther north an the ridge of 31DRIS. Even if they were separate, the two sites appear to have been roughly contemporaneous Middle Woodland period occupations. The artifacts and shells observed at the site, and the pre-construction description from the earlier work, both indicate (Page 46) a temporary subsistence camp oriented toward Albemarle Sound with an intrasits pattern basically the same as that at Rush Point.