Although the Cape Island investigations are considered "limited" testing and data recovery, data sufficient to address several issues presently under inquiry in coastal North Carolina archaeology were recovered. The research issues for this investigation are presented below.
One of the distinguishing traits of Algonquian occupation along the North Carolina coast is the longhouse (Loftfield and Jones 1995). Longhouses can be seen in the John White drawings and De Bry engravings. The archaeological evidence for structures on the coast is limited by the sandy nature of the soils, the high rainfall, and a long history of plowing, all of which have reduced the preservation of soil stains and have disturbed features indicative of buildings.
Several structures have been uncovered on the coast, however. These have, for the most part, consisted of two design types. The first is a small rectangular structure, which appears in the White and De Bry renderings of Secotan and Pomeiooc. Archaeologically, this structure form is seen at the Amity site (31HY43) in Hyde County (Gardner 1990), Permuda Island (31ON196) in Onslow County (Loftfield 1985), and the Broad Reach site (31CR218) in Carteret County (Mathis 1993). They are 6 x 9 m at Amity, 4 x 8 m at Permuda Island, and 4 x 6 m at Broad Reach. The small rectangular structures are composed of individually set posts.
The second design form is the Algonquian longhouse. This style is also seen in the drawings of Secotan and Pomeiooc. The longhouse has been observed at the Amity site (31HY43) in Hyde County (Gardner 1990), the Broad Reach site (31CR218) in Carteret County (Mathis 1993), and the Uniflite site (31ON33) in Onslow County (Loftfield 1979). These structures measured 6.5 x 14 m at Amity, 6 x 15+ m at Broad Reach, and 5 x 13 m at Uniflite.
Mathis (1993) reports at least one round to oval structure at the Broad Reach site. Mathis (personal communication 1997) suspects it is a Middle Woodland structure. Pits immediately adjacent to the structure have Middle Woodland dates, and Hanover is the prevalent ceramic in that part of the site. Loftfield (personal communication 1997) had some possible circular patterns at Permuda Island, but additional work discerned no definite structures.
The data available from the central North Carolina coast regarding structures leads to several questions: 1) Are the small rectangular and true longhouses the only design associated with the Algonquians? 2) Are the round to oval structures observed at Broad Reach and possibly Permuda Island associated with the Algonquians, or with earlier groups? 3) Were longhouses used by other Late Woodland groups on the coast, or are they exclusively an Algonquian characteristic? 4) Are there enough structure patterns remaining on the coast to address community plan, and will such a plan be consistent with those seen in the White drawings and De Bry engravings?
Another hallmark of the Carolina Algonquians was the ossuary burials (Phelps 1983). Previous research along the coast has indicated three ossuary burial patterns. Ossuaries associated with the Iroquoian speakers on the inner Coastal Plain in the north generally contain two to five individuals interred as secondary bundle burials (Phelps 1983), assumed to represent family units. The Iroquoian ossuaries usually contain marginella beads as grave goods and are inside villages.
In the north and central parts of the outer coastal plain, the Algonquian ossuaries are associated with shell-tempered ceramics and generally represent community burials. They contain many individuals, a significant proportion of whom are represented by isolated skeletal elements. McCall (1987) noted Felis phalanges in two Algonquian ossuaries in the north region. Generally, however, no unequivocal grave goods are present. Algonquian ossuaries have generally been assumed to be a short distance from the villages.
In the southern part of the coast, ossuaries associated Siouan peoples are placed in mounds or on high sand ridges. This practice began in the Middle Woodland and apparently continued into the Late Woodland. These mounds harbor secondary burials placed in bundles. Cremations, or least charred skeletal elements, are often present, and grave goods are sometimes present. These burials appear to be far from village sites, but this contention needs to be tested with intensive survey designed to locate village sites.
As laid out above, the Broad Reach excavations have raised some questions about Late Woodland burial patterns along the North Carolina coast. Is the broad array of burial types at Broad Reach representative of cultural exchange between Siouan and Algonquian groups, as suggested by Loftfield (1990)? Or do they represent alternating occupation of the central coast by both Siouan and Algonquian groups?
Hargrove (1993) recovered limestone-tempered ceramics from the Hamp's Landing site. His report represents the first recognition of this ceramic type. Hargrove was unable to obtain a radiocarbon date for this ware but suggested, based on the stratigraphic position of the sherds, that it was a Middle Woodland occurrence.
Can Hamp's Landing ceramics be recovered from a datable context to either confirm or refute Hargrove's Middle Woodland attribution? How does the Hamp's Landing series compare technologically and stylistically with other area series? Herbert and Mathis (1996) suspect that the previously defined Oak Island series (South 1976) may actually be the limestone-tempered Hamp's Landing ware. This possibility plays into the next research issue.
The distribution of Late Woodland peoples along the North Carolina coast has generally been interpreted from linguistic studies and ethnohistoric accounts (e.g., Hariot 1991; Speck 1924; Mook 1944; Paschal 1953; Lawson 1967). North of the central North Carolina coast, populations spoke dialects of the Algonquian language family. South of the central coast, the inhabitants used a Siouan language. In the central coast, there are no surviving early observations from which a language affiliation can be inferred. This raises the question of exactly which group occupied the central coast at the time of European contact. To further confuse the issue, Lawson (1967) indicates that the Coree, living along Core Sound in central North Carolina in the early eighteenth century, were speaking an Iroquoian language. However, one must keep in mind that much happened between the time of Hariot's visit in 1585 and Lawson's visit in 1700. For example, disease brought in by Europeans had resulted in great movements of people and much cultural disruption.
There are discrepancies between the ethnohistoric accounts and the archaeological record. For example, the early European visitors leave the impression that agriculture, large mammals, and large fish provided the bulk of the diet for the prehistoric inhabitants (Hariot 1991; Quinn 1985). But very few large fish or mammal remains are recovered from midden or feature contexts. Instead, the bulk of the diet appears to come from small fingerling fish and shellfish. It is suggested that the focus of those early European visitors was on public relations instead of ethnographic accuracy.
The extent of Algonquian expansion to the south lies somewhere along the central North Carolina coast. Exactly how far did the Algonquians expand? How long did they occupy the southernmost extent? Following European contact (or prior to that) did the occupation of the area shift intermittently from Algonquian to Siouan to Iroquoian?
Data from the physical remains of individuals interred in ossuaries can lend some evidence to address the question. The Algonquians are generally considered to be a large-statured population.
Did Hamp's Landing ceramics appear during the Middle Woodland and continue through the Late Woodland? That is, if the previously defined Oak Island ware (South 1976) is indeed Hamp's Landing, does that suggest a Siouan trait which continued along the southern coast right up to contact? Is the presence of Hamp's Landing ceramics in the central part of the coast indicative of an established Siouan presence there or just evidence of material exchange?
Loftfield (1987a) has suggested that by the Middle Woodland period, a pattern of adaptation to the coastal area was evolving based upon exploitation of estuarine resources, with sites becoming less frequently located in inland areas. Does the round structure uncovered at Broad Reach represent a Middle Woodland permanent occupation? If so, does it indicate a complete and total reliance on the estuarine environment earlier than presently suspected? That is, are we seeing an abandonment of seasonal rounds in preference for year-round occupations in the littoral zone before the Late Woodland?
Can changes in the frequency of wild vs. domesticated plants be detected in Middle Woodland components? Can such data, combined with structural data, argue for an earlier adaptation to the estuarine environment than presently known? That is, does evidence of sedentary life exist for the Middle Woodland?
Site 31ON190 was tested by using a 120G road grader to remove all but approximately 23 inches of the plow zone from areas along the proposed road running through the center of island (Sailview Drive) and the proposed road serving as the entrance to the development (Cape Lane). To complete the stripping, a tractor with a 7-foot box blade was used to remove the remaining plow zone. Six areas measuring 65 x 10 m were stripped along the proposed Sailview Drive. One area measuring approximately 65 x 10 m was stripped along the proposed Cape Lane. These were designated Areas AG.
Additionally, six areas measuring approximately 15 x 3 m were stripped perpendicular to the proposed Sailview Drive. These smaller areas were aligned along proposed lot lines. They were designated Trenches HM. All of these smaller trenches were backfilled at the end of fieldwork.
The removal of the plow zone was monitored by at least one person at all times. This person(s) also shovel-shaved the stripped surface in order to better recognize features and possible post holes. Any artifacts observed during the plow zone removal were collected, bagged, and labeled with the appropriate Area or Trench designation. All features and possible post holes were marked with pin flags. Features were then numbered consecutively. Possible post holes were not numbered unless they appeared to represent a pattern indicative of a structure, but all post holes were mapped. Numerous post holes or possible post holes were present in all areas and trenches. Not all of the possible posts could be verified as such. A conservative estimate of the true post holes uncovered during the fieldwork is approximately 1,500.
Features were photographed in both black and white print and color slide format and drawn to scale in plan view. Each feature was then bisected, and one-half was removed and screened through 0.125 inch mesh hardware cloth. The resulting profile was then photographed in both black and white print and color slide format and drawn to scale. The other half of the feature was taken as a soil sample to be processed by flotation.
All stripped areas, trenches, and features were plotted with a Leica T1010 total station. The field map was updated and field-checked daily.
All artifacts were returned to Garrow & Associates' archaeological laboratory in Atlanta, Georgia, for analysis and temporary curation. The first step in analysis involved inventory control measures such as updating bag lists, correlating level and feature forms with the bag lists, and cross-checking artifact and feature provenience. After these had been reviewed and verified, accession numbers were assigned and listed in a master accession log.
All flotation soil samples were processed using a Dausman self-contained flotation system. Ethnobotanical and zooarchaeological remains were sorted from the flotation residues. Methods of the ethnobotanical and zooarchaeological analyses are discussed separately below.
Artifact analysis forms detailing the provenience, bag number, quantity, artifact raw material, and, in some cases, the total weight of each artifact type, as well as artifact descriptions and classifications, were completed during initial sorting and typing of artifacts. All artifact data were entered into a specially designed coded database using 4th Dimension and Microsoft Excel software on Apple Macintosh computers.
The prehistoric artifact analysis focused on identifying assemblages and technological attributes diagnostic of particular temporal and cultural trends along the central North Carolina coast. The artifacts were identified according to established regional types or styles. Ceramics were typed according to paste, temper, and surface decoration. Haag (1958), Loftfield (1975), South (1976), Phelps (1983), and Herbert and Mathis (1996) were followed for ceramic identification.
The term "sample vessel" is used in this analysis. A sample vessel represents portions of an individual pot that can be separated from other vessels or sherds based on decorative techniques and modes, thickness, vessel form, and paste characteristics. Since the cumulative assemblage of sample vessels cannot be a true minimum vessel assemblage, the term "minimum vessel" is not applied. The adjective "sample" is appropriate because the sample vessels represent only a portion of all vessels at the site, because the utilized sherd collection does not represent 100 percent of the site area and because a site assemblage will represent only a sample (i.e., those that broke) of the vessels used at the site.
Sample vessel analysis can yield significant data on site function and use intensity, stylistic change through time, and diachronic changes in technology. However, the limited scope of the present data recovery resulted in a small number of sample vessels. The pottery from the dated Features 4, 15, and 35 was examined. In addition, sherds from Features 11 and 37 were analyzed. It was opportune that a direct cross-mend was recognized between sherds from Features 35 and 37, thereby providing indirect dating for Feature 37.
All sherds were pulled during the primary analysis. The sherds were labeled and then placed according to surface treatment. Cross-mending was attempted based on surface treatment, decorative mode within a surface treatment (e.g., width of cords), thickness, vessel form, and paste characteristics. When identified, cross-mends were glued together. Following the cross-mending, sherds were sorted into sample vessels based on the same attributes as the cross-mends.
At the end of the sorting for sample vessels, not all of the sherds were assigned to a given vessel. Many sherds lacked sufficient decorative or stylistic discriminators to allow them to either be placed with a given sample vessel or be used to define a sample vessel of their own. However, the sample vessels represent over 90 percent (by weight) of all feature ceramics recovered.
The contributing sherds for each sample vessel were recorded, as was the estimated percentage of the vessel represented. The latter attribute is important in addressing issues of refuse patterning. Each sample vessel had the following attributes recorded to provide a detailed technological description of the wares:
Type, size, shape, and density of major aplastic
Type and size of minority aplastics
Degree of carbon core retention
Sherd core cross-section configuration
Dominant paste color
Thickness 3 cm below rim
Rim production step
Presence of coil breaks, sooting, and interior abrasions
Exterior and interior surface treatments
Inferred vessel form
Decorative mode details
These attributes have been examined in numerous ceramic analyses in the Southeast (Blanton et al. 1986; Espenshade and Brockington 1989; Espenshade, Foss et al. 1994; Espenshade, Kennedy et al. 1994; Poplin et al. 1993). This level of analysis has been argued to be the minimal acceptable level for data recovery of pottery-bearing sites (Espenshade 1996:44).
The type, size, shape, and density of major aplastics will reflect the nature of the clay utilized and the type of temper added (if any). Major aplastic was defined as the prevalent (by count) aplastic that was fine or larger on the Wentworth scale. Aplastic type was described as geological stone type (e.g., quartz, feldspar) or type of organic temper (e.g., fiber).
Aplastic size was defined by comparison of the sample vessel to a clay bar with known aplastic content. Aplastic size was recorded on the Wentworth scale as finemedium, coarse, very coarse, and granule sizes. Sherds with no aplastics finemedium or larger were labeled as No Apparent Temper.
Aplastic shape was described by geological grain shape labels including angular, subangular, subrounded, rounded, platy, and irregular. Shape was examined on a fresh break under 30X magnification.
Aplastic density was recorded as the number of major aplastics in an 8 mm diameter view field. Density was examined in several locations on a fresh break, and an average was estimated. In addition, an estimate was made of the percentage of the body cross section comprised of aplastics (rather than paste).
The type and size of minority aplastics were also recorded. Minority aplastics may indicate natural aplastics specific to a given clay source type. Alternatively, minority aplastics may reflect the natural range, across Wentworth size class lines, of natural or added aplastics. Typing and sizing followed the methods described for majority aplastics.
The degree of carbon core retention is related to the nature of the clay; the temperature, oxygen environment, and duration of the firing; and the postproduction use of the vessel. In general, the incomplete firing of vessels produced of native clays will result in the retention of carbon-rich clay in part of the wall cross section (Rice 1987). Shepard (1980) considers dark grays and blacks typical of carbon-rich clays. Carbon core retention was recorded as the percentage of the vessel wall cross section comprised of dark gray or black colors.
Two other attributes, sherd core cross-section configuration and dominant paste color, are related to degree of carbon core retention. The type of core configuration (homogeneous, two-layer, three-layer, etc.) and the paste colors will reflect the general nature of the prehistoric firing. The presence and location of a dark core relative to other paste colors can reflect the orientation of the vessel during firing. Sherd core cross-section configuration was recorded as the color layers present from interior to exterior.
As stated above, dark grays and blacks are indicative of carbon-rich clays fired at too low a temperature, for too short a time, or with insufficient air flow to allow carbon to be completely burned off. Tans, buffs, and reds are indicative of oxygen-rich firings, while light grays and browns are indicative of reduced oxygen firings (Shepard 1980).
Thickness was measured 3 cm below the rim. When measured consistently, thickness can reflect technological traditions and functional considerations (Rice 1987:227228). Three centimeters below the rim is far enough down the vessel wall to avoid rim-specific modifications (e.g., folded or thickened rims). This measurement point has been used by numerous studies in the Southeast (e.g., Sassaman 1993a, 1993b; Blanton et al. 1986; Espenshade, Foss et al. 1994; Espenshade, Kennedy et al. 1994; Espenshade and Brockington 1989; Poplin et al. 1993), assuring comparable data.
Rims were addressed through two related attributes, rim form and rim production step. Rim form is simply the basic description of the rim shape (e.g., round, square, round with exterior lip). In this analysis, rim refers only to the extreme edge around the vessel orifice; it does not include the vessel neck or shoulder attributes. Other analysts refer to rim as lip.
Rim production step involves the process than went into making a rim form. There are many different ways of producing a given rim form. A square rim, for example, can be produced through smoothing, scraping, cutting, and paddling. By examining both rim form and production step, it is possible to distinguish technological traditions within the seemingly limited range of rims in southeastern prehistory.
Coil breaks, sooting, and interior abrasions were recorded as present, possibly present, or absent. Coil breaks reflect breakage along weaknesses in the vessel wall that follow coil junctures (Rice 1987). Not all coiled vessels exhibit coil breaks, and the frequency of coil breaks is relevant to addressing the relative use strength of vessels. Coil breaks are recognizable as latitudinal breaks with either concave or convex, regular cross sections.
Sooting is generally the direct result of pot use over an open fire (Skibo 1992). Sooting can occur in postdepositional context (e.g., a sherd used as a cooking pot support during a later occupation), so this attribute has its best interpretive value when large vessel sections are present. Soot is a surface deposit of dense carbonaceous material.
Interior abrasions represent direct evidence of mechanical use impact to a vessel (Hally 1983; Skibo 1992). Interior abrasions have been linked to vessels commonly used for mixing, food preparation, parching, and boiling and stirring. Interior abrasions are not expected on vessels used for bulk storage of liquids or dry foodstuffs. Interior abrasions are recognized as patterned damage to the interior surface of a vessel.
Exterior and interior surface treatments were also recorded. Exterior surface treatments are clearly of interest in typology and chronology. The examination of interior surface treatments can be useful in addressing potential technological continuity between types or series. Interior surface treatment was recorded as poorly smoothed, smoothed, semiburnished, burnished, and scraped.
As feasible, the form of the vessel shoulder and/or base was recorded. This attribute was directly linked to the inferred vessel form. Because vessel morphology has been shown to reflect the primary intended functions of the vessel, the recognition of vessel forms can provide important data for site interpretations (Braun 1980; Hendrickson and McDonald 1983; Smith 1985). The only two vessel forms recognized were shallow bowls and deep bowls. Vessel profiles were drawn as feasible.
Rim diameter was measured on a 1-cm-increment chart when a sufficient rim section was available to assure reliability of the derived measurement. Studies have indicated that vessel size is often linked to site function and size of the social units (Shapiro 1983, 1984; Smith 1985). Rim diameter data complement vessel form information.
Because lithic resources are sparse on the North Carolina coast, very few stone tools are recovered from coastal sites. The majority of tools used by prehistoric inhabitants were probably fashioned from perishable materials such as wood or bone. Tools made from whelk shells are often found on coastal sites consisting of both complete and fragmented specimens. Whelks from the Cape Island site were examined as whole unmodified shells, modified shells, and fragments of columnella.
Tool types were defined based on the location and extent of modifications observed on particular shells. Probable function of such tools was inferred based on the observations.
Charred plant remains from the general site fill were placed in acid-free bags and labeled with the appropriate provenience information. Soil samples taken from features were individually processed through the flotation system described above, which carries away only those soil particles which are less than 0.0625 inch in size. The light fraction floats to the top and is siphoned off. The heavy fraction is retained. Both fractions were allowed to air-dry and were then examined for charred botanical remains. All charred botanical remains from the flotation system were placed in acid-free bags and appropriately labeled.
The plant remains were identified using the botanical collection held at Garrow & Associates' laboratory. Each specimen was identified to the species level when possible. Of the 795 charred botanical remains recovered, 270 (33.9 percent) were highly fragmented seeds that could not be identified. In 10.0 percent of the cases (80 seeds), severe charring had obliterated characteristic physical traits, such as surface textures and attachment areas, precluding the identification of these specimens. Of the 795 plant remains from 31ON190, a total of 445 (55.9 percent) were identified.
Zooarchaeological specimens from the general site fill were collected in acid-free bags and labeled. Soil samples taken from features were individually processed through a flotation system that carries away only those soil particles less than 0.0625 inch in size. The light fraction floats to the top and is siphoned off, while the heavy fraction is retained. Both fractions were allowed to air-dry and examined for zooarchaeological specimens, which were placed in acid-free bags and appropriately labeled.
Animal remains were identified using the zooarchaeological reference collection held by Garrow & Associates. A subsample of fish bones was identified by Mr. Daniel C. Weinand at the Natural History and Zooarchaeological Museum, University of Georgia, Athens, under the supervision of Dr. E. J. Reitz.
In addition to species identification, other factors such as
element, side, degree of burning, animal gnaw marks, butchering
marks, and state of bone preservation were also recorded for each
specimen in accordance with generally accepted archaeological
procedures (Olsen 1968; White 1953).