A total of 3742 potsherds was recovered from the Amity Site, of which 907 or 24% are greater than two centimeters in maximum dimension (Table 3; Appendix B). Of the sherds larger than two centimeters, 68% are Colington Series and 27% are Mount Pleasant Series. The remaining 5% is comprised of Mockley Ware, and Hanover. The Amity Site thus contains ceramic series derived from the Middle and Late Woodland Periods.

Middle Woodland Period Ceramics (Figure 27): Three ceramic series or wares characteristic of the Middle Woodland Period in the Middle Atlantic-Carolina Sounds Region are found in the Amity Site collection. Shell-tempered Mockley Ware (Stephenson et al. 1963) is represented by 13 net-impressed (Figure 27j) and one cordmarked sherd. Mockley Ware is frequently recovered in the Coastal Plain of Virginia, where it is dated to the later Middle Woodland Period, circa A.D. 200-900 (Egloff and Potter 1982). No dates for Mockley Ware are available from [Page 47] North Carolina, nor is Mockley Ware ever found in abundance here. However, sizable collections of Middle Woodland ceramics from the North Carolina coast generally contain a few Mockley ware sherds (David Phelps, personal communication 1989).

Twenty sherds greater than two centimeters are assignable to the Middle Woodland Period Hanover Series (South 1976) based on their temper of aplastic clay (Figure 27h-i). Hanover Series pottery -- here subsuming the clay- tempered Carteret Series pottery (Loftfield 1976) as well -- is common in the southern Coastal Plain of North Carolina, but rarely occurs in large amounts north of the Neuse River. Radiocarbon dates from oyster shell have suggested an early Middle Woodland Period (circa 300 B.C. - 100 B.C.) date for Hanover Series pottery in South Carolina (South 1976:1). In North Carolina, clay-tempered pottery has been dated to much later times. In the shell midden of Permuda Island, Hanover pottery co-occurred with shell-tempered pottery in zones II and III dated (uncorrected) respectively to A.D. 500 ± 80 and A.D. 1300 ± 50 (Loftfield 1985). A.D. 1300 seems too late a date for Hanover pottery. It is likely that the Hanover material form Permuda Island is properly associated with the earlier Zone III and that the Hanover specimens in Zone II are ones that migrated upwards through the shell midden (cf. Waselkov 1987:147-148). A date of circa A.D. 500 would seem reasonable for Hanover pottery in North Carolina and is supported by its frequent association with Mockley Ware and Mount Pleasant Series pottery.

Sand-tempered Mount Pleasant Series pottery is represented at the Amity Site by 241 sherds larger than 2 centimeters. Mount Pleasant Net-Impressed makes up 47% (Figure 27a-c) of the series, Mount Pleasant Cord-Marked 26% (Figure 27d-e) and Mount Pleasant Plain 5% (Figure 27f-g). Sherds with surfaces too eroded to identify make up 22% of the series.

Mount Pleasant Series pottery is considered the horizon marker for the Middle Woodland Period in the northern Coastal Plain of North Carolina (Phelps 1983). Although described in a widely cited publication (Phelps 1983) as a ware tempered with sand and pebbles, the published definition of Mount Pleasant Series allows for a much broader range of tempering material (Phelps 1984:41). The typical Mount Pleasant pottery has "variable amounts of fine to medium sand with frequent particles of coarse sand and pebbles (2-7mm)" (Phelps 1984:41). Also included, however, are sherds "with only fine to medium sand temper and others which contain primarily coarse sand and pebbles" (Phelps 1984:41). It is anticipated that further research will establish the pattern of this variation and allow the definition of historically-culturally valid varieties of the Mount Pleasant types (David Phelps, personal communication 1989).

The Mount Pleasant Series is the one reasonably well-dated Middle Woodland ceramic series from the Coastal Plain of North Carolina. Uncorrected dates from the Rush Point site (31Drl5) of A.D. 265 ±65 and A.D. 890 ± 80 bracket the Mount Pleasant phase occupation (Phelps 1981a). Uncorrected dates of A.D. 360 ± 65 from the Baum Site (31Ck9) and A.D. 460 ±85 from the Tillet Site (31Dr35) are also associated with Mount Pleasant pottery. The radiocarbon evidence thus far indicates a later Middle Woodland Period placement for Mount Pleasant Ware, circa A.D. 250 to A.D. 900. Earlier dates would not be surprising, however (Phelps 1983).

The Mount Pleasant Ware from the Amity site is tempered with fine to medium sand with no inclusion of larger aplastics. Sherd thickness averages eight millimeters. Rim sherds make up only about four percent of the collection, suggesting that the typical vessel was deep bodied. Rims were straight to slightly everted, and lips were either rounded or flattened. Occasionally, the lips bear the impressions of the paddle used to treat the surface of the pot. Overall, except for the absence of larger sized temper particles, the Amity site Mount Pleasant sherds are typical of the ware. Certainly it is included within the range of variation expected of the ware as formally defined (Phelps 1984).

[Page 49] Late Woodland Period Ceramics (Figure 28): Of the 907 sherds greater than two centimeters in size, 68% are shell-tempered Colington Series. (In this report, the Colington Series subsumes Roanoke Simple-stamped (Blaker 1952) as a component type). Of the Colington sherds 69% are simple-stamped (Figure 27a-f), 24% are plain (Figure 27g-h) and 7% have exterior surface treatments too eroded or damaged to identify. Remarkably, the Amity site Colington assemblage contains no fabric-impressed sherds. Elsewhere in coastal North Carolina, Colington Fabric-impressed is virtually ubiquitous at Late Woodland period sites (Loftfield 1976; Phelps 1983).

The Amity site Colington sherds are typical of the ware as formally defined (Phelps 1984:44-49). Sherds average 7 millimeters in thickness. Rims make up only three percent of the collection, so a deep-bodied vessel is indicated as the typical form. Rims are either slightly everted or straight. Lips are typically flattened, and on the simple- stamped vessels usually show the imprint of the paddle used to treat the surface.

Interestingly, although temper and surface treatment are completely different between the Middle Woodland period Mount Pleasant ware and the Historic period Colington ware at the Amity site, the indicated vessel form and the rim and lip treatments appear to have remained constant for over a millennium. In general, ceramic assemblages from the outer Coastal Plain seem to display a quite limited range of vessel forms that do not vary through time. This is in marked contrast to the South Appalachian Mississippian assemblages described by Hally (1983), and differs as well even from the more limited temporal variation described for North Carolina Piedmont assemblages (Coe 1964:107).

Colington Simple-stamped (Figure 28a-f): The recovery of Colington Simple-stamped pottery from the Amity Site during its initial surface collection in 1985 was met with great enthusiasm, as shell-tempered simple-stamped pottery has been considered the horizon marker for the Historic Period in coastal North Carolina ever since its. recovery from Fort Raleigh in 1948 (Harrington 1948). Thus, its presence at the Amity Site was considered a quite promising development towards the identification of Pomeiooc.

In addition to its recovery from the late sixteenth century Fort Raleigh, Colington Simple-stamped pottery has been found in an early sixteenth century context at the Riding Ring site (44Vb7) at Great Neck, Virginia. There a pit containing a mixture of Colington Simple-stamped and fabric-impressed pottery (presumably Rappahanock Fabric-impressed) has been dated to 435 ±70 years: uncorrected A.D. 1515 ± 70 (UGA-3294) (Painter 1981). Seventeenth century occurrences of Colington Simple-stamped pottery are common as well (Keith Egloff, personal communication 1989). In particular, the original type definition of shell-tempered simple-stamped pottery as Roanoke Simple-stamped was based on collections recovered from the early seventeenth century trading post of Kecougtan near Hampton Roads, Virginia (Blaker 1952). Shell-tempered simple-stamped pottery is also abundant at the seventeenth century site of Flowerdew Hundred on the James River, Virginia (Egloff and Potter 1982). The temporal placement of Colington Simple Stamped pottery seems to span the Proto-historic and Historic periods, but the dates of its initial appearance and final disappearance are yet to be established. Unfortunately, Harrington’s quibble of nearly 30 years ago can still stand as the final word on the chronology of Colington Simple-stamped pottery: It dates to "an indeterminate period prior to and after the colonizing attempts of the late 16th century" (Harrington 1962:46).

In North Carolina no site other than Amity has produced a sizable collection of Colington Simple-stamped pottery. Haag (1958) found the type at only 14 sites during his survey of the North Carolina coast and only three sites produced more than ten sherds. Two of these sites, Haag’s P14 and P19, are in the Pungo River area and Haag considered them possible candidates for being the sixteenth century village of Aquascogoc (1958:120). The third site, Haag’s H1, is located at Cape Hatteras and Haag (1958:120) considered it a likely candidate for being the [Page 50] sixteenth century settlement of Croatoan. Haag’s speculations may be correct, but a later date for the three sites cannot be discounted. The Pungo River area was inhabited throughout the seventeenth and early eighteenth centuries by the Machapungo Indians, who gave the river its name (Powell 1968), and Cape Hatteras was the home of the Hatteras Indians as late as 1733 as indicated by the Moseley Map (Figure 4). Either of these groups could be responsible for the Colington Simple-stamped pottery.

The limited work done in the Chowan River area by David Phelps (1982) and Paul Green (1986) recovered surprisingly little Colington Simple-stamped pottery. At neither the "Chowanoke" Site (31Hf20) (Green 1986) or the Robert’s Wharf site (31Ga1) (Phelps 1982) did Colington Simple stamped make up as much as 2% of the ceramic collection, in spite of the fact that both sites are assumed to include Historic period occupations. Judging from the limited excavations on Colington Island (Phelps 1981a), Colington Simple-stamped is rare there as well. It made up only 2% of the ceramic assemblage from both 31Dr14 and 31Dr33 and was totally absent from 31Dr15 (Phelps 1981a). On Roanoke Island the excavation of the Tillet Site produced only three Colington simple stamped sherds from an assemblage of 1535 sherds (Phelps 1984). From the Bath Creek area, the excavations at 31Bf115 recovered a ceramic assemblage that included only 4% Colington Simple stamped (Payne and Dahlin 1987).

From the central North Carolina coast, Loftfield has likewise found little Colington Simple-stamped pottery. From his extensive survey (Loftfield 1976), only 30 of 21,135 sherds were classified as White-Oak Thong-marked (here subsumed under Colington Simple-stamped). No simple-stamped sherds were found at the Uniflite Site (Loftfield 1979) and less than 3% of the excavated ceramics from Permuda Island were of this type (Loftfield 1985).

No shell-tempered simple-stamped pottery was recovered from South’s survey of the southern North Carolina coast (1976).

Collington Plain (Figure 28g-h): Colington Plain is somewhat more frequently recovered than Colington Simple-stamped in North Carolina, although it is far from common. Colington Plain makes up 13% of the ceramics collected by Haag (1958), but over one-half of these derive from just three sites, Haag’s H4, A1, and C4. In addition, Haag cautioned that Plain sherds were overrepresented, since much of his material was collected from the surf which had probably eroded some other surface treatments so that they appeared plain (1958:72).

In the Chowan River area Colington Plain made up less than 2% of both the "Chowanoke" (Green 1986) and Robert’s Wharf (Phelps 1982) ceramic assemblages. On Colington Island it made up 3% of the ceramic assemblages of 31Drl14 and 31Dr15 and 4% of the 31Dr33 assemblage (Phelps 1981a). On Roanoke Island it made up 3% of the Tillet Site assemblage (Phelps 1984). Colington Plain makes up 25% of the total assemblage of ceramics from 31Bf115 on Bath Creek, but only 10% of the material from pit features (Payne and Dahlin 1987). Given the disparity between the total assemblage and the pit feature percentages, the total assemblage from 31Bf115 may contain some sherds speciously classified as plain due to their eroded surfaces. Still, 10% is a substantial amount of this type and may suggest a late occupation, perhaps Pomeiooc’s "sister city", Secotan.

From the central North Carolina coast, Colington Plain makes up 5% of Loftfield’s survey collection (1976), 7% of the Permuda Island collection (1985), and 9% of the Uniflite assemblage (1979). From the southern North Carolina coast, shell-tempered plain pottery makes up 6% of South’s survey collection (1976). Surprisingly, within the shell-tempered ware, a plain surface treatment is the most common. South considers shell-tempered plain to be a late type coeval with the Colington Simple-stamped of the northern North Carolina coast (1976:45). The evidence from the Amity site supports South’s conclusion. However, his shell-tempered assemblage is dominated by net-impressed and cord-marked types suggestive of Middle [Page 51] Woodland period Mockley Ware (Stephenson et al. 1963). This makes any temporal inferences drawn from the surface collections even more than usually suspect.

Compared to the dearth of potsherds, the Amity site produced a relative abundance of smoking pipe fragments (Table 4). A total of 149 pipe fragments was recovered. The ratio of [Page 52] pipe fragments to potsherds is thus 1:25 (Table 5). The reason for this abundance of pipe fragments is not known, but the figures from Table 5 suggest that pipe fragments are more common in the Postcontact period.

Unfortunately, the prehistoric sites from the coast contain both Middle Woodland and Late Woodland components and most pipe fragments cannot be assigned confidently to either. Hence it is not possible to say if the increase in pipe fragments is an overall trend of the Woodland period or if the increase in pipes is a phenomenon of the Postcontact period.

It is tempting to speculate that the role of tobacco in Algonquian culture might have changed during the historic period. Hariot makes clear that at the time of the Roanoke colony, tobacco was considered a sacred plant (Quinn 1955:345). The Euroamencans adopted it, of course, as a medicinal and recreational drug. It is possible that following Euroamerican contact, Native Americans, too, might have come to view tobacco smoking as a more secular activity, thus leading to an increase in smoking and a concomitant proliferation of smoking pipes. Given the widespread and far-reaching cultural changes that occurred within Native American societies following contact, such a change in attitudes towards smoking would be hardly surprising.

A somewhat more elaborate version of this hypothesis would hold that not only had Euroamerican attitudes about tobacco smoking spread to the Native Americans, but that a different tobacco might have been introduced as well. Indian tobacco (Nicotiana rustica) is pungent and somewhat narcotic (Janiger and Dobkins de Rio 1976). It was the tobacco described by Hariot (Quinn 1955:345) and the original tobacco grown by the Virginia colonists of the early seventeenth century. It was replaced as a crop by the more flavorful and less narcotic Sweet Caribbean tobacco (Nicotiana tabacum) following the latter’s importation from Trinidad by John Rolfe in 1611 (Barbour 1969). It is known that among Native American groups in the western United States, Nicotiana tabacum obtained from white traders replaced natively grown tobacco for general smoking purposes (Haberman 1984:270). It seems possible that Native American groups in close proximity to the tobacco growing colonies of Virginia might have adopted the smoking and/or growing of this milder tobacco. The adoption of a milder tasting species might well have promoted more widespread recreational smoking, thus leading to a proliferation of smoking pipes.

The latter hypothesis is less appealing than the preceding one in that it is more complex, entailing a change in the plant grown as well as a change in attitudes about it. It does have the advantage of being testable, however, as tobacco seeds can be identified to species under high magnification (Haberman 1984), and pipe dottle is identifiable to species based on its relative proportion of nornicotine to total alkaloid content (Kirby 1982). Thus, given adequate preservation and considerable attention to recovery of plant remains, it should be possible to determine if Nicotiana tabacum is indeed present at sites producing large numbers of smoking pipes. Unfortunately no Amity site pipes contained dottle, nor were tobacco seeds recovered from the flotation samples.

The smoking pipes from Amity are highly fragmentary. Bowl fragments greater than one square centimeter are rare. The vast majority of the stems and about one-half the bowl fragments are undecorated (Table 4), but the small average size likely means that many undecorated fragments are actually plain portions of decorated pipes. Decoration was applied to the pipes in three ways: incision, rouletting and punctation (Figure 28i-l). The fragmentary nature of the specimens precluded the recognition of any complete design motifs.

Incision was the most frequent decorative element employed, occurring on 15 fragments (Figure 28i). Incisions are usually quite fine and shallow. Both single lines and fragments of complex patterns were observed.

Fine "watchwheel" rouletting was observed on 11 fragments. Judging from the quite small fragments at hand, rouletting seemed to be used mostly to form simple designs such as [Page 53] single and double lines (Figure 28k) or chevrons. Although a few fragments bore filled zones of rouletting (Figure 281), nothing that appeared to be part of a "running deer," star or other typical design motif (Mitchell 1976; Henry 1979) could be recognized.

Small circular punctations possibly made with a plant stem occurred on four fragments (Figure 28j). In all but one case, punctation occurred as part of a larger design incorporating incision or rouletting.

The rouietted pipes are the most temporally diagnostic. Such pipes are frequently found in sites of the mid- to late seventeenth century in coastal Virginia and Maryland and more rarely from the Piedmont of North Carolina as well (Table 6 and references therein).

An apparent exception to the seventeenth century placement of rouletted pipes is reported from the Susquehanna area. Kent reports a ‘tidewater" pipe from a Shenks Ferry Phase pit at the Schultz-Funk site in Pennsylvania (1984:148). If properly associated with the Shenks Ferry Phase, this would imply a date of 1550 or earlier. However, the Shultz-Funk site itself dates to the 1645 to 1665 period (Kent 1984), just the time from which rouletted pipes are expected to occur. Thus it seems that the pipe might actually date to the mid-seventeenth century and be only speciously associated with the earlier phase.

Rouletted pipes are rarely reported from the North Carolina coast. Green reports a surface find of pipe with a rouletted "running deer" design from area 20B of the "Chowanoke" site (Green 1986). Haag illustrates five rouletted pipe fragments from Cape Hatteras surface collections (1958:Plate 10). Although both Cape Hatteras and "Chowanoke" are locales where sixteenth century contact period sites should occur (the towns of Croatoan and Chowanoke respectively), they are also areas where Native Americans continued to live at least into the eighteenth century. The Moseley map of 1733 shows the "Hatteras Ind" at Cape Hatteras and the ‘Meherin Ind" at "Chowanoke". Hence, surface finds of rouletted pipes in these areas cannot be ascribed necessarily to the sixteenth century.

Finally it is notable that the one pipe recovered from Harrington’s excavations at Fort Raleigh is not rouletted but is of plain design with a quite shallow elbow angle (1962:Figure 32). it is quite unlike the typical "tidewater" pipe (cf. Henry 1979:Figure 3). Overall, the frequent [Page 54] recovery of rouletted pipes from mid- to late seventeenth century contexts would suggest that this is the most likely period of occupation of the Amity Site.

As is typically the case with coastal area sites in North Carolina, the Amity Site yielded a very small assemblage of stone artifacts (Figure 29; Table 7). Only 91 pieces of chipped stone and a single piece of ground stone were recovered. The ground stone was a fragment of a pipe bowl (Figure 29d) identified as polished lignite or chlorite by Paul Green. It should be recognized that any sizable lithic artifact occurring in the stone-free soils of Hyde County would invite collection by farmers. The absence of groundstone artifacts in particular might easily be explained in this fashion, although the local people know of no such collecting from the Amity site area. Mr. Seth Henry Collins, one of the current landowners of the site, recalls seeing tobacco pipes while plowing the field as a child, but does not recall noticing other artifacts.

Four types of raw material were used to fashion chipped stone tools at the site: chert, jasper, quartzite, and quartz. All occur in the form of small, water-rolled pebbles on the beaches of the Outer Banks. No stone occurs naturally in the immediate site environs.

Only 10 pieces that can be identified as tools are present in the Amity assemblage. Three chipped stone projectile points were recovered, all from plowzone. A small triangular point made of chert (Figure 29a) is probably associated with the Postcontact period occupation. An "eared variety" Yadkin point (Coe 1964:Figure 42) made of quartzite is almost certainly associated with the Middle Woodland Mount Pleasant component (Figure 29c). A quartzite Roanoke Large [Page 55] Triangular point (Coe 1964:110-111) (Figure 29b) may also be associated with the Mount Pleasant component. Its size (38 X 21 X 8), however, is equivalent to the largest of the glass points recovered from the site. Hence an Historic Period association cannot be ruled out. Four flakes exhibit marginal "nibbling" indicative of use as a cutting implement. Three pieces of bifacially worked chert were recovered, but all appear to be small fragments of projectile points. No formalized cutting or scraping tools were recovered.

The by-products of stone tool manufacture dominate the Amity assemblage. Decortication and secondary flakes make up 86% of the assemblage. Two pebbles of grey chert have a single flake removed from one end, as if they have been tested for their flaking properties. Only one core was recovered (Figure 29f). This is a small pebble of brown chert which exhibits bipolar reduction. In addition one piece esquillee was recovered (Figure 29e). This appears to be a fragment of a grey chert biface that was recycled as a source for bipolar flakes. Opposite edges of the biface display extreme battering and crushing, as if attempts to remove flakes continued well after successful results were achieved.

With such a small lithic assemblage any generalization is suspect. An obvious speculation is to ascribe the paucity of cutting implements to the introduction of metal tools acquired by trade. This idea certainly cannot be rejected, but no metal tools were recovered from the Amity site, and McManus (1985:36) found the introduction of metal tools to have had only a limited effect on Postcontact period Piedmont Siouan lithic assemblages. It is also tempting to wonder if the overall paucity and limited variety of lithics from the site might not indicate a special-purpose rather than year-round habitation site (cf. Davis 1985). However, coastal sites often produce few lithics, and southeastern Hyde County is a particularly stone-free area. The use of bipolar techniques at Amity suggests the site inhabitants were attempting to conserve a scarce resource, and this natural scarcity of raw material is probably sufficient to explain the paucity of worked stone at the site.

Glass Beads: Nine glass trade beads were recovered from the Amity site excavations. Three were found in the midden zone of square 140R202 and six were recovered from postmolds. No postmolds defining longhouses or the palisade yielded beads. Surprisingly, the postmolds yielding beads were located near to, but outside of, the longhouses.

[Page 56] No "necklace" or "fancy" beads were recovered. All of the beads from Amity are of drawn manufacture, circular, and small (2 mm to 4 mm in diameter). Six are of simple white seed beads (Kidd and Kidd variety IIa14), and one is a simple blue seed bead (Kidd and Kidd variety IIa41). (All comparisons to Kidd and Kidd were made to the 1983 reprint (Hayes 1983)). The other two beads appear at first glance as simple white beads but closer inspection shows them to be of compound construction. One is Kidd and Kidd variety IVa11, a seed bead with an opaque white layer sandwiched between two clear layers. The final bead is Kidd and Kidd variety IVa13, a seed bead with an opaque white layer surrounding a clear core. Superficially, both of the compound beads mimic the variety IIa14 beads, and I suspect they may be frequently misclassified as such during cursory inspections of large samples of white seed beads (cf. Huey 1983:86)

Although one pale green, spheroidal glass bead was recovered from the ditch fill at Fort Raleigh (Harrington 1962), it is noteworthy that in none of the accounts (Quinn 1955) left by the Roanoke colonists are beads mentioned as a trade item. Instead copper (presumably in the form of jetons) seems to have been the principal trade good (Quinn 1955:282). This contrasts sharply with the later accounts of the Jamestown colonists where blue glass beads are frequently mentioned as items of trade (see for example Barbour 1986:28, 71, 168, 242). It is irresistible to speculate that this change in the English trading inventories between the 1580’s and 1609 was related to the establishment of the Dutch bead manufacturing industry in the 1590’s (Karklins 1983; Baart 1988), a development that can reasonably be expected to have made glass beads more readily and cheaply available to English colonial outfitters.

Cross-dating the Amity site bead assemblage is problematic. Nine beads constitute a small sample, and all nine are of common varieties apparently in use throughout the seventeenth and eighteenth centuries. Furthermore, the composition of a bead assemblage found at any Native American archaeological site is structured not only by the date of the site, but also by the aesthetic preferences of its inhabitants concerning color and form (cf. Smith 1983:151) and by the nationality of the Europeans with which they dealt -- different nationalities proffering different beads (cf. Kenyon and Kenyon 1983:69). However, comparisons with other archaeologically derived bead assemblages do yield some clues as to the most likely period of occupation of the Amity Site.

Archaeologically recovered beads from sixteenth and seventeenth century Amsterdam indicate a wide diversity of bead types, as would be expected from a manufacturing center. Bead assemblages clearly assignable to the sixteenth century are dominated by tubular and striped varieties (Baart 1988). These do not occur at the Amity site. Kidd and Kidd variety IVa11, which does occur at Amity, makes up 19 percent of a pre-1610 assemblage that is otherwise dominated by redwood beads of simple and layered construction, simple black beads and various striped varieties (Baart 1988). Monochrome seed beads in colors other than redwood do not seem to have been popular until the mid-seventeenth century. However, a post-i 647 assemblage is reported to be composed of "primarily of blue, white and black circular varieties" (Baart 1988:72). This description would, of course, encompass the nine Amity site beads.

Not surprisingly, the most complete studies of sixteenth century trade beads in the New World concern Spanish dominated areas, as the Spanish were the most active colonial power during that period. Although seed beads occur throughout the period of Spanish influence, bead assemblages of the late sixteenth century were dominated by "fancy" or "necklace" beads. In particular, Florida cut crystal, true amber, and eye beads are specified as markers of the late sixteenth century (Smith 1983), while a tumbled spherical or barrel-shaped blue bead is the most common variety during the 1560 to 1620 period (Smith and Good 1982:53; Smith 1987:32). None of these varieties occur in the Amity Site sample.

Studies of beads from early Contact Period sites outside the area of Spanish hegemony are more spotty. In New England (Bradley 1983) glass beads are rare prior to 1600. The few [Page 57] beads recovered from sixteenth century contexts are mostly of tubular form, although two round blue beads are also known. From 1600 to 1620, oval and round monochrome white and dark blue beads are most common, while circular varieties in black and dark blue are the most popular types from 1620 to 1650. Small amounts of circular white and light blue varieties are also recovered from the 1620 to 1650 period, as are examples of variety IVa11. In the following period (1650 to 1675), white or grey circular seed beads and blue circular seed beads are abundant, and both IVa11 and lVa13 occur in sites of this period as well (Bradley 1983). Hence, the Amity site bead assemblage is matched entirely by New England assemblages of the third quarter of the seventeenth century.

From Fort Orange, a Dutch trading post at modern Albany, New York dating from 1624 to 1676, the most common bead variety is IVa11, which makes up 27 percent of the bead assemblage (Huey 1983). The other compound bead variety recovered from the Amity Site, IVa13, makes up 2 percent of the Fort Orange assemblage and is first recovered from a pre-1 624 context. Simple circular white seed beads were not recovered, and simple circular blue seed beads were recovered only in trace amounts (Huey 1983).

Likewise in the well-studied Susquehanna bead assemblages, simple circular white seed beads are also apparently absent (Kent 1983,1984:211f.). Simple circular dark blue (actually "ultramarine") beads occur throughout the Susquehanna sequence, but Kent considers them the hallmark of the 1600 to 1630 period (Kent 1984:213). Variety IVa11 occurs from 1600 to the 1680, but variety IVa13 occurs only after 1690 (Kent 1983).

The Seneca and Oneida Iroquois trade bead sequences shed little light on the Amity site material. Circular forms do not seem to have ever been important, and red and black are more common colors than white or blue (Wray 1983; Bennett 1983). Variety IVa13 is reported from early contexts (ca. 1595 to 1625) (Wray 1983; Bennett 1983). However, the dating of this phase is not secure, and it may not date prior to 1615 or later (Kenyon and Kenyon 1983).

From the Cheaspeake Bay area glass trade beads are surprisingly rare (Miller et al. 1983). No sixteenth century sites in the Chesapeake area have yielded trade beads, and from seventeenth century sites monochrome blue beads are the most common sort. Simple monochrome blue beads (mostly round but including circular as well as oval varieties) are abundant at the St. John’s site, a trader’s house dating from 1638 to 1680 in St. Mary’s City, Maryland, and from two Potomac River ossuaries, Patawomeke and Piscatoway Creek, presumably dating to the second third of the seventeenth century. An assemblage of beads from a Rappahanock River ossuary dating to the first half of the seventeenth century was comprised overwhelmingly of a white-with-red-stripes variety but also contained about 25 percent circular blue beads as well as a few examples of variety IVa13 (Miller et al. 1983). From the James River area, Native American sites of the first third of the seventeenth century have yet to be identified (Miller et al. 1983; Keith Egloff personal communication 1989). Two ossuaries dating to the late seventeenth-early eighteenth century have yielded 11 varieties of glass beads. Eight are simple monochromes (variety IIa). Of these, four are white varieties and two are blue. The earliest reported recovery of circular blue and circular white beads from Anglo-American sites on the James River is from contexts dating circa 1630 to 1650 (Miller et al. 1983). The two compound beads recovered from the Amity site have yet to be reported from the James River area.

Finally, from the North Carolina Piedmont two Contact Period ‘Siouan’ sites have analyzed bead assemblages (Carnes 1986,1987,1988). The Mitchum site (ca. 1660 to 1680) is dominated by monochrome blue seed beads (97% of a total of 1960 beads) with white seed beads the next most popular (2% of total) (Carnes 1987). From the Fredricks site (ca. 1670 to 1710) (Carnes 1987) more than 11,000 beads have been analyzed with simple white seed beads overwhelmingly the most common, followed by black, blue and redwood varieties (Carnes 1986, 1987,1988). "Necklace" and "fancy" beads are rare at both sites, and from neither site are the two compound varieties recovered from Amity reported. However, it seems possible that both [Page 58] IVa11 and IVa13 varieties occur, but were lumped into a general "white seed bead" category (cf. Bradley 1983:33) or perhaps overlooked amid the thousands of white seed beads of (presumably) simple construction.

To summarize briefly, the Amity site has produced no beads such as "tumbled purple" (Smith 1983) or black and white frit-cored beads (Kenyon and Kenyon 1983) that might be considered "hallmarks" of the late sixteenth century. However, some of the Amity beads do appear in quite early contexts. In particular, the IVa11 variety was manufactured prior to 1610 in Amsterdam (Baart 1988), and variety IVa13 was found in pre-1624 contexts at Fort Orange (Huey 1983). Circular blue beads have been recovered from the Schultz site (ca. 1575 to 1600) in Pennsylvania (Kent 1984), and circular white beads are relatively common in New England in 1620 to 1650 contexts. By the same token, all four varieties have also been recovered from post-1680 contexts: IVa11, IVa13, and circular blue varieties from Conestoga Town (Kent 1984:Table 8), and IIa14 from the Fredricks site (Carnes 1988).

The best fit with the Amity site assemblage is provided by the Later Plantation Period (1650 to 1675) sites in New England, as these are the only other sites in which all four of the Amity site varieties co-occur as an assemblage. This is also the time during which the Carolina Sounds region began to be settled by colonists from Virginia (Powell 1975). It seems likely that the Native Americans of the region would have acquired small amounts of trade goods during this time.

Glass Projectile Points: Two fragmentary and two complete glass projectile points were recovered during the Amity Site excavations (Figure 30a-d). All are fashioned from transparent green glass by pressure flaking along the margins. The faces of the points are largely unworked. The glass used to fashion the points is apparently derived from square section bottles such as were in use from 1625 to 1675 (Noel Hume 1970).

The two examples from the 1986 excavations are distal fragments. One was recovered from the fill of feature 6, an intrusive Euroamerican period ditch (Green 1987:25). The second was found following the removal of plowzone from "tractor trench II" (Green 1987:Appendix A).

A small triangular glass projectile point recovered in 1988 from the plowzone of square 170R210 measures 16mm X 14mm X 2mm and is deeply basally indented. If chipped from stone, it would be easily assignable to the Clarksville Small Triangular point type (Coe 1964). The final glass point was recovered during the 1988 season by an Operation Raleigh Venturer while flatshoveling the top of subsoil in square 180R208. He reported it "popped out from one of the dark spots" (i.e. a postmold) but could not recall from which one. The point is an isosceles triangle measuring 30mm X 18mm X 2mm. Its shape is that of a Roanoke Large Triangular point (Coe 1964).

Glass projectile points are known from two dated sites in coastal Virginia. Both sites are characterized by Colono-lndian pottery and numerous trade goods (MacCord 1969,1976). A single triangular point of transparent glass measuring 15mm long X 16mm wide is reported from the Camden Site on the Rappahanock River. Coins from this site provide a terminus post quem date of 1672 (MacCord 1969). Another glass projectile point, measuring 27mm X 13mm X 2mm, was recovered from the possible site of Portobacko also on the Rappahanock River in Essex County, Virginia (MacCord 1973). This site is dated to the 1660 to 1680 period (MacCord 1973).

Three chipped glass projectile points and three chipped glass scrapers are reported as surface finds from coastal Virginia (McCary 1962). McCary offers no discussion of the sites and only general locational information ("Pamunkey River shore", "Jamestown Island", and "confluence of the Accokeek and Potomac Creeks"). He dates the artifacts to the 1625 to 1650 period based on their manufacture from square section bottles (McCary 1962), but such bottles are now considered to have remained in use for at least another 25 years (Noel Hume 1970).

[Page 60] In addition, three glass projectile points have been recovered from the site of Upper Saratown (31Sk1a) in the Piedmont of North Carolina (Jane McManus, personal communication 1989). This site dates circa 1660 to 1680 (Davis and Ward 1989).

Documentary evidence for the use of glass by Native Americans to fashion projectile points is scant. There is reference in the colonial records for a trade in glass bottles to Virginia Indians in 1628 (Mcllwaine 1924:165). Presumably the bottles were used by the acculturated Native Americans of coastal Virginia for tool production. John Lawson in his travels through the Carolina backcounty in 1701 observed at the town of Lower Quarter "very long arrows, headed with pieces of glass, which they had broken from bottles" (Lefler 1967:63). Lower Quarter is presumed to be near the Piedmont-Coastal Plain fall line but has yet to be identified archaeologically (Simpkins 1986).

Other Aboriginally Utilized Glass: The plowzone of the Amity Site produced 30 fragments of green glass seemingly identical to that used to fashion the glass points. It seems likely that these date to the time of the aboriginal occupation, although this cannot be demonstrated. In fact, the only glass from a firm context is one pressure flake of green glass recovered from a postmold in square 180R210.

Kaolin Pipe Fragments: Six kaolin pipe bowl fragments were recovered from the Amity Site. All are undecorated and from plow disturbed contexts. Only one retains sufficient rim circumference to allow calculation of bowl diameter (Figure 30f). It has an interior diameter of only one-half inch (12 mm) which suggests a seventeenth century date (Noel Hume 1970).

A dozen kaoline pipe stem fragments, all undecorated, were recovered as well (Figure 30g). All were found in historically disturbed contexts. Two fragments have bore diameters of 9/64 inch, four of 8/64 inch, two of 7/64 inch, one of 6/64 inch, and three of 5/64 inch. Application of the Binford (1978) formula (Date = 1931.85 - 38.26 X Mean Diameter) indicates an occupation date of 1661. This agrees well with the evidence of the rouletted terra cotta pipes, the trade beads and the glass projectile points in indicating the seventeenth century as the most likely period of site occupation. However, the sample size is quite small, making the date suspect. Furthermore, bore diameter exhibits a bimodal distribution. It seems possible that some of the fragments with small bores may not be associated with the aboriginal occupation. Eliminating them from the calculation produces, of course, earlier dates. Eliminating the three 5/64 inch diameter stems produces a date of 1634, while eliminating the 5/64 inch and the 6/64 inch stems produces a date of 1626.

Obviously, with a sample this small and no archaeological context to separate stems from different components, use of the Binford formula is little more than a game. It is significant, however, that 8/64 inch and 9/64 inch stems occur at Amity. According to Harrington (1978:Figure 1), these predate 1680. As this is prior to the Euroamerican settlement of the Hyde County region, it seems most likely that these pipes reached the site as trade goods. The absence of seventeen century Euroamencan ceramics from the site further supports the contention that the kaolin pipes are trade goods and not part of a Euroamerican occupation.

Kaolin pipes were not manufactured during the sixteenth century, and so they cannot have derived from Fort Raleigh colonists. Therefore, it seems likely that they, like the glass beads and bottle glass, were received by the Native Americans as trade goods during the period subsequent to the founding of Jamestown and prior to the Euroamerican settlement of eastern Hyde County after the Tuscarora War. Although suspect on statistical grounds, the mean date of 1661 derived from the Binford formula seems as good an estimate as any.