The Amity Site: Final Report, Chapter 2



[Page 13]

The English established their sixteenth century colony on the north end of Roanoke Island. Their principal interactions with Algonquians were with the occupants of the town of Roanoke, with whom they shared the island, and with the people of Dasemonkepeuc, a town located on the mainland opposite Roanoke Island and united with Roanoke under the werowance Wingina. From the Roanokes the English quickly began gathering geographical information on the surrounding regions. Barlowe (Quinn 1955:11 0), reporting on the first voyage of 1584, provides us with the earliest mention of Pomeiooc.

The following year Sir Richard Grenville, Ralph Lane, Thomas Hariot, John White and several others explored the Pamlico Sound region. They departed on June 11, 1585.

From this terse account we learn that Pomeiooc was not visited by Barlowe in the previous year and that the English spent the night there in 1585. This initial exploration of the Pamlico Sound apparently was not followed by further explorations to the south of Roanoke Island (Quinn 1955:256), the interests of the English shifting to the Chesapeake Bay and Chowan River regions.

During 1586 the English were involved in considerable machinations, intrigues and conflicts with the Algonquians of the Chowan River, Albemarle Sound and Roanoke Island areas, but not once did Colony Governor Ralph Lane refer to Pomeiooc or any other Pamlico Sound town (Quinn 1955:255-294).

Pomeiooc reappears in the colony narratives in 1587. John White, the governor of the final "Lost" colony, attempted to re-establish peaceful relations with the Algonquians.

[Page 14]

These five brief comments are the total mention of Pomeiooc in the narratives of the Roanoke colony. Obviously, compared to Roanoke and Dasemonkepeuc, the two towns nearest the colony, or even to the more distant Croatoan and Chowanoc, Pomeiooc played a negligible role in the history of the English colony. Although there is to North Carolinians an undeniable romance attached to everything surrounding Sir Walter Raleigh�s "Lost Colony", the significance of Pomeiooc is less historical than ethnographic. It was a typical example of a coastal Algonquian town as they existed before the disruptions caused by the European intrusion. General scholarly interest in the town derives from Thomas Hariot�s rich description of Algonquian culture (Quinn 1955:314-387) and particularly from John White�s watercolors of the town and its inhabitants. In addition, the maps created by Hariot and White provide the potential means of locating some of the Algonquian towns. Pomeiooc, as an apparently inland rather than coastal town, has seemed a particularly inviting object for search, since it offers the best chance of having avoided destruction by coastal erosion and rising sea level.


Pomeiooc was visited by the artist John White on July 12, 1585. Whatever sketches, drawings or paintings he made on that day have not survived (Quinn 1955:392). Extant today are a set of watercolor drawings made by White probably between 1586 and 1590 (the manuscript drawings), a set of engravings published by Theodor De Bry in 1590 and based on a version of the White watercolors different from the manuscript drawings, and finally a set of copies (the "Sloan copies") made prior to 1614 presumably by members of White�s family and based on a version of the drawings different from both the manuscript drawings and the version used by De Bry (Quinn 1955:396). These three versions of the drawings provide three slightly different depictions of Pomeiooc. In addition, De Bry incorporated Pomeiooc into the background of White�s figure of an "Old Indian Man" (Hulton 1984: Figure 13).

The Manuscript Drawing: In the manuscript version of the White drawings (Figure 5), Pomeiooc is shown as a palisaded village of 18 pole-and-mat longhouses. The palisade has two entrances formed by overlapping sections of the palisade wall. Some of the longhouses show interior platforms (benches or beds) along the walls. One large longhouse has a roof with a central peak; whereas the other longhouse roofs are of the typical Algonquian "arbor" style. The longhouses are arranged roughly into two concentric circles surrounding an open courtyard with a large central fire. Around the fire are about a dozen individuals, and about another two dozen--and a dog--are scattered about the village. White�s caption reads (Quinn 1955:415):

The De Bry Engraving: De Bry�s engraving of Pomeiooc (Figure 6) shows a few minor differences. The palisade is made of much more substantial poles and has only one entrance. The house with the peaked roof is shown with a hexagonal floor plan. De Bry adds a background that includes an open woods, a cornfield, some sunflowers and a pond from which three figures draw water. The foreground includes a ridge with a few stylized plants. Thomas Hariot�s caption (Hulton 1984:187) reads (Quinn 1955:417):

Although the above passage is extremely informative concerning the spatial organization of Carolina Algonquian chiefly villages, it does not make clear what the temple actually represents. Harlot in his Briefe and True Report (Quinn 1955:373) explains that the Algonquian temples are called Machicomuck and are the repository of the Algonquian idols called Kewasowak (singular Kewas). In his caption for the engraving of Kiwasa [sic] (Hulton 1984: Figure 25), Hariot states that the illustrated kewas "is placed in the temple of the town of Secotan as the keeper of the kings� dead corpses" (Quinn 1955:425). In the engraving of "The Tomb of Their Werowans or Chief Lords" (Hulton 1984: Figure 26), kewas and the arrayed corpses of the werowances are shown placed on a mortuary scaffold enclosed within a longhouse wherein lives a priest. It would seem a reasonable inference that Hariot�s temple at Pomeiooc is such a longhouse enclosing a mortuary scaffold.

However, this is confounded by the fact that the mortuary longhouse shown at Secotan does not display the distinctive peaked roof of the Pomeiooc temple, nor does Hariot explicitly refer to the Secotan structure as a temple. Hence, it remains possible that the Pomeiooc temple and the Secotan mortuary house represent two functionally distinctive structures. On the other hand it may be that the Pomeiooc temple is merely a more architecturally elaborate version of the Secotan mortuary house, just as the palisaded village of Pomeiooc is a slightly more elaborate version of the unenclosed Secotan. The latter interpretation seems more likely but leads to the conclusion that unburied corpses were kept within the village palisade. Although perhaps repugnant by modern standards, such a situation is in accord with Beverly�s observation that the Virginia Algonquians "never fail to secure within their palisado all their religious reliques and the remains of their princes" (Wright 1947:177).

The Sloan Copy: The depiction of Pomeiooc in the Sloan Copies (Figure 7) omits the foreground, and the background shows extensive cornfields and three ponds. The palisade shows two entrances. The figures placed around the village are modified and the dog is omitted. The arrangement of the houses is modified and a nineteenth structure is added. This is a large, apparently hexagonal structure with a peaked roof. On its interior platform sit two figures.

De Bry�s "An Aged Manne...": A final depiction of Pomeiooc is incorporated into the background of De Bry�s "An Aged Manne in His Winter Garment" (Figure 8). As the figure of the man eclipses most of the town, little detail can be gathered, but the palisade entrance is apparently shown as a gap rather than as a passage between overlapping wall sections. The man is shown on a high bluff overlooking Pomeiooc and its cornfields. Beyond the town is an orchard­like forest, then water, then an island chain. This topography is impossible in that no such high bluffs exist on the Pamlico-Albemanle peninsula. Further, there is no such island chain extant [Page 18] today which can be seen from the Hyde County mainland. Lane indicates that in 1586 as well, the Pamlico Sound was "without kenning of land" (Quinn 1955:256). Overall the background of the picture seems to be a melange of elements from different drawings (cf. Quinn 1955:419 note 3) combined fancifully by the engraver. It does not inform us about Pomeiooc.


To summarize the pictorial information, it seems safe to conclude minimally that Pomeiooc was a palisaded village of substantial size made up of 18 or so longhouses arranged around a central courtyard. One longhouse, the �kings�, may be noticeably larger. The palisade entrances were one or two in number and were formed by overlapping wall sections. One or more ponds were likely located nearby. These traits should be archaeologically recognizable.

The temple is problematic. Its distinctive root would, of course, leave no archaeological trace. A hexagonal floor plan should be archaeologically recognizable, and may serve to distinguish such temples, provided they were, in fact, hexagonal (contra the manuscript drawing), [Page 19] and the "kingly" residences were not (contra the Sloan copy). If the interpretation of the temple as a charnel house is correct, then a subterranean ossuary may be located nearby.

As of town of 17 or 18 domestic structures, Pomeiooc likely had a population of about 175 people (see above p.9 and Potter 1982:49 for a discussion of Algonquian population estimations). A population of this size should have produced a fair amount of debris. Of course, the amount of archaeological material present at a site depends not only on the size of the population, but also on the length of time the site was occupied, refuse disposal patterns, and post-depositional factors governing preservation (see discussions in Ward 1980; Hassan 1981). Certainly there is no inevitable mathematical relationship between the population of a town and the amount of archaeological material preserved as its remains. Nonetheless some insight can be obtained by examining other sites.

The early eighteenth century Fredricks site in the North Carolina Piedmont, although the remains of a society culturally distinct from the coastal Algonquians, is especially informative in that the entire site has been exposed allowing the estimation of both its population size and its duration of occupation. The Fredericks site contained about 11 houses and is considered to have had a population of about 75 people and to have been occupied for less than a decade (Davis and Ward 1989). It is, then, considerably smaller than Pomeiooc, and there is no reason to think it was occupied for a longer period of time. The excavation of the site yielded potsherds at a density of 42 per square meter of excavated area; the plowzone alone yielding 37 per square meter (Davis 1988).

In the South Carolina Coastal Plain, excavation of the early eighteenth century site of Wachesaw Landing yielded potsherds at a density of 40 per square meter of excavated area (Trinkley, et al. 1983).

Within the Carolina Sounds region, two sites yielding evidence of structures have been excavated. Excavation of the prehistoric seasonal village of Uniflite yielded potsherds at a density of 26 per square meter of excavated area (Loftfield 1979). The initial archaeological exploration of Permuda island, designed to test both apparently unproductive as well as obviously productive areas of the island, yielded potsherds at a density of 32 per square meter of excavated area (Loftfield 1984). Excavation of two sites considered to be seasonal subsistence camps, the Tillet (Phelps 1984) and the Rush Point (Phelps 1981a) sites, yielded potsherds at respective densities of 28 and 11 sherds per square meter of excavated area.

Judging by the above sample of sites, it would appear that village sites produce about 25 to 40 sherds per square meter of excavation and seasonal camps about half so many. Although refuse might be distributed in a markedly uneven pattern at a village that incorporated both sacred and secular space (Ward 1980, 1985), there is no prima facie reason to think the overall yield of material from Pomeiooc should be marked less than other village sites.

Finally, as a site visited by Englishmen in the late sixteenth century, Pomeiooc should be expected to yield trade goods diagnostic of the period. Unfortunately we have scant knowledge of the objects offered by the English (Quinn 1985:104, 265). Metal in general and edged tools in particular seem to have been highly sought by the Carolina Algonquians (Quinn 1955:100-101), and the documents left by the colonists suggest "copper� to have been an important item of trade (Quinn 1955:282). The copper may have been in the form of jetons, casting counters made of a brass-like alloy, as jetons have been recovered from Fort Raleigh (Harrington 1962), Jamestown and the possible sixteenth century site of Croatoan (Hulton 1979). I would suggest that jetons are the closest thing we have to a "smoking gun" denoting the period of the Roanoke colony.

[Page 20] Glass beads are likely to have been traded to the Algonquians as well, and a single large green glass bead was recovered from the Fort Raleigh excavations (Harrington 1962). Trade bead assemblages from the late sixteenth century Native American sites are dominated by large, often multi-colored necklace beads, but these beads are mostly derived from the Spanish (Smith and Good 1982; Smith 1987). The documents left by the Fort Raleigh colonists do not mention beads at all. Still, necklace beads would be an expected find at Pomeiooc.


There are extant today five maps of the Carolina Sounds region based on the work of John White and Thomas Hariot during the 1584-1590 period (Hulton 1984:32). Two do not concern us: De Bry�s "Arrival of the Englishmen" (Hulton 1984: Figure 6) does not encompass the Pamlico Sound region, and White�s map of Florida to Chesapeake Bay (Hulton 1984: Plate 59) adds no new information concerning the Carolina Sounds but merely incorporates a small scale version of his map of Virginia (Hulton 1984: Plate 60). Potentially useful for locating Pomeiooc are White�s manuscript map of Virginia (Hulton 1984: Plate 60), the De Bry engraving based on it (Hulton 1984: Figure 5) and the pilot�s sketch map of Raleigh�s Virginia (Hulton 1984:10).

Although the De Bry engraving (Figure 9) includes 16 villages not shown on the manuscript map and differs in the locations of others, the two maps agree on the location of Pomeiooc. It is shown as being near the southwestern "corner" of Lake Mattamuskeet between the lake and Pamlico Sound. In both maps Pomeiooc is shown set well back from the Pamlico shoreline. It could be suggested that this placement is a cartographic convention designed to keep the town symbol from merging with the shoreline. (All of White�s and De Bry�s towns are, in fact, recessed from the shore to some degree, although no other to such an extent as Pomeiooc.) However, the inland location is confirmed by the pilot�s sketch map, which also shows Pomeiooc much closer to the lake than to the sound. Too "sketchy" to be otherwise useful as a locational tool, the pilot�s sketch map can hardly be said to be conventionalized (Hulton 1984:10). Confusingly, an inland location appears at odds with the verbal descriptions of Pomeiooc�s location. Barlowe described the river Occam as being the one "on which standeth" Pomeiooc (Quinn 1955:110). Quinn (1955:110 note 2) states, however, that Occam refers to the land beyond the water and not the water itself. This would seem to remove any necessity of inferring a waterside location from Barlowe�s passage. Furthermore, as Barlowe never visited Pomeiooc (see above p.13), his locational information should not carry the weight of White�s map based on firsthand knowledge.

Much more difficult to reconcile is Hariot�s caption accompanying De Bry�s engraving of the "Cheiff Ladye of Pomeiooc" (Hulton 1984: Figure 12), wherein Hariot states "...neere the lake of Paquippe, ther is another towne called Pomeioock hard by the sea." The Oxford English Dictionary offers as denotations for "hard by" the phrases "close by, in close proximity to, close to, and very near to" (OED 1933:Vol. 5:87). The inclusion of "very near to" in this group would suggest it to be an intensified version of "neere", implying a location closer to the sound than to the lake. Certainly to twentieth century ears, the caption suggests a waterside location at odds with the maps. However, it has been recognized for some time that there are discrepancies between some of White�s captions for the manuscript drawings and Hariot�s captions for their De Bry counterparts (Quinn 1955:419 note 4). Quinn speculates that Hariot may have been misled occasionally by De Bry�s embellishments of White�s drawings (Quinn 1955:41 4-415 note 5). Since the engraving of the "Cheiff Ladye" includes a prominent shoreside background not present in the White painting, it seems quite possible that Hariot used this rather than his personal recollections of the place as the basis for an erroneous "hard by the sea" description.

[Page 22] Although the cartographic evidence offers no support for considering Pomeiooc a shore-side village, it might be argued on logical grounds that an inland location for a village is implausible in the Carolina Sounds region, since until the era of modem road building, transportation by water was always more practical. In answer, it can be offered that Hariot in his report states that all villages "but a few" were near the coast (Quinn 1955:369). The qualification of the statement indicates that there were at least a few inland villages. It seems feasible that the relatively large expanse of mineral soils suitable for native agriculture between Lake Mattamuskeet and the Sound might have been sufficient motivation to locate a village there. Finally, the place name "Pomeiooc" seems to translate "they take refuge" (Quinn 1955:870). This might allude to one useful function of a palisaded village located away from the major water routes.


Historians have speculated about the modern location of Pomeiooc for over a century (Hawkes 1858; Mook 1939; Quinn 1955), and all agree the area of Wysocking Bay to Engelhard is the general location. Archaeologists have made a much slower start in addressing the problem. The first archaeological survey of Hyde County was in 1954-1955 when William Haag (1958) surveyed portions of the Outer Banks and the nearby mainland in conjunction with the establishment of the Cape Hatteras National Seashore. Haag was looking primarily for evidence of the Lost Colony, but attempted to acquire a representative sample of site types (1958). Haag did not attempt to locate Pomeiooc. His survey was limited mostly to shoreline locations, and judging from his map of site distributions (1958:Figure 2), it was in Hyde County apparently restricted to areas west of Rose Bay.

In 1983 the National Park Service funded through the North Carolina State Historic Preservation Office a general archaeological survey of Hyde County (Cook 1984). This eight month survey was aimed at providing the basic information on site distributions, size, temporal range, and significance necessary for management of the county�s cultural resources (Cook 1984). Like Haag, Cook emphasized shoreline locations, but did survey some inland areas, especially along the ridge surrounding Lake Mattamuskeet. Although discovery of Pomeiooc was not an expressed goal of this survey, considerable acreage in the Engelhard region was examined. In all, 3760 acres were surveyed and 15 sites were located, mostly along the shores of the Pungo River, Rose Bay and Swanquarter Bay. Two small scatters of Native American material (31Hy19 and 31Hy20) were discovered along the ridge northeast of Lake Mattamuskeet (Cook 1984: Figure 12.). Neither site is substantial enough to be Pomeiooc. Site 31Hy19 yielded only 14 sherds, none clearly assignable to the Late Woodland period, and 31Hy20 yielded but two flakes (Cook 1984:15).

In 1984 the Hyde County Historical Society funded a ten day survey that was aimed expressly towards discovering the remains of Pomeiooc (Lautzenheiser 1984). The survey emphasized the area southeast of Lake Mattamuskeet, especially the portions from Gray Ditch to the Waupopin Canal. In all, 1340 acres were surveyed, and four sites were located, two Euroamerican farmsteads and two very small Native American sites. The Native American Sites (31 Hy31 and 31 Hy32) produced less than 10 flakes, a lithic artifact and a single sherd apiece (Lautzenheiser 1984:33). The sherds are both of the Middle Woodland period. Hence the sites are both too early, as well as much too small, to be Pomeiooc.

In 1985 Paul Green, then of the Department of Sociology and Anthropology at East Carolina University, continued the survey for Pomeiooc. Funding was provided by the America�s Four Hundredth Anniversary Committee, the Z. Smith Reynold�s Foundation and the East Carolina Bank (Green 1985). Green surveyed about 600 acres and found four Euroamerican sites and one Native American site. Among the artifacts recovered from the Native American site were sherds of Colington Simple-stamped and Plain pottery and fragments of rouletted smoking pipes.

[Page 23] Such pottery (Harrington 1948; Blaker 1952) and pipes (Emerson 1988) are horizon markers of the Postcontact period in the coastal regions of Virginia and North Carolina; hence the site was considered a strong candidate for Pomeiooc. The site was given the designation 31Hy43 in the state�s permanent site registry and later given the name "Amity Site" after a nearby Methodist church.

With the identification of the Amity Site as a "possible Pomeiooc", the emphasis of the archaeological endeavors in Hyde County shifted to excavation in order to establish if the Amity Site was indeed the famed town. However, during the winter of 1988-1989, the author and David Jones conducted a survey of another 520 acres considered to have high probability of Native American settlement. These areas were along the ridge south of Lake Mattamuskeet and along the western shore of Middletown Creek. Two Native American sites were found. The first, assumed to be of the Middle Woodland Period, yielded only seven sherds, all sand tempered and only one greater than two centimeters in diameter. The second site produced only 20 sherds with seven greater than two centimeters. The sherds are shell-tempered Colington Ware, but no artifacts typical of the Historic period were found. Both sites are quite small and likely represent short term campsites or small farmsteads. Neither is of sufficient size to represent a village of 18 houses.

Chapter 3 > > >


Copyright 2001
Carolina Algonkian Project, All Rights Reserved