Archaeology of Coastal North Carolina, Part 5


Site Descriptions (Continued)

Part 5

NOTE: Information in this section has been edited for the protection of archaeological sites. Please visit:

North Carolina Archaeological Resources Protection!!!

Pamlico Sound

The shores of this more southernly estuary were searched with greater zeal than Albemarle or Currituck for two reasons. First, the archeological materials were more numerous and, second, the historical data have shown that the English colonists at Roanoke had enjoyed better relations with Indians here than elsewhere.

(Page 48) If the colonists left Roanoke willingly, this river might be one they would look upon with favor as a new homesite.

Conditions of distribution along Pamlico Sound shores are similar to those along Albemarle. The south shore was only sparsely inhabited while many rather large villages dotted the north shore. Pamlico Sound is the broadest and deepest of the estuaries but these features would not attract aborigines. Its width today is probably greater than in the sixteenth century. It widens only through violent storm action; it tends to build up a tidal flat of grass and accompanying silt during quiet periods.

Along Pamlico River bluffs for about eight hundred or more feet is an excellent shell midden. Its width is variable but surface indications are apparent six hundred feet back from the bluff. It has been known to local residents for many years as an Indian village site, and previous investigations have been reported but none could be specifically determined. The shell content Is heavy; some areas are almost total shell with little included soil. Because of the apparent extent of the site, a stratigraphic test was made here. (See below.)

(Page 49) A thin but extensive midden that is judged by many to be the location of Seca as indicated on John White's map. The sparseness of the midden prevents an accurate delineation of the site, but seven hundred feet back from the bluff--the site is about fifteen feet above water--shell fragments and potsherds are to be found.

Encroaching Pamlico River waters have cut a low escarpment at this point. A small midden is revealed about six feet above water thinly scattered along five hundred feet of share.

Nt more than five feet above water line is a thin midden. Presently, it cannot be more than one hundred feet in diameter but a larger size in previous times is indicated. Storm waves easily break over this low bank which is the only remaining high spot along the point.

On a point is this long, thin shell midden. The majority of the sherds recovered from the site were picked up from the water. Hurricane waves have concentrated the (Page 50) shells in a line along the shore which is not more than four feet above normal water level. The shells and sherds are confined to a zone some five hundred to six hundred feet long and nowhere more than one hundred feet in width. The size of the site today probably bears no resemblance to its former condition. In 1955 the "tide" produced by hurricane Connie of August raised the water six feet over the point and carried inland one-third of a mile. The midden remains are probably being moved inland by each storm but eventually will be entirely under water. (Compare this evolution with Gum Point Site, P 26.)

The land stands nearly ten feet above the water of Pamlico River. Here a midden marked by black humus earth to depths of two feet extends nearly one thousand feet along the wave-cut bank and about two hundred feet inland. Shell lenses (also about two feet thick) are cross sectioned in the bank. The bank is being cut back at a rate rapid enough to topple live trees in the river or leave them standing as isolated, elevated stumps. The sherds in the midden are not waterworn, and it is probable that only the most severe storm sends water over this part of the point.

A thin scattered midden of a few sherds and shells extending over in area nearly one thousand feet long but not more than one hundred feet (Page 51) wide.

This may hardly be labelled a "site" since it consists of a few sherds scattered along the beach. (P 10 was the designation given a "tarpot" investigated in the area.)

A sparse distribution of a few sherds.

Along the shore for a distance of eight hundred feet, midden materials were recovered from the water.

Both Indian materials and modern white artifacts were found. The owners of this property believe this to be the remains of an early village now entirely destroyed by encroaching Pungo River waters. In March, when the westerly winds blow the waters away from this shore, brick piles and (Page 52) even streets are observed. Certainly a tremendous amount of old china, glass, and crockery could be obtained from this site. The Indian midden is still preserved, probably only in part, on the shore. Elevated about eight feet above the water level, the midden of shells, bone, and other refuse is scattered over a one-thousand-by-two-hundred-foot zone that must have been a large village at one time.

This site is one of the largest middens in the area with much of the shell content far enough back from the edge of that it in not broken to bits. Its size, stated conservatively, in fifteen hundred feet north-south, and one thousand feet wide. Nowhere was the midden determined to be more than eighteen inches deep. The possible relationship of this midden, and others in the vicinity, to the historic Indian village Asquatock will be considered in a later section.

Fragments and patches of a shell midden were found. Its present size of five hundred by two hundred feet gives little evidence of its original extent. For years much of the shell content has been removed for road metal. The remainder is nearly wholly oyster. Presently the midden is one hundred feet from the shore; the beach ridge here is about fifteen feet high. A house on the beach ridge built in 1947 has been moved back from the beach twice in recent years because of the attrition along the (Page 53) shore. Hurricane Hazel of 1954 eroded large areas away in this region.

In an area 350 by 150 feet, sherds and flint fragments were picked up. Shells were virtually absent.

A shell midden now about four hundred to five hundred feet long. It does not extend inland more than fifty feet. Situated directly in the path of Hurricane Hazel (1954) the site suffered removal of twenty or more feet in a single night.

In the midst of the only cleared area for miles around is a slight promentory on which this shell midden occurs. It is roughly circular, perhaps seven hundred feet maximum diameter. It is generally one foot thick and varies in shell concentration, some spots having little soil content.

(Page 54) On the bank is a shell midden. Its exact size was difficult to ascertain when visited as it is heavily covered with a rank growth of grass and shrubs. It is at least three hundred feet in diameter. The shell is quite thick; in a few places it is nearly three feet. The site has been the location of a later modern residence, and chinaware fragments and other modern debris in mixed with the aboriginal.

A thin midden is scattered over two noses of land that project into the creek. The total area is not more than five hundred feet in diameter.

The shell is more than three feet thick in those spots where it remains protected as underneath the dwelling. In the field the shell is nearly three feet thick. Its maximuma length is eight hundred feet and perhaps it is three hundred feet wide.

Across from P 19 is a small shallow midden that is probably an outlier of P 19. It is about two hundred by (Page 55) fifty feet.

This shallow midden is interesting because it is presently nearly one-half mile from any active stream. The midden is about 250 feet in diameter.

The land stands as a low terrace five to ten feet above the water. There is no shell in this site but over an area 700 by 250 feet sherds are quite common.

On a point formed on the bank a typical shell midden is to be seen. It extends more than twelve hundred feet along the shore and extends inland 250 feet. Wave action is turning the end of the point inland and the shells concentrated at the point are water worn and concentrated into piles. Depth of midden is difficult to determine but seems to be not more than one foot generally.

The artifacts recovered at this site were picked up from a low (Page 56) sandy beach which is there because the sand has been dredged from Pamlico River and spread over an area about seven hundred by three hundred feet. Apparently, all the aboriginal materials found on the beach were dredged from the river. The sherds and projectile points from P 26 do not show water wear to any extent readily discernible. A likely conclusion is that the materials were transported by the river from close by; however, there presently is little current in the river here. It is a possibility that the materials were dredged from a site now flooded by a rising Pamlico River. Finally, it may be conjectured that the materials could have been left more or less in place by an expanding River that encroached on a site of low elevation and removed the soil and sand while transporting the heavier shell and artifacts but little.

There is a scattering of evidence of a village site. Only a few sherds were found but they extended over an area about four hundred by one hundred feet.

An extensive midden with little shell evident is spread over an 850-by 350-foot area. Pottery is about all that gives indication (Page 57) of this village site.

On the Pamlico River is a well-known high bluff overlooking the river. The earthworks of an old fort is the site of an aboriginal village. Potsherds and chert fragments cover centration that might be a large area but nowhere is there any concentration that might be called a midden.

A rather large village area. Potsherds were found over an area about 900 feet by 250 feet but shells were very scarce. At no place was any depth indicated but only a shallow humus was found.

A midden area extends along the bank of the creek for nearly one-fourth of a mile. It is probably related to the site across the creek.

(Page 58) On the high bank of the shore, seems ideally suited for an Indian village. However, only a few sherds, scattered over a wide area of several hundred feet diameter, indicates aboriginal occupation.

A prominent sand ridge, a village or dwelling area is indicated. A scattering of potsherds and flint chips extends along the ridge for 1000 feet. It is less than seventy five foot wide and seems entirely superficial.

High flat ground, the area is easily 1000 by 750 feet and potsherds, flint fragments, broken shell, and bone occur over it. Projectile points and even a gun flint were found. The potsherds are generally small since the area has been cultivated for many years. Nonetheless, this area is by far the most evident Indian village in the whole vicinity. As indicated later, the site in judged to be that called Secotan.

(Page 59) A thin scattering of sherds over an area about five hundred feet in diameter on the bank.

Historicall interesting and should be significant since it is here that Lawton located the Core Indians. Because of this the region was examined with great care. Although pieces of early European chinaware were found, only a single fabric-impressed sherd was picked up. Local residents insist that Indian materials have been found but nothing was evident.

Locality suggests ideal conditions for an aboriginal village but very little was found. Over an area 750 by 250 feet sherds are scattered. Shells are not to be seen. The source of the name of this landing is now probably in the river.

On the shore a small midden was found. This is now largely in the water and the few sherds found were picked from the shells at the water's edge.

(Page 60) On Indian Island an extensive midden of much shell and numerous potsherds has been known for some years. Word of our archeological activities stimulated interest in the area and zealous local visitors to Indian Island removed most of the surface pottery. As might be expected, the head of the island (upstream) is being eroded and it is at this point that the remains of the site occurs. Within the knowledge of present residents of the vicinity the island has been reduced one hundred yards at the upper end with resulting accretion at the lower end.

Neuse River

It would be desirable to determine how far south extend the dominant archeological manifestations of the Banks area. Hence, a cursory examination wasmade of a part of the Neuse River region, namely from New Bern eastward. By no means a complete survey was made, but certain conclusions may be reached because of this investigation. Four sites were located, but only two of these were middens.

On the bank of the Neuse the land slopes gradually to the water's edge. No extensive site was located but sherds were found in a localized area.

Along the bank of the Neuse, sherds may be found. The bluffs here are often more than twenty feet high and are similar in all respects to the south shore of (Page 61) the Pamlico River. The terrain presents ideal localities for aboriginal occupation but the south side of the Neuse is scarcely an archeological arena.

The north shore of the Neuse with its several embayments and tributary estuaries is quite a contrast to the south shore. Along the shore one of the better collections was made with large potsherds, shells, and even a whole pipe (see Figure 10) being recovered from the water. Not much midden proper was evident, yet the large number of sherds and other debris indicates a large population. No part of the shore at this point is more than three feet above tide so it would not be difficult to effect considerable erosion. It is judged that the site proper extends about one thousand feet along the shore.

The small middens comprising this site are significant beyond their size. The site is two patches of shell midden on both banks of a small embayment, each about two hundred feet in diameter, with a few sherds included. The importance of the site lies in its position which is virtually at sea level. The conclusion most readily reached is that (Page 62) the area has subsided somewhat since it was used by Indians. An alternate conclusion is that silting has nearly covered the shell midden. However, it may also be that the site was used only on rare occasions and not as a permanent dwelling. Whatever is the correct answer, it is certain that the site is not unique; aborigines did occupy coastal habitations in very exposed positions without fresh water or any forest cover.

Roanoke Island

Not a great deal more can be added to the known archaeology of Roanoke island. The investigations of Harrington at the site of Fort Raleigh are as complete and as revealing as was possible under the conditions characteristic of the soil at the site. Nothing could be found of Williams' "mounds" despite diligent search. Northwest Point, one of the areas suggested by previous investigators as the location of the village of the 1587 Indian residents of the island, is much eroded. Despite these facts, some generalizations may be made about Roanoke island's aboriginal occupants.

It is not likely that Roanoke could support a very large Indian population, anymore than it could support a small colony of white people. Certainly the whole northern end of the island gives varying evidences of Indian occupation. In the immediate vicinity of the Fort, the old humus zone--that is, the humus layer at the surface of the ground four hundred years ago--is near or at the present-day surface. Immediately northwest of the Fort dune sand has accumulated on this old level to depths of over ten feet. The recent construction of a replica of an Elizabethan sunken garden in this area west of the Fort has been fortunate from the archeological standpoint. By chance, the six-foot depth to which the garden is sunk carried just to the level of the old humus, that is, the ground level (Page 63) of 1587. Many small potsherds, a fragment Of an Indian pipe, and several fragments of china were found at this level.

For many years it has been known that the tip of land on Roanoke Island was the source of Indian artifacts. When representatives of the National Park Service visited here in 1938, twenty five potsherds were picked up along the shore. A number of white clay English pipe stems were found on this and subsequent investigations.

The specimens found at this locality have all been recovered from the water's edge. The most careful search failed to reveal any evidence of an Indian "village." Even in the immediate shore area no midden is to be found. Test pits dug at random along a mile of the shore border failed to reveal any aboriginal remains. The probable explanation for the situation is that the whole of the northern part must have been occupied by Indians but nowhere very intensively. Harrington found many potsherds at a few places in his excavations in and about the Fort, but nothing was found that suggested concentrated village midden. The reason potsherds may be found along the shore is that the shore line is constantly retreating here, and wave action redistributes the sand and silt with resulting concentration of larger particles such as sherds, shells, and stones. This situation was constantly met throughout much of this archeological survey. Often all that remains of a shell midden is a strandline of shell concentration. A few other objects than sherds were found. An "arrowshaft smoother" and a hammerstone seems hardly enough to justify calling the point a village site. It was nearly twenty years ago, 1938, that an archeological (Page 64) reconnaissance of the Point produced about twenty five potsherds. Several visits during 1954 and 1955 produced thirty nine sherds.

The area excavated for a sunken garden has been designated a "site", although it is clearly an extension of the aboriginal remains associated with the Fort. Sherds and pipe fragments attest the Indian occupation and a piece of iron and some fragments of china recovered from the old surface indicates European contact before the accumulation of the sand dune cover. Nothing of a cultural nature was found in the dune sand. Fortunately, Robert Cox, one of the maintenance crew of the Lost Colony who was assisting in the excavation of the garden, saved some potsherds and a pipe stem exposed during the operation.

Throughout most of its length, a dark band of humus is exposed in the bank. Much of this is the same layer exposed at the Fort and now in the bottom of the garden. Hence, the zone was searched for aboriginal materials. Only one sherd was found. At frequent intervals charcoal appeared. Also many charred roots were seen but this seemed to be the result of carbonization rather than fire. However, there did seem to be frequent evidence of some forest fire such as sighted by John White on his doleful return to Roanoke in 1590.

Copyright 2001
Carolina Algonkian Project