Other than pottery there is little material upon which cultural analysis may be based. None of the projectile points is particularly diagnostic of any known archeological horizon or of any ethnic group. However, it has been demonstrated repeatedly that the triangular points are most recent in the coastal region. Unfortunately, this is not a very profound finding, since triangular points are generally late in the whole eastern North America. Nevertheless, some pertinent facts may be learned froma the projectile points alone.

(Page 105) It would be inconsistent with statistical techniques to seriate so small a sample of projectile points as we have from the Banks. Despite this deficiency they may be equated with previous work. The most recent analysis is that of Holland (1955) for the Virginia area. He has seriated nearly three thousand projectile points and compared the result with the ceramic remains for the same state as described by Evans (1955).

Projectile point classification is never a very satisfying activity primarily because projectile points are rather sluggish in morphologic change through time. Only a rare type is culturally diagnostic. Those forms associated with eastern coastal regions are not unique but blend with many other areas. Only broad generalizations may be made. As pointed out above, only two major types of projectile points were found in the Carolina Sound, namely, the triangular form and the "lightly" corner-notched.

Triangular projectile points are clearly associated in greater frequency with the sites on which simple-stamped or shell-tempered pottery is dominant. As mentioned before, Mr. Stick's collection from B 5 is most numerous yet it was made over several years. Discounting its larger size, it still indicates that triangular points are culturally associated with shell-tempered pottery. H 6 had one point and P 35 had eight. P 35, Handy's Point site (Secotan), disclosed no shell-tempered pottery but 54 sherds (15 per cent of the total) were simple stamped, grit tempered.

Holland's seriation, as in any seriational study, was wholly arbitrary but logically began with non-ceramic sites, assuming these were the oldest. On this basis the small triangular projectile points (Page 106) were last (most recent) to become popular. This point was most often associated with grit- or sand-tempered pottery types; in Virginia, never with shell-tempered pottery. This startling and unfortunate fact is largely because of the defective data, to wit, the sites included in Evans report which were predominantly or wholly shell tempered had no flint artifact collections for comparison. Such sites as Kecoughtan (Kicotan) with a clear historic dating had hundreds of shell-tempered sherds but did not enter into the flint study. Hence, Holland's study is of limited value in the coastal Carolina area.

Bullen has made the clearest statements that are useful in this context (1950). Above it has been said that bison (buffalo) bones are totally absent from eastern North American archeological sites except within the historical horizon. Bullen (ibid., p. 3) notes that the Brumback Mound, excavated in 1894 by Gerard Fowke, falls within the historic period because, among other things, of the presence of bison bones triangular points. He finds further that triangular points are the major type at the historic Indian towns of Occaneechi and Keyanne in the Roanoke Valley and at Potawomeke on the Potomac. Such findings could be cited many times over, but it may be concluded that triangular projectile points within the coastal area are quite late prehistoric and continued into the historic period.

It is often assumed that man is a reasonable, logically-motivated animal; this is often assumed by anthropologists who should know better. Thus it seems reasonable and logical that Indians would use whatever rock substances were near at hand for making projectile points. Since sand comprises the Carolina coastal lands, flint, quartzite, granite, and other stones are absent. We must conclude these substances were brought in or traded. In this connection it may be pertinent that Barlowe made no (Page 107) mention of stone points for arrows but said that a fish tooth or piece of shell was used (Quinn, loc. cit., p. 112). Perhaps the conical antler-tine point (one found at H 1) would have been thought a "tooth" by Barlowe. Sharks' teeth are known to have been used by southeastern Indians but none is recorded in this area.

The other Projectile-point type which occurred more than once, namely, the corner notched, was a form recognized by Holland in Virginia. He named it "eared" and found but eleven specimens. Seriation would suggest an intermediate position for this type but nothing definite may be said from his data. Our records would imply that the form was associated with mixed sites; that is, sites with recent and earlier pottery types. Hence, an intermediate age for this type may be valid.

The remaining projectile points are either so generalized or unique that their use for diagnosis or comparison is limited.

It would seem, then, that the pottery remains may produce better comparative data. Our original statement about pottery differentiation may be a guide; only two significant wares are culturally important, namely, sand-grit tempered and shell-tempered. It is apparent that the temper differentiation is warranted because sites were found in which the pottery assemblage was all of one temper type. No site was found with the pottery having only one surface finish. Certainly surface finishes of several kinds are rather valueless for time markers. Fabric-impressed pottery occurs on sand-tempered, clay-(sherd-?) tempered, crystalline-grit tempered, and shell-tempered paste. It has as wide a geographic distribution as any pottery type in the New World. It certainly is the basic pottery type manufactured by (Page 108) historic Algonkian Indians from their northern climes to their southern limits. As a consequence nothing very diagnostic may be gleaned from fabric-impressed shards, except to strengthen the historical fact that we are within an Algonkian domain.

Perhaps similar remarks may be pertinent to the plain-surface finish, although this is a minority treatment or at least less popular than fabric impressing. Cord marking and net impressing occur on all sand-grit paste variations; both are absent from the shell-tempered ware. Hence, here are two problems to be explained away. Third problem: simple stamping, like fabric impressing, occurs on all temper types but is decidedly different in that it has a limited distribution in space and time. Finally, incising and punctating, although occurring very infrequently, may cast some light on the various pottery relationships.

Cord-marking surface treatment is about as extensive in distribution in eastern North America as fabric impressing. It may be demonstrated repeatedly that fabric impressing is somewhat older than cord marking throughout much of the southeast and certainly it had a different cultural history. Speaking of the eastern United States, Griffin says, "The introduction of pottery was probably first made in the north-central and the northeastern area where it is a heavy thick, coarse granular- tempered ware with plain surfaces, or, more often cord-marked on both the interior and exterior surfaces. In the south-central area the surfaces of the pottery are plain or cord-marked, but much more commonly are fabric impressed." (Griffin, 1952, p. 356). Fabric-impressed pottery is first in the Carolina Piedmont (Coe, 1952, p. 306). It is first in Kentucky (Webb and Haag, 1947, p. 27) and first in Mississippi followed by cord marking, a slowly southward-moving technique in late times (Page 109) (Haag, 1955, p. 24). A summarizing statement (Sears and Griffin, loc. cit.) regarding fabric-impressed pottery indicates the types are virtually the same throughout the southeast. As a matter of fact, there is something remarkable about this similarity. If differences in paste and temper (which would reflect local materials available) are discounted, the types of fabric, the method of application, and the shape of the vessel are nearly identical. A final fillip is that the kind of over-incising illustrated for the Banks area in Figure 4 is exactly duplicated as far away as north Alabama (Haag, 1942, p. 514).

Returning to cord marking, this technique never replaced fabric impressing in the Banks area and environs. In fact, cord marking probably died out as an accepted surface finish before fabric impressing, since the latter was present in quantity on late sites where shell was the dominant temper. Thus cord marking represents a disappearing art and those sites upon which it is found must have older occupation than those from which it is absent. Of course, this generalization holds true for those sites from which a sufficient number of sherds were recovered to assure random selecting would have given a fair sample of the total typological range. A "sufficient number" is judged to be about one hundred sherds. (See Table 2.) Coe and Lewis (1952) have been able to determine that cord marking disappeared from the Carolina Piedmont before 1700 A.D.

Cord marking, along with fabric impressing, is an integral trait of the pottery making of Algonkians. As mentioned above, it is about as widespread as fabric-marked pottery and nearly coterminous in distribution. Cord-marked pottery moved from a general northeastern (Page 110) point of origin, (although its ultimate source was almost certainly Asia and Europe) toward the south and west, and toward the lower Mississippi Valley. The time depth decreases along this path, so that (see above) its introduction in the lower Valley is much the more recent.

The origin of net impressing is not clear. Looking to adjacent areas where some sequences have been established, the Piedmont presents the clearest picture. This new pottery decoration was introduced during the development and expansion of the Uwharrie culture. Coe (loc. cit., p. 307) believes it began about 1200 A.D. and for the Piedmont became "The parent culture out of which the later ones emerged, specialized, and diverges." It became fully developed by 1500 A.D. Net impressing was at the "peak of popularity" among the Sara (Cheraw) in 1625-1675 but declined rapidly thereafter (Coe and Lewis, loc. cit.). Net impressing did not persist much longer in the Coastal area.

Coe and Lewis think net impressing survived longer in Virginia. Evans (1955, p. 136) points out that certain differences exist between the Dan River Focus (Sara Indians of North Carolina) and the Clarksville pottery which he describes, but the two (Dan River and Clarksville) are probably contemporaneous. A substantiation of this relationship seems to be that of sites reported by Evans as having Clarksville pottery predominating, all but two have the greatest percentage net impressed. This simply means that not impressing is a relatively recent (1500 to 1700) pottery technique, at least for the Piedmont and Virginia. Since there is nothing to suggest that it originated in the Carolina Sound, it must date from about that two-century span there.

Farther afield very little net impressing of the coastal-Carolina type may be found. It never becomes an important type anywhere in Georgia (Page 111) although it was spread through much of South Carolina. Net impressing thus becomes a very good horizon marker in coastal Carolina.

Simple-stamped pottery presents a far more difficult problem. It is not so complex a treatment that its occurrence everywhere may be taken as examples of the same thing or expressions by related peoples. Thus, within the earliest pottery ware known for the southeastern area, simple stamped is a numerous type (Haag, 1942, p. 523; Sears and Griffin, 1950). However, this first type has its impressions applied to the vessel by individual strokes of a single-edged implement, whereas later simple-stamped types were "paddled" with a tool having parallel grooves incised upon it. The latter types for the sand-grit and shell wares in the Carolina Banks are of this second category. Harrington (1948) has given support to the sixteenth century date for simple-stamped pottery at Fort Raleigh, so we have a great many centuries spanned by this technique.

The origin of simple stamping in the eastern United States is still unknown. There is a suggestion that the earlier fiber-tempered type may be the prototype for, at least, the simple stamping on sand- tempered pottery (Fairbanks, 1952, p. 286). In north Georgia this latter type, called Mossy-Oak Simple-Stamped, is associated with Early Woodland materials. Griffin and Sears (1950) say "This complex, then, of Plain, Simple stamped, and Fabric-Marked types constitute an Early Woodland complex which is moving into the far Southeast from the north." The site data in the Carolina Sound would lend support to this idea but with a major proviso, namely, that the complex was loose. Simple stamping came long after plain and fabric marked and from not (Page 112) very far north. The support for this statement may be seen in Table 2. Many sites exhibit no simple-stamped type and those sites in which it does appear in quantity have some other evidence for lateness. This would imply that P 5, P 13, P 14, P 16, A 1, and H 1 were late sites.

Adjacent areas give some evidence of the time of introduction of simple stamping. In Virginia it seems evident, although not very clearly, that two periods of simple stamping are represented in the coastal area--an earlier with sand and gravel temper, and a recent with shell temper (Evans, loc. cit., p. 95). In other parts of Virginia the situation is obscure. In the Piedmont simple-stamped pottery does not appear until 1700 when it was introduced by the Occaneechi (Coe, loc. cit., p. 310). Both shell and sand tempering were used (Coe, personal communication). Harrington found the potsherds from the Fort Raleigh village area to be predominately simple stamped and both sand and shell temper were used. The evidence clearly places simple stamping in the Carolina Sound well within a historic time frame.

Some of the unique sherds with incising and punctating, are important because each one represents a "trade" vessel. The Carolina coastal Indians did considerable trading with other aborigines and pottery was one of the trade items. Trading was never so extensive that such introduced objects would make up a large part of any cultural assemblage but generally the intruded material can be recognized as such. In the opening paragraphs of the pottery chapter, it was pointed out that a few sherds of a pottery type characterized by mica tempering were found at C 4. The practice of mica tempering may not be identified with any one group of people. The evidence available suggests the Piedmont as the source of the technique as well as the material.

(Page 113) The sherd illustrated in Figure 8, upper row middle, is a trade sherd found at site A 1. The sherd is from a vessel quite remineseent of those made by the Occaneechi around 1700 A.D. (Coe, ibid., Figure 166, G). Site A 1 has considerable shell-tempered pottery as well as clay-grit-tempered simple-stamped type; both of these items suggest a later date for A 1.

In the middle row of Figure 8 are punctated and incised sherds. Surely these are trade sherds but their point of origin is only a guess. A good guess, though, is that these represent other evidences of contact between the coastal aborgines and the Siouan peoples of the Piedmont. Coe (ibid., p. 310) notes that the Dan River Focus materials are characterized, in part, by punctate decoration borrowed from the Fort Ancient culture of the Ohio Valley. Also, these peoples made some use of shell for tempering pottery. The Dan River Focus is the archeological remains of "the Sara Indians in the Dan River area along the Virginia-North Carolina boundary between 1625 and 1675" (Coe and Lewis, 1952)

One other important item in the analysis and comparison of these data are the pipes. The obtuse-angle pottery pipe of cruder form is Algonkian and is nearly the same from New York to Carolina. In the former area it first appears in the Middlesex Focus, an archeological horizon in New York dated at about 300 A.D. This pipe type must have survived until historic times in the Carolina Sound but none has been reported by early chroniclers. It would represent the "older style" and would be found only in localities not closely associated with the newer f orm.

The "newer" form is that illustrated by the pipes from Hatteras (Page 114) Island. This is undoubtedly the same pipe as that found along the Atlantic Coast from Maryland to Carolina. Schmitt (1952, p. 63 and Figure 23H) shows that pipes of this fine dentate-stamped decoration are common in the Potomac area. They are equated with the Late Woodland period which in the Potomac and areas to the north would have a historic contact date. This casts more interesting light on the legend that Barlowe, et al, brought smoking to England. Quinn (loc. cit., p. 345) notes "That the colonists of 1585-6 -- and Sir Walter Raleigh -- introduced smoking into England is one of those myths which it appears almost impossible to kill. There is no doubt that Raleigh and the colonists alike did much to popularize what was already an established custom . . . . what the colonists apparently introduced was the smoking pipe used on Roanoke Island as a model for English pipe-makers." The Hatteras-Roanoke pipe was clearly of historic date.

The analysis of the material culture may be summarized as follows. The dominant type of flint projectile point in the Banks proper is the triangular point; elsewhere, stemmed points are found but nowhere are stone projectile points common.

The pottery represents a diversified assemblage but several facts may be established. Fabric-impressed ware is the hallmark of the Algonkian and it is ubiquitous in the Carolina Sound. It is of great time span yet where a relative date is possible it clearly has survived locally until historic usages. Cord-marked pottery may have survived until historic times on certain Pamlico and Albemarle sites but it did not continue in use on Hatteras Island sites. Net impressing arrived late in (Page 115) the Coastal region yet it did not constitute a pottery technique of Hatteras in the contact period. Net impressing may be a historic-contact-period time-marker for some Pamlico and Albemarle sites. Simple stamping is of two types, sand-tempered and shell-tempered. Although the former has some possible greater antiquity, the two kinds occur together at the Fort Raleigh site. Also, historic sites in nearby areas, such as Kicotan in Virginia, have a preponderance of simple-stamped shell-tempered pottery.

Unique sherds demonstrate that trade was carried on with interior peoples by the aboriginal Bankers. This trade was never great but probably was steady since many of the innovations of their neighbors eventually reached the coastal inhabitants.

All the evidence at hand--stone work, pottery, pipes, etc.--presents an overwhelming picture of recency. Everything else, such as shallow depth of middens, suggests no great time depth. However, occupation of the Carolina Sound probably has encompassed more than one thousand years and has a transition from hunting, shell-fish-eating, semi-nomadic groups to agriculturists of rather large population.

From the archeological remains we may differentiate two areas characterized by certain tangible remains. One area may be called the Banks since it includes the Banks proper from Hatteras to the Virginia line plus Roanoke Island. Among the features that distinguish it is shell-tempered pottery, most specifically simple-stamped. Also, triangular projectile points of chert are most frequently found in the Banks.. And, finally, the small, delicately-stamped pottery pipes are an important criterion of this area. These items may not seem abundant (Page 116) nor important but considered against the whole archeological assemblage they are both of these things.

The second area centers in the farther reaches of the sounds and it may be called the Sounds area. Here the features that are distinctive are cord-marked and net-impressed pottery, stemmed projectile points, and plain, obtuse-angle pipes.

These are not major cultural areas in the sense of Wissler's or Kroeber's usage, for the similarities overshadow the differences, and, of course, there are transitions and overlaps. Perhaps it would be more apt to call them subareas. No time differential can be established between the two subareas but certain sites within them may be reasoned to be older or younger. For example, on Hatteras Island Cape Creek site, H 1, is judged to be Croatoan, a village occupied in 1584-1590. A few thousand feet to the west of H 1 is H 4, the remains of a shell midden in water or directly at the edge. H 4 has 37 per cent sand-grit-tempered pottery (Table 2) while H 1 has exactly 10 per cent sand-grit tempered. H 4 is low enough to be washing away; H 1 is higher and still preserved. Similar conclusions may be reached for B 3, probably once under Nags Head dunes, and B 5 up on the Duck dunes. B 3 shows a considerable part of its pottery to be grit-tempered (Table 1), whereas all B 5 pottery was shell-tempered.

Carrying this idea into the Sounds culture subarea, we can discern from Table 2 that A 12 is a fairly pure representation. Also it is, in part, earlier than many within the group. Table 5 reveals no shell-tempered sherds and only two sand-grit simple-stamped examples well up toward the top of the four-foot midden. Several other sites, P 2, P 24, P 30, and N 3 were early within the subarea. Of these only N 3 had any (Page 117) shell-tempered pottery and only P 2 had simple-stamped type.

Between the typical sites of the two subareas are sites that have strong representations of all the important pottery types. A 1, at the south tip of Currituck Peninsula is such a site. A test pit there showed complete admixture of the types. Three significant sites are P 13, P 14, and P 19. They are close together and probably in the area where Aquascogoc must have been located. These sites show good representation of nearly all types including simple stamping on both grit temper and shell temper.

P 16 and P 35 are two revealing sites also. They are both in the Bath Creek area far up Pamlico River. Neither site had any shell-tempered pottery but both had a fair percentage of simple-stamped type. They seem to belong to the transitional sites.

The major sites listed in Table 2 and some others considered important enough may be categorized thus: Banks subarea includes, H 1, H 6, R 1, R 2, B 5, A 9, and C 3; Sounds subarea includes P 1, P 2, P 5, P 6, P 30, A 2, A 12, and N 3; Transitional are P 13, P 14, P 16, P 18, P 19, P 24, P 25, P 35, C 4, A 1, H 4, H 5, and B 3.