William F. Stanyard
TRC Garrow Associates
The prehistory of Georgia begins sometime before 9000 b.c. and ends with the de Soto entrada of a.d. 1540. During this time, major changes took place in the societies that occupied this portion of the Southeast—changes which are documented in the archaeological record (see Summary Table page). This document summarizes the technological, economic, social, and political processes that unfolded over the millennia since humans first inhabited the region.
The Paleoindian period marks the beginning of human occupation in the New World. Exactly when the first human populations permanently settled the western hemisphere is uncertain; most Americanist archaeologists believe it was sometime between 20,000 and 14,000 years ago, during the last stages of the Pleistocene glaciation. The earliest securely dated Paleoindian site is in Monte Verde, Chile, where dates as early as ca. 11,800 b.c. have been obtained (Dillehay 1989). The end of the Paleoindian period coincides with the Pleistocene/Holocene transition and in most areas of the Southeast is estimated to be ca. 8000 b.c.
By 8000 b.c. environmental conditions were approaching those that exist today. North of 33° N,[1] “patchy” enclaves of xeric boreal forest/parkland vegetational communities were gradually replaced by widespread stands of mesic oak-hickory forests. This forest type lasted until large-scale Afro/Euro-American agriculture and construction severely modified the landscape. South of that parallel, the oak-hickory canopy was present much earlier (Delcourt and Delcourt 1985). The project area lies just north of the postulated vegetational interface (33°40’ N) and, given the coarse-grained nature of this reconstruction, it is not possible from the available data to determine whether the oak-hickory regime was present in the area during most or all of the Paleoindian period, or whether there was a change from the boreal forest/parkland regime to oak-hickory during that time.
The Paleoindian lithic tool kit was based on a highly refined flake and blade technology. Examples of Paleoindian lithic tool types include unspecialized flake tools, formal side and end scrapers, gravers, denticulates, specialized hafted unifacial knives, large bifacial knives, and specialized lanceolate projectile points, which were sometimes “fluted.” The best known of these is the Clovis point, the earliest recognized projectile point type in the western hemisphere (dating 9800–9000 b.c.). Clovis variants have been found from Canada to the southern tip of South America.
Formal variation in projectile point morphology began to emerge in regions of the Southeast by about 9000 b.c., probably due to restricted movement and the formation of loosely defined social networks and habitual use areas (Anderson 1995; Anderson et al. 1992). These new forms include the Cumberland, Suwannee, Simpson, Beaver Lake, and Quad types (Anderson et al. 1990; Justice 1987:17–43; Milanich and Fairbanks 1980).
A significant wood, bone, and antler technology was present as well. These organic items do not preserve well in the acidic soils that cover much of the Southeast, and they are rarely found in such contexts. However, at sites where they have been preserved, primarily in Florida, it is clear that organic media such as wood, bone, and antler were very important. These materials were manufactured into projectile points, foreshafts, leisters, awls, and needles, to name just a few tool categories (Milanich and Fairbanks 1980: Figures 3, 5, and 6).
Original views of the Paleoindian subsistence economy were based on observations from a series of sites in the western United States where Paleoindian artifacts, particularly large, lanceolate, “fluted” points, were recovered in direct association with the remains of several species of now extinct Pleistocene megafauna. Initial interpretations of Paleoindian subsistence suggested that these early inhabitants focused primarily on hunting such large mammals as mammoth, mastodon, bison, ground sloth, giant armadillo, tapir, horse, wild pig, and caribou. Resources such as arboreal seed and nut crops as well as small mammals, birds, and fish were, until recently, assumed to have been minor dietary constituents.
Because of the striking similarity in Paleoindian technological organization that pervaded most regions of the western hemisphere until ca. 8500 b.c., the large game–oriented subsistence model devised from the western United States evidence was initially assumed to have applied to all Paleoindian economic systems, including those associated with groups in Georgia. However, archaeologists working in Georgia have yet to document a clear association between Paleoindian tools and the remains of displaced and extinct animal species known to have been present in the state as late as 11,000–10,200 b.p.—mastodon, bison, giant ground sloth, and giant armadillo, for example (Holman 1985:569–570).
Over the past 15 years there has been a reevaluation of Paleoindian subsistence, particularly for eastern North America, based upon data from sites such as the Meadowcroft Rockshelter in southwestern Pennsylvania. Cushman’s (1982:207–220) analysis of the Paleoindian occupation at Meadowcroft Rockshelter suggests that the occupants were geared toward the type of “broad spectrum” resource utilization traditionally associated with the subsequent Archaic period. Her examination of the botanical remains indicates that a variety of leafy plants, seeds, nuts, and berries (Cushman 1982:207–220) were important dietary components.
Broad-based Paleoindian subsistence is also indicated by evidence from Florida. At Little Salt Spring, an important underwater site in Sarasota County, Florida, a variety of smaller mammals, fish, plants, and reptiles (including a now extinct form of giant land tortoise) have been shown to be constituents of the Paleoindian diet in that region (Clausen et al. 1979).
There is very little evidence of resource exploitation in the littoral by Paleoindian peoples living in the Southeast. This very likely is because of site obfuscation and destruction caused by coastal submergence during the Holocene, and not because the resources these ecozones contained were not utilized (e.g., Dunbar et al. 1988, 1991).
In summary, new perspectives on Paleoindian subsistence economy emphasize the utilization of a broader spectrum of ecotones and resources and deemphasize the degree to which Paleoindians relied on large-game hunting for sustenance.
In the Eastern Woodlands, the majority of Paleoindian sites consist largely of diffuse lithic scatters at open locations, with more intensive occupations in rockshelter or cave settings. No conclusive evidence of permanent structures or long-term encampments has been located for this time period in the Southeast. The majority of the Paleoindian data recovered in Georgia to date is derived from surface scatters of projectile points and a small assortment of chipped stone implements collected from settings in which the depositional integrity has been compromised. However, a limited amount of data has been recovered from intact contexts (Anderson and Schuldenrein 1985; Elliott and Doyon 1981; Gresham et al. 1985; Kelly 1938; O’Steen et al. 1983; O’Steen et al. 1986).
Several models of early Paleoindian settlement patterning have been advanced in the past quarter-century (see Anderson et al. 1992 for an overview). Some are concerned with Paleoindians in general (Anderson 1990a; Kelly and Todd 1988; Martin 1973), and others with regional trends (Anderson 1995; Gardner 1983; Morse and Morse 1983). Most are mechanistic models that portray specific economic strategies as primary reasons for how Paleoindians settled upon and utilized the landscape. Each is slightly different in its focus, with primacy placed on one of three major influences: (1) the need to maintain access to prominent, high-quality raw material sources (e.g., Gardner 1983); (2) a preference for exploiting specific habitual use zones and staging areas (e.g., Anderson 1995); or (3) a nomadic or seminomadic existence dictated to a large degree by the movements and availability of large game (e.g., Kelly and Todd 1988).
An attempt to review and assess each model is impractical in this context; however, there is a general consensus among archaeologists involved in Paleoindian research regarding Paleoindian settlement. Groups were probably each comprised of four or five extended families and counted 25–50 individuals. Marriage was almost certainly exogamous and residence was likely extralocal. This would have assured that primary social groups remained small enough to remain economically sustainable but linked with a larger, interactive social network that provided information, cooperation, and mates of suitable kin distance.
Primary social groups very likely met at predetermined locations with other groups at specific times of the year to cooperate in large-scale food acquisition (nut harvesting, fishing, shellfish gathering, etc.) and/or lithic resource extraction, as well as to exchange information, renew or create alliances, fulfill social obligations, find mates, and perform rituals. For most of the year, however, primary groups appear to have dispersed into loosely defined habitual use areas. They probably exploited a wide variety of economic resources, moving often to take advantage of seasonal resources. It is also possible that they periodically established logistical base camps and used them as staging areas for special activity forays.
The end of the Paleoindian period (ca. 8000 b.c.) is associated with the end of the Wisconsin Ice Age and the onslaught of new environmental conditions, which influenced how humans organized their society and coped with the environmental and social pressures that came about during the climatic transition. New settlement and subsistence patterns were established and regional technological innovations were developed. These trends are associated with the subsequent Archaic culture period.
The transition from Paleoindian to Archaic is loosely defined; in the Southeast the chronological interface ranges from ca. 8000 to 6500 b.c. In Georgia, the transition has been arbitrarily designated as 8000 b.c. In addition to rapid changes in environmental conditions that were nearing completion by 8000 b.c. (Delcourt and Delcourt 1985), and the changes in utilitarian technology that were developed to cope with those changes, population demography and diversity in social organization distinguish the Archaic experience. A tripartite scheme dividing the Archaic period into Early, Middle, and Late subperiods is traditionally used to demarcate some of the important developments of this time. It should be emphasized, however, that these subdivisions are heuristic devices; changes were more gradual and non-uniform across the Southeast than a discussion with these limitations intimates.
Early Archaic (ca. 8000–6000 b.c.). Tool assemblages associated with the Early Archaic period are similar to those of the preceding Paleoindian period, although a variety of groundstone tools first appear at this time. Notched and/or stemmed hafted bifaces replace lanceolate forms by 8000 b.c. in the Southeast. Big Sandy, Palmer-Kirk series, Kirk Corner Notched, Kirk Stemmed, and several bifurcate styles are the Early Archaic types known in the project area. Wear patterns suggest that these tools were used for activities such as killing, butchering, and skinning game, as well as woodworking.
The Early Archaic lifeway is represented by social, settlement, and subsistence strategies designed to take advantage of the biotic diversity of the early Holocene environment, and also to cope with movement restrictions placed upon some Early Archaic populations because of increased population. Environmental conditions were approaching those that the first Europeans encountered in the sixteenth century. Hardwood primary forests and extensive palustrine swamps provided large and small game as well as a variety of plants for medicine, subsistence, clothing, and shelter. Rivers were used as travel corridors and provided fresh water, fish, and shellfish. The only areas of low productivity would have been the pine stands that began to emerge in the uplands by about 6000 b.c. (Delcourt and Delcourt 1985).
As population apparently increased dramatically, the social landscape became more complex. Several models of Early Archaic social organization have been proposed for the region (Anderson et al. 1992: Part II; Anderson and Hanson 1988); again, this is not the proper context to explore and assess the merits of each. In general, it is hypothesized that Early Archaic societies in Georgia and the Carolinas were organized into band-sized communities (population 25–50) whose main territory surrounded a segment of a major river (the Ocmulgee, for example). These bands are postulated to have been organized into larger “macrobands” that gathered on special occasions for community food harvesting, rituals, and the exchange of mates and information. These activities probably took place at or near the heads of rivers close to the Fall Line, or at the mouth of the rivers on the coast. The similarity in certain tool forms throughout and across drainages—projectile points, for example—and the apparent movement of raw materials over long distances support this argument.
Early Archaic settlement patterns are not well understood, but two types of settlements have been especially noted: small, short-term “camps” and large, densely occupied areas that appear to have been base camps or congregation sites (see above). As before, high-quality cherts were accessible and were the raw material of choice for stone tools. Also, specific point types, such as Palmer-Kirk series and bifurcate styles, were widely distributed across the Southeast and the Eastern Woodlands. This suggests that territories were large and/or that the exchange of information, ideas, and material culture took place frequently and over large distances.
Middle Archaic (ca. 6000–3000 b.c.). As in the final stages of the Early Archaic, climax hardwood forests were established in the lowlands, and upland pine stands became mature and fairly widespread.
Diagnostic bifaces dated to this period include the Stanly and Morrow Mountain types, as well as Benton and Guilford-like forms (MALA) that have not been formally typed. Unremarkable quartz ovate hafted bifaces are common as well. Although all of these are known to occur in Georgia, the Morrow Mountain styles are the most frequently encountered diagnostic hafted bifaces in north and north-central Georgia. Morrow Mountain hafted bifaces also occur in Middle Archaic contexts on the coastal plain and coast, but in much smaller numbers.
The Middle Archaic period tool kit was, for the most part, expedient and manufactured from locally available raw materials. Quartz, which is ubiquitous in northern Georgia, was the preferred source of lithic raw material in the region during this period. Chert tools or debitage are not frequently encountered in Middle Archaic contexts in northern Georgia. Compared to chert, quartz is difficult to work, yields a dull edge, and requires frequent resharpening. Chert was probably not used to any great extent because of limited access to or knowledge of source areas. On the coastal plain and coast, locally available chert was the preferred raw material for stone tool manufacture.
Piedmont Middle Archaic sites have been described as small, randomly distributed occupations exhibiting very little intersite technological variability. Local raw materials were used almost exclusively, and the vast majority of tools were technologically expedient (Blanton and Sassaman 1989; Sassaman 1993a). In terms of social organization, small hunting and gathering bands of 25–50 people probably still formed the primary social and economic units. Residences were moved frequently, subsistence was generalized, and social groups were small, mobile, and likely coresidential. Long-term investments and social obligations were probably kept to a minimum, insuring that there were very few restrictions on group movement or fissioning (Sassaman 1993b).
By contrast, large-scale tool production and intensive occupation characterize many Middle Archaic habitations in the coastal plain, especially in the latter half of the period (Sassaman 1988). This is likely due to the patchy distribution of both lithic and organic resources in that region, as opposed to the relatively homogeneous distribution of resources that characterized the Georgia piedmont (Sassaman et al. 1990).
Subsistence data is scarce, but it is assumed that a variety of interior floral and faunal resources were exploited on both a general (e.g., white-tailed deer) and seasonal (e.g., nuts, fish, and migratory waterfowl) basis. It is probable that coastal and riverine resources—marine shellfish, freshwater shellfish, and anadromous fish, for example—were exploited to some degree, but their economic importance is unknown due to the lack of Middle Archaic components that can be unequivocally associated with these types of remains. This void can be partially attributed to coastal submergence and rising sea level, which has inundated previously exposed coastline and obfuscated the importance of littoral resources in this and earlier eras (Brooks et al. 1990).
Late Archaic (ca. 3000–1000 b.c.). The hafted biface most commonly associated with the Late Archaic period in Georgia is the Savannah River point. These point types are often very large (12+ cm in length is not uncommon) and exhibit a straight stem, straight base, and triangular blade.
Other Late Archaic varieties are known by various names, such as Appalachian Stemmed, Elora, Kiokee Creek, Ledbetter, Limestone, Otarre, and Paris Island (Bullen and Greene 1970; Cambron and Hulse 1983; Chapman 1981; Coe 1964; Elliott 1994; Harwood 1973; Keel 1976; Sassaman 1985; Whatley 1985). Except for the Ledbetter hafted biface, which appears to have had a specialized function—it exhibits a heavily reworked, asymmetrical blade—these latter type names are more a product of parochial terminology than actual morphological differences; they all are characterized by triangular blades, straight or slightly contracting stems, and straight bases.
The earliest ceramics in the region were tempered with fiber. According to radiocarbon evidence obtained from Rabbit Mount, a Late Archaic shell midden along the southern portion of the Savannah River in Allendale County, South Carolina, this ceramic technology may have been introduced as early as 2500 b.c. (4465±95 b.p.). This date, and another of 4450±150 b.p., were obtained from wood charcoal recovered from excavation levels containing fiber-tempered sherds (Stoltman 1966).
The earliest ceramic-bearing components on the Georgia coast date to approximately 2200 b.c. (Sassaman 1993b). These components are also the oldest known along the current Georgia coastline, so this evidence does not necessarily demonstrate that coastal groups did not produce and use pottery prior to 2200 b.c. Sea level rise may have inundated earlier ceramic-bearing assemblages; ecological restrictions may also have prevented humans from occupying the present-day coastline before 2200 b.c. (Sassaman 1993b:19).
The Late Archaic ceramic sequence has been refined over the years and a detailed chronology for both the interior and coastal zone has been developed. The coastal sequence is known as the St. Simons phase, which is a term first used by Holder (1938) to describe the Late Archaic ceramics he recovered during excavations on St. Simons Island in the 1930s. Based on his extensive research on Late Archaic ceramics from various sites along the Georgia and South Carolina coasts, DePratter (1979) identified a time-transgressive trend in surface decoration techniques that led him to divide St. Simons into two subphases, St. Simons I and St. Simons II.
In DePratter’s scheme, St. Simons I dates to ca. 2200–1700 b.c. This phase is characterized by the production of fiber-tempered pottery with plain surfaces (DePratter 1979:114). St. Simons II dates to ca. 1700–1000 b.c. The ceramics produced in this era are also fiber-tempered and exhibit plain, punctated, incised, incised and punctated, and grooved surface designs. Vessel form is limited to simple bowls with round or flattened bases. Rims are straight or slightly incurving, and the lip is rounded or flattened (DePratter 1979:114).
The term Stallings Culture was introduced by Stoltman (1974) to describe the material culture associated with Late Archaic populations residing in the central Savannah River region. Stallings I was defined as a preceramic phase that dated to the earliest part of the period (ca. 3000–2500 b.c.). The hallmark of Stallings II (ca. 2500–1700 b.c.) was the production of fiber-tempered pottery with plain surfaces. Stallings III lasted from 1700 b.c. until the end of the period (ca. 1000 b.c.), and it is characterized by fiber-tempered vessels with plain, punctated, incised, incised and punctated, and grooved surface treatments. Vessel form is limited to the simple bowls discussed above.
The close similarity between St. Simons I and II and Stallings II and III is evident. Sassaman’s research has led him to refine the St. Simons and Stallings ceramic series and incorporate them into a single chronological sequence. This new chronology is provisionally defined as Group I, Group II, and Group III (Sassaman 1993b:102–110). Group I assemblages date to approximately 2500–1800 b.c. in the interior and 2200–1800 b.c. on the coast. Most of the pottery manufactured in this era exhibits plain surfaces. When designs are present they are usually simple and limited to a single simple-stamped, incised, or punctated motif. Vessels are simple bowls with thickened and flanged lips.
Group II (ca. 1800–1400 b.c.) is defined by a marked increase in decorated vessels. Incising, punctations, and grooving are common surface treatments; simple stamping is almost absent, however. Many vessels exhibit multiple design motifs. Wares with thickened and flanged lips occur less often; by the end of the phase this technology is no longer utilized. Vessel form is restricted to the simple bowl.
Group III dates to between ca. 1400 and 1000 b.c. The ceramics produced in that era exhibit plain, incised, punctated, and simple-stamped designs. Plain ceramics were the most common; the relative frequency of decorated wares is much lower than for Group II. Multiple design motifs are not evident. Plain vessels in Group III can be distinguished from Group I wares by the absence of thickened and flanged lips. Simple bowls continue to be the exclusive vessel form.
The most intensively occupied Late Archaic site yet discovered in Georgia is on Stallings Island, located in the Savannnah River in Columbia County (Bullen and Greene 1970; Claflin 1931; Crusoe and DePratter 1976; Fairbanks 1942; Jones 1873). One type of bone tool found at Stallings Island is the bone “pin.” These objects are intricately decorated and highly prized by artifact collectors. Unfortunately, they were “mined” at the site until recent measures were taken to prevent unauthorized access to the site. The mining has devastated the site; large “potholes” and mining trenches have destroyed much of its integrity.
This unfortunate circumstance notwithstanding, a great deal has been learned from professional excavations at Stallings Island. Large quantities of projectile points, drills, grooved axes, perforated soapstone slabs, and other formal lithic, bone, and antler tools have been discovered. Plain and punctated fiber-tempered ceramics, which bear the type name Stallings Island, have also been recovered.
The earliest Late Archaic levels at Stallings Island have been dated to between 2700 and 2450 b.c. (Williams 1968). These basal levels lacked ceramics but, among many other tool types, contained “classic” Savannah River projectile points (Coe 1964). Subsequent excavations elsewhere in the region have shown that these large “classic” Savannah River points are associated with the incipient use of fiber-tempered ceramics (Elliott 1994:370). Large Savannah River bifaces were often manufactured from metavolcanic rock; some assemblages—from the Mill Branch, Toliver, and Chase sites, for example—are dominated by points of this material (Ledbetter 1991, 1994; Stanyard and Stoops 1995).
This particular form of Late Archaic technology is associated with a suite of traits that are spatially and chronologically specific (ca. 2200–1600 b.c.). In the Savannah River region, it was manifest between ca. 2200 and 1850 b.c., and is referred to as the Mill Branch phase (Elliott 1994; Elliott et al. 1994; Ledbetter 1994; Stanyard 1997; Stanyard and Stoops 1995). As the Stallings influence took hold in the Savannah River drainage by 1850 b.c., Mill Branch people moved out of the area and permanently settled into the surrounding region. Mill Branch culture persisted until ca. 1500 b.c. in those places. In order to accentuate the geographical and chronological separation between the two “episodes” of the Mill Branch phase, a recent proposal has suggested that the latter expression be designated the Black Shoals phase (Stanyard 2000)[2].
Though ceramics have been dated as early as 2500 b.c. in the Southeast (Stoltman 1966; Sassaman 1993b), they do not appear at Stallings Island until about 1730 b.c. Projectile point styles associated with the ceramic levels at Stallings Island are smaller than Savannah River point types and tend to have slightly contracting, rather than straight, stems (Bullen and Greene 1970). Beginning about this time, the use of ceramics intensified in the region. Elliott (1994) refers to this technological expression of the Late Archaic period as the Lovers Lane phase and frames it between approximately 1800 and 1350 b.c.
Curiously, soapstone vessels, a hallmark of the Late Archaic in the interior of Georgia, are almost absent in the archeological record at Stallings Island specifically (n = 1) (Elliott et al. 1994) and in the central Savannah River valley in general. This is despite the existence of several nearby sources of soapstone that were used to obtain raw material for perforated slabs, gorgets, and bannerstones.
Most Late Archaic groups surrounding the central Savannah River valley, on the other hand, preferred soapstone for bowls and other containers. Steatite bowl fragments are common at Late Archaic sites in these areas, and fiber-tempered ceramics are uncommon (Sassaman 1991, 1993b).
The
discrepancy between sites that contain ceramics and those that contain
soapstone vessels may not reflect an absence of technological knowledge
concerning ceramics, but actions that are politically, economically, and
socially motivated instead (Sassaman 1991, 1993b). New radiocarbon data
obtained from soot adhering to soapstone sherds found in the region supports
this contention; no dates precede the known or suspected date for the local
adoption of pottery (Ken Sassaman, personal communication 1996).
The Late Archaic period witnessed several significant changes that anticipated the cultural developments of the following Woodland period. Information gathered from hundreds of Late Archaic period sites in northern and central Georgia presents a fairly clear picture of demography and settlement. Seasonal single-household occupations and special activity camps related to those occupations dotted the uplands throughout north-central and northeast Georgia, as well as the western Carolinas, while large and intensively occupied special-purpose aggregation and multiseasonal village sites are associated with the central Savannah River basin.
Late Archaic architecture is not well understood, for only a few examples have been investigated in northern Georgia. Excavations at 9WR4, in Warren County, Georgia, discovered a Late Archaic pit house measuring approximately 4 x 5 m (Ledbetter 1991:200). It was subrectangular in plan and approximately 35 cm deep (Ledbetter 1991:200). Large corner posts and a few wall posts defined the perimeter. A large hearth area was discovered in the eastern portion of the structure. It is interpreted as a hearth and earth oven that may have been partitioned (Ledbetter 1991:201); three “caches” of debitage surrounded the hearth area.
Six structures associated with the Late Archaic occupation of the Lovers Lane site have been documented (Elliott et al. 1994). All were subrectangular or oval in plan; only one structure (Structure 6) was determined to be a pit house similar to the one at 9WR4. The smallest structure measured 5 x 8 m and the two largest 8 x 8 m. None of the structures contained discernible hearths. Pit features used as storage or discard pits for quartz debitage were found in the vicinity of Structure 4, but the association is suspect (Elliott et al. 1994:335).
In terms of subsistence, a wide variety of large and small mammals, reptiles (including sea turtle), birds, and amphibians have been recovered in Late Archaic contexts. Shellfish were very important to Late Archaic populations that inhabited and/or exploited the coast and major drainage systems, as evidenced by the large shell middens at Stallings Island (Claflin 1931), Bilbo (Williams 1968), St. Simons Island (Holder 1938), and elsewhere. The bone fishhooks and foreshafts recovered at these and other sites indicate that fishing was also important.
A broad spectrum of plant materials is assumed to have been used for sustenance, medicine, fabric, and construction. There is no conclusive evidence of horticulture in Late Archaic societies in Georgia. It is possible that the growth of certain useful opportunistic plants, such as weeds containing starchy seeds (e.g., Chenopodium sp.), and possibly cucurbits (Cucurbita sp.), was encouraged by clearing overstory and not disturbing established communities of these plant types.
The end of the Archaic period and advent of the Woodland era is an arbitrary demarcation created by archaeologists. It is a consensus that recognizes the widespread adoption of an improved ceramic technology by 1000 b.c.
The improvement in ceramic technology that became widely available by 1000 b.c. in the Southeast greatly altered food storage and preparation capabilities, though it did not have an immediate effect on subsistence. Throughout most of the Woodland period, subsistence strategies were a continuation of earlier hunter-fisher-gatherer ways; cultigens did not begin to play an important role until approximately a.d. 900.
In Georgia, the nature of Woodland peoples’ ideological and nonsubsistence-related economic systems are more accessible to modern researchers than those of earlier peoples because they involved activities, architecture, and artifacts that are more visible in the archaeological record. For example, large mounds associated with mortuary, ceremonial, and status-related domestic domains first appear by about a.d. 1. Also, large quantities of magico-religious and prestige goods manufactured from such durable media as stone and unsmelted metal were deposited in and around these mounds beginning at approximately this time. The Woodland period also witnessed intensified participation in long-distance trade and exchange in exotic materials such as copper, mica, obsidian, and marine shell.
The introduction of very small triangular projectile points (<1–3 cm in length) around a.d. 600 suggests that bow and arrow technology was adopted in the southeastern United States at about this time.
Ceramics became more refined, and regional differentiation of wares, particularly with respect to temper, paste, and surface decoration, became manifest during the period. Woodland cultures in the interior of northern Georgia are often discussed and categorized by reference to established ceramic typologies and related developments. Common Woodland ceramic types include Dunlap Fabric Impressed; Cartersville Simple Stamped; Cartersville Check Stamped; Swift Creek Complicated Stamped; Swift Creek Plain; Napier Complicated Stamped; Woodstock Plain; Woodstock Incised, Woodstock Complicated Stamped; Vining Plain; and Vining Simple Stamped.
Diagnostic projectile point styles attributable to Woodland developments north of the Fall Line in Georgia include small-stemmed specimens, large and small triangular types, and miscellaneous notched specimens.
The Woodland period, like the preceding Archaic, is divided into three subperiods—Early, Middle, and Late—based upon major demarcations in general social patterns. As with the Archaic period, it should be emphasized that changes were more gradual and nonuniform across the Southeast than the discussion intimates.
Early Woodland (ca. 1000–300 b.c.). Early Woodland occupations are thought to reflect a more or less unchanged continuation of Late Archaic lifeways, except for the widespread adoption of a much improved ceramic technology. Dunlap Fabric Impressed pottery, which is associated most closely with the Early Woodland, is tempered with sand or crushed quartz, and the vessel exteriors usually are decorated entirely with impressions of fabric or basketry (Caldwell 1957:166). The most common vessel form is a large, conoidal-based jar. Toward the end of the Early Woodland period, another ceramic type, Cartersville Check Stamped, was manufactured and used along with the earlier Dunlap Fabric Impressed wares (Caldwell 1957:287). Cartersville Check Stamped, as the name implies, is characterized by a checked design stamped on the exterior of the vessels. Vessel types include large jars and, for the first time, smaller bowls. These vessels often had small podal supports on their bases that are termed tetrapods. Cartersville Simple Stamped ceramics began to be produced at about the same time as check stamped vessels, though in the Early Woodland they were a minority ware. Vessel morphology and technology are identical to those of check stamped vessels.
A diagnostic tool that first appeared in the Early Woodland is the triangular hafted biface. This tool form was popular throughout the Southeast until the Contact period. Early Woodland specimens are generally large and sometimes have incurvate bases or small “ears.” These latter two types are known as Yadkin and Eared Yadkin, respectively. Small, stemmed hafted bifaces were also produced during this era. Although various names have been given to these types, such as Coosa, New Market, and Otarre, the terms are basically parochial in nature and do not reflect significant technological differences.
Soapstone, a popular raw material in the Late Archaic period, was reduced to a very minor constituent of the overall Early Woodland artifact assemblage. It was used to make utilitarian items such as line weights, gorgets, and works of decorative or ritualistic art.
Villages were built primarily in the floodplains of large to medium-sized rivers. Archaeologically, they occur as isolated entities (Bowen 1989; Wood and Ledbetter 1990) or in concentrations along river stretches (Stanyard and Pietak 1997). Hunting, fishing, seasonal foraging (especially in the fall), and lithic reduction were conducted in the uplands, on levees, and at river shoals (Stanyard and Baker 1992; Stanyard and Pietak 1997). Burial mounds, a hallmark of Middle and Late Woodland mortuary practices, appear to be lacking in the Early Woodland.
A variety of nut crops, especially acorns, were a major subsistence preference during the Early Woodland in north Georgia. Nut processing and roasting pits are much more common at Early Woodland sites than at any other type of site in the region (Bowen 1989; Wood and Ledbetter 1990). The remainder of the subsistence base encompassed a broad spectrum of species acquired by hunting, fishing, and gathering.
Along the Georgia coast and on the coastal plain, the onset of the Early Woodland period is recognized archaeologically by the appearance of Refuge ceramics. This pottery complex was defined by Waring (1968) based on data obtained from the Refuge site, which is on the southern South Carolina coast. Waring described four types of surface decorations associated with the Refuge ceramic series: Refuge Punctate, Refuge Incised, Refuge Simple Stamped, and Refuge Dentate. These surface treatments are still used as diagnostic criteria for identifying Early Woodland occupations in the region.
Simple stamping was a technique developed in the Late Archaic but was used as a decorative motif until the end of the Middle Woodland. As a result, the sand/grit-tempered wares associated with the Early Woodland (Refuge) and Middle Woodland (Deptford) periods are difficult to distinguish. Waring (1968:200) noted that Refuge simple stamping tended to be haphazard and that the lips of these vessels were sometimes notched. Deptford wares, by contrast, primarily exhibit parallel or crossed designs that were applied with more control.
Recent stratigraphic evidence obtained from 38AK157, which is on the Aiken Plateau at the Savannah River Site in Aiken County, South Carolina, concurs with Waring’s supposition that Refuge simple stamping can be distinguished from Deptford on the basis of the decorating technique (Anderson in Sassaman et al. 1990). Based on stratigraphic analysis, Anderson (Sassaman et al. 1990) observed that there were proportionally more sherds with V-shaped grooves than U-shaped grooves. In addition, parallel and evenly crossed designs were present in proportionally larger numbers in the upper levels, while sloppy designs were more prevalent in the lower levels. It is suggested that the earlier, more haphazard designs were applied with a dowel or stick, while the more controlled designs were applied with a carved paddle (Anderson in Sassaman et al. 1990). Since the check-stamped pottery associated with Deptford was definitely created with a carved paddle, Sassaman postulates that simple stamping characterized by the even, U-shaped groove is associated with Deptford, while the haphazard V-shaped design is associated with Refuge ceramics. Although the data from 38AK157 is useful when dealing with large assemblages, distinguishing between Refuge and Deptford simple-stamped designs is an imprecise exercise when sample sizes are small.
Refuge ceramics from sites on the Georgia coastal plain and coast are usually grit-tempered and generally have a very sandy paste. Grog tempering occurs in a minority of wares found in the South Carolina coastal plain (Anderson 1982), but is dominant on the Refuge series of the Santee River (Espenshade and Brockington 1989). The predominant vessel form is a hemispherical bowl with a rounded base. Deep, straight-sided jars were also produced but in lesser numbers. Rims are incurving or straight; the lips are rounded or squared and are occasionally decorated (DePratter 1979). Punctations and incising sometimes occur on vessel interiors (Anderson in Sassaman et al. 1990)
With the exception of ceramics, very little is known about Refuge material culture. Diagnostic lithics associated with the Refuge phase consist of small, stemmed hafted bifaces that are similar to the varieties manufactured in the later part of the Late Archaic period. Lithics occur in low frequencies at Refuge sites, which may indicate that the lithic sources in the interior were not easily accessible (Hanson and DePratter 1985). Another reason lithics are not abundant, one that may or may not be directly related to the availability of lithic material, is the ready accessibility of shell. Shell and bone tools are commonly associated with components of this age (Lepionka 1983).
The Refuge phase has been divided into subphases based on temporal differences in the popularity of ceramic surface design types. DePratter (1979) describes three subphases, Refuge I, Refuge II, and Refuge III. In his scheme, Refuge I dates to ca. 1100–1000 b.c. and is defined by punctate and incised wares. Dentate stamping appears approximately 1000–900 b.c., and the appearance of that design demarcates Refuge II. According to DePratter, Refuge III (ca. 900–400 b.c.) is defined by the manufacture of linear check and check-stamped wares. Plain and simple-stamped pottery was manufactured throughout all three subphases.
Anderson (Sassaman et al. 1990) argues that, given the general lack of radiocarbon dates, DePratter’s chronology is too refined. In addition, there is no conclusive evidence that linear check and check-stamped designs were in use as early as 900 b.c. Sassaman (1993c:190) suggests that only two subphases are recognizable within the Refuge ceramic complex. Refuge I dates to ca. 1000–800 b.c. and is defined by Refuge Punctate and Refuge Dentate designs. Refuge II is characterized by the absence of punctate and dentate surface designs and by the emergence of plain and simple-stamped surfaces as the primary design types. According to this chronology, Refuge II occurred between approximately 800 and 600 b.c.
The social transformations at the end of the Late Archaic on the coastal plain resulted in population decentralization (Sassaman 1991, 1993c; Stanyard 1997). Small groups disengaged from their social obligations to the larger community and created dispersed year-round settlements. People that produced Refuge ceramics settled the fall zone uplands, the lower coastal plain interior, and the coast. Upland and interior sites tend to be on well-drained ridges, while coastal sites are often situated near marshes in riverine and estuarine settings (DePratter 1976). The upland and interior sites are usually small and lack evidence of intensive utilization (Hanson and DePratter 1985; Sassaman 1993c). The coastal sites usually contain large middens and appear to have been utilized more intensively and extensively (Hanson and DePratter 1985). This pattern suggests that coastal and lower coastal plain sites functioned as permanent or semipermanent villages, while interior sites perhaps served as single-household seasonal base camps.
Subsistence was generalized, and the resource base was very similar to that of the Late Archaic period, with the possible exception of shellfish. White-tailed deer, bear, a variety of small mammal species, reptiles, freshwater fish, marine fish, anadromous fish, and mollusks have been recovered from Refuge contexts (Marrinan 1975; Lepionka 1983; Hanson and DePratter 1985). Although shellfish were harvested during the Refuge phase, their degree of dietary importance appears to have been dramatically lower than it was in the Late Archaic. This may be due to lower productivity caused by sea level fluctuations (DePratter 1977). It is also possible that the larger shell midden sites are currently inundated, as sea level has risen about three meters since the early Woodland period (Hanson and DePratter 1985).