An Archaeological Overview and Assessment of the
Main Unit, Petersburg National Battlefield, Virginia

Chapter Three:  Paleoenvironment and Prehistoric Occupation (Part A)

Paleoenvironmental and Prehistoric Overview

The Pleistocene Epoch witnessed a series of cold periods and associated "ice ages," the most recent of which terminated approximately 14,000 to 12,000 years ago.  One of the most dramatic effects of these "ice ages" was the lowering of ocean levels worldwide as sea water was frozen and trapped in glaciers and continental ice sheets.  Milliman and Emery (1968) argued on the basis of 80 radiocarbon determinations from samples taken along the Atlantic continental shelf that sea levels 30,000 to 35,000 years ago were close to those at present.  Sea levels dropped subsequently as much as 130 meters during the final glaciation c.16,000 years ago.  Along the Atlantic coast, ocean beaches lay at the edge of the modern continental shelf, perhaps 100 kilometers east of the current Virginia coastline.  Belknap and Kraft (1977) questioned the maximum depth of sea level drop, but agreed with the overall pattern.

Climatic patterns have changed on regional and continental scales during the Holocene Epoch (for some only an interglacial), which began at the end of the Pleistocene.  Sea levels have continued to rise as a result of the release of water from melting ice sheets.  As the sea level rose, it began to transgress, or cover, the western land mass of the former coastal plain, now the modern submerged Atlantic continental shelf.  The Holocene marine transgression, or sea level rise, began c.14,000 years ago and proceeded rapidly until c.7000 years ago (Milliman and Emery 1968; Kraft et al. 1983).

The implications of such dynamic changes for any paleoenvironmental reconstruction of the Chesapeake Bay region in general and the lower Appomattox River area in particular are profound.  The Pleistocene Susquehanna River, which underlies the modern Chesapeake, became an estuary as sea levels rose.  Climatic changes resulted in a succession of vegetation types moving northward, while the coastline and associated marine and eustatic environments were approaching from the east.  As temperatures warmed and the climate alternated between dry and moister periods during the Holocene, open grassy environments were replaced by boreal evergreen forests and then by deciduous forests.  As the coastline approached, local environments shifted in the study area from inland riverine forests to a forested plateau overlooking a tidal river.  A paleoenvironmental reconstruction must therefore consider both the generally northward-moving vegetational patterns arising from the regional climatic shifts, and westward-moving coastal geomorphological changes associated with rising sea levels.

The occupancy of prehistoric peoples within these dynamic and mobile environments is the focus of this chapter.  Paleoindian occupation of the southeastern United States may date c.12,000 BP (before present), and thus may predate the earliest sites in the northeastern United States (Dent 1995:102).  This chronology, however, ignores the suggested earlier dates of 18,000-20,000 BP for a human presence at Meadowcroft Shelter in southwestern Pennsylvania (Adovasio et al. 1978; Adovasio 1993), or the suggested pre-Clovis occupation provisionally dated c.15,000 BP at Cactus Hill along the Nottoway River in southeastern Virginia (Marshall 2001:1731-1732).

During the period 12,000-10,000 BP, the western portion of the Chesapeake region was covered with a boreal forest composed primarily of pine and birch which shifted, as temperatures warmed, to pine and oak (Dent 1995:131).  Similar vegetation cover extended throughout much of the region, although the presence of favorable microenvironments arising due to topography, solar exposure, and surface water (ponds, lakes, and rivers) exerted a considerable influence on prehistoric subsistence and adaptations.

Paleoindian occupants would have co-inhabited the region with a rich fauna.  The mammoth, oriented to more open habitats, disappeared from the area prior to the arrival of humans.  A few forest mastodon may have been contemporaries of the earliest Paleoindians.  Therefore, the image of early humans as hunters of megafauna requires substantial revision throughout the eastern United States (Meltzer 1993; Custer 1994:332-333).  Deer and probably caribou would have been common inhabitants of the early Holocene forests, as would a range of smaller fauna.  The proximity of stream and riverine habitats would have supported aquatic resources, both animal and plant in nature.

Numerous isolated fluted projectile points have been recovered throughout the region, but larger concentrations of Paleoindian-period artifactsCat times much largerChave also been found.  No isolated fluted points or larger loci have been identified within the Main Unit, but such loci do occur in the general vicinity.  The Williamson site near Cattail Creek in eastern Dinwiddie County is a major resource extraction and habitation locus.  The site is located along the extreme western margin of the inner coastal plain, near the piedmont.  Decades of excavation have yielded 175 fluted points and more than 2000 end and side scrapers, as well as other retouched tool forms.  Factors that undoubtedly led to repeated visitations included local sources of chert for tools, water, a southern exposure, and topographic shelter from winds (McCary 1951; Dent 1995:107-109; Hill 1997a,b).

The Conover site lies slightly to the east of Williamson, also on the inner coastal plain.  Approximately 1100 artifacts have been recovered, including nine fluted points and preformed points.  These artifacts are made, for the most part, on the cherts found in the Williamson vicinity.  The Point-of-Rocks site is positioned in Chesterfield County near a swampy area of the Appomattox River between City Point and Petersburg.  This smaller locus has yielded five fluted point fragments and one reworked complete point (McAvoy 1979; Dent 1995:111-113).

A relatively high number of Paleoindian sites south of the James River is attributed to the presence of a more diverse ecosystem.  Dent has argued that the inner coastal plain —particularly along its western boundary with the piedmont—represented an ecotone setting with abundant and diverse ecological resources.  He suggests that these inner coastal plain loci may represent colder season habitations, with warmer season sites located on the outer coastal plain and the banks of the ancestral Susquehanna River, now submerged beneath the Chesapeake (1995:133-139).  It should be noted, however, that other settlement models exist.  Gardner (1974, 1977) argued that Paleoindian groups were often tethered to sources of high-quality raw materials, with elements of the settlement system conditioned by the availability of such materials.  Group movements and catchment areas in the resulting "cyclical" pattern were centered upon specific raw material sources.  Although Gardner's model emerged from studies at the Flint Run complex of sites in the Shenandoah Valley, a comparable argue could be made for the Williamson site.

Prehistoric occupation during the Archaic period is traditionally divided chronologically as follows (Dent 1995:168, 173, 178):

These chronological divisions are general ones, and subject to some exceptions.  The Early Archaic period has been combined by Gardner and others (Custer 1989, 1994) with the Paleoindian period into a broad late Pleistocene-early Holocene adaptational continuum.  Custer also favored combining the late Archaic with the subsequent early and middle Woodland phases.  Dent (1995:188-190) argued for a post-Paleoindian adaptational break that extended into the late Archaic until c.4200 BP.  The final phases of the late Archaic (4200-3000 BP) represents for Dent an "intensification" of resource extraction efforts.  Various scholars therefore perceive a cultural change prior to or during the late Archaic.

A paleoenvironmental reconstruction for the Chesapeake area of the Middle Atlantic region has been extrapolated from the Indian Creek V site in Prince George's County, Maryland, where a pollen core was extracted from an abandoned stream channel of the creek (LeeDecker and Holt 1991:Table 1).  Seven pollen zones were defined:

Environmental changes undoubtedly exerted an influence upon the cultural adaptations, the material remnants of which are reflected in the archaeological record.  As the flooded channel of the ancestral Susquehanna became increasingly estuarine with sea level rise, the Appomattox and James rivers had become subestuaries by 6000 BP, with at least slightly saline waters up to the fall lines.  The Chesapeake had formed as a coastal bay by 3000 BP (Dent 1995:191).  The following vegetational sequence has been defined by Dent (1995:190) from pollen core data such as those from Indian Creek V:

Vegetation thus changed from a transitional forest association to a temperate deciduous forest, as the environmental setting changed from a freshwater riverine location to a variably saline tidal river.

The initial portion of the Archaic, which to Dent encompasses the Early, Middle, and initial Late Archaic, witnessed heavier occupation of the inner coastal plain and piedmont, probably because the full extent of estuarine resources had not been developed along the bay and coastal rivers (Dent 1995:197).  Dent does acknowledge, however, that the advancing marine transgression may have covered some of these earlier sites on the outer coastal plain.

Custer (1994:337) has argued that a shift from oak-hemlock to oak-hickory forests (Atlantic to Sub-Boreal phases) c.5000 BP may be correlated with the cultural changes perceived as the Late Archaic.  The oak-hickory forests had a higher carrying capacity that resulted in an expanded number of habitable areas.  An intensification of resource utilization is reflected in various manners during the Late and Terminal Archaic (Dent 1995:188, 200-208).  The technological component reflects an expanded use of ground stone tools, the appearance of steatite (soapstone) vessels and, in riverine and coastal areas, fishing implements in the form of notched cobble netsinkers.  The presence of storage pit features is noted and shellfish exploitation occurs, although not on the scale seen later in the Woodland.

Johnson (Rudolph et al. 1996:24-26) has argued that the depiction of the Sub-Boreal phase as the warmest and driest period of the Holocene may be inaccurate.  He suggested that the Xerothermic maximum may have actually occurred c.6500-6000 BP within the generally moist Atlantic phase.  By proposing the boundary of the Middle-Late Archaic at 6000 BP, he therefore placed less importance that did Custer on environmental changes associated with the expansion of oak-hickory forests.

The Woodland period may be subdivided chronologically into the following phases (Dent 1995:224):

As mentioned above, Custer (1989) has challenged this traditional tripartite classification, choosing instead to redesignate the Late Archaic and Early-Middle Woodland as Woodland I and the Late Woodland as Woodland II.

A general classification of resource extraction activities and associated broadening of the subsistence base is indicated during the Early and Middle Woodland phases.  Subsistence from the Early Woodland occupation sought to utilize a fuller range of resources such as shellfish from the expanding estuary.  Fishing and nut gathering were important activities during the Early phase, as well as the possible exploitation of seed plants.  The importance of sedentary occupation is indicated in the appearance of ceramics.  A certain technological "homogenization" has been noted during the Middle phase, with a more limited range of projectile point and ceramic types (Dent 1995:228-235, 268).  The fall line emerged as a "boundary" between Middle Woodland groups in the coastal plain to the east and piedmont to the west (McLearen 1992; Stewart 1992), and this boundary was maintained through the end of the prehistoric period.

Nevertheless, evidence of interregional contact within eastern North America is also indicated.  The appearance in the Chesapeake region of Adena-related materials from Ohio serves to reenforce the identification of these contacts, and suggests greater social complexity during the Early and Middle Woodland (Dent 1995:242-243; 268-269).

During the Late Woodland, the emergence of interacting regional groups is indicated by the presence of ceramic types in specific core areas with wider areas of distribution.  Technological changes include smaller triangular points evidently associated with bow and arrow propulsion.  Increased evidence of sedentism is reflected in estuarine shell middens, refuse middens, storage pits, and house patterns.  Late Woodland sites in the vicinity of the Main Unit that contain house patterns include Jordan's Point, Flowerdew Hundred, and Hatch along the James River, as well as sites along tributaries of the Appomattox River (Dent 1995:245-249; Gregory 1980; Mouer et al. 1992; McLearen and Mouer 1994).