The late archaic to early woodland transition in Northwest Georgia : evidence for terminal archaic (ca. 1100-600 B.C.) period occupation in the region / prepared for Georgia Department of Transportation, Office of Environment/Location ... by Jerald Ledbetter, Lisa O'Steen, Scott Jones, with contributions by David S. Leigh ; Southeastern Archeological Services, Inc

The Late Archaic to Early Woodland Transition in Northwest Georgia
Evidence for Terminal Archaic (ca. 1100 - 600 B.C.) Period Occupation in the Region
Georgia Department of Transportation. Office of Environment/Location Atlanta, Georgia
OCCASIONAL PAPERS IN CULTURAL RESOURCE MANAGEMENT # 14

The Late Archaic to Early Woodland Transition in Northwest Georgia
Evidence for Terminal Archaic (ca. 1100 - 600 B.C.) Period Occupation in the Region
Prepared for: Georgia Department of Transportation
Office of Environment/Location 3993 Aviation Circle Atlanta, GA 30336
By: Jerald Ledbetter Lisa D. O'Steen
Scott Jones with contributions by David S. Leigh
April 1, 2009
SOUTHEASTERN ARCHEOLOGICAL SERVICES, INC.
P.O. Box 8086 Athens, Georgia 30603
Georgia Department of Transportation. Office of Environment/Location Atlanta, Georgia
OCCASIONAL PAPERS IN CULTURAL RESOURCE MANAGEMENT # 14

Georgia Department of Transportation Occasional Papers in Cultural Resource Management
The Georgia Department of Transportation's (DOT) Occasional Papers in Cultural Resource Management series consists of archaeological and historical research sponsored by the Georgia DOT. These reports have been produced by Georgia DOT in-house cultural resource staff and by consultants under contract with the Georgia DOT. Reports within the series present information ranging from cultural resource contextual themes to specific data associated with historic properties considered eligible for the National Register of Historic Places that would be affected by transportation projects.
Each report addresses research questions and the conclusions and interpretations contained therein reflect the theoretical orientation, background, and assorted biases of the authors. Each manuscript has been prepared as a result of a project with Georgia DOT. The reports are distributed by the Office of Environment/Location, Georgia Department of Transportation.
For a copy of any or all of the reports, please indicate the specific report; there may be a minimal charge for copying if the report is out of print.
Georgia Department of Transportation Occasional Papers in Cultural Resource Management
No. 1 - Archaeological Investigations at 9CK(DOT)7, Cherokee County, Georgia; by William R. Bowen (1982).
No. 2 - Cagle Site Report: Archaic and Early Woodland Period Manifestations in the North Georgia Piedmont; by Morgan R. Crook, Jr. (1984).
No. 3 - Lowe Site Report: A Contribution to Archaeology of the Georgia Coastal Plain; by Morgan R. Crook, Jr. (1987).
No. 4 - Rush: An Early Woodland Period Site in Northwest Georgia; by W. Dean Wood and R. Jerald Ledbetter (1990).
No. 5 - A Few Visits in Prehistory: Data Recovery at 9RH18, Randolph County, Georgia; by Christopher T. Espenshade (1993).
No. 6 - The Pig Pen Site: Archeological Investigations at 9RI158, Richmond County, Georgia; by R. Jerald Ledbetter (1988).
No. 7 - Data Recovery at Lovers Lane, Phinizy Swamp and the Old Dike Sites, Bobby Jones Expressway Extension Corridor, Augusta, Georgia; by Daniel T. Elliott, R. Jerald Ledbetter, and Elizabeth A. Gordon (1994).
No. 8 - The Victory Drive Site, 9ME50, Muscogee County, Georgia; by R. Jerald Ledbetter (1997).
No. 9 - The Bull Creek Site, 9ME1, Muscogee County, Georgia; by R. Jerald Ledbetter (1997).
No. 10 - An Archeological Survey of the Shoulderbone Tract, Hancock County, Georgia; by Thomas J. Pluckhahn (1997).
No. 11 - Prehistory of the Stuckey Tract, Bleckley County, Georgia; by Scot Keith (2006).
No. 12 - Archeological and Historical Investigations at Battery Hamilton (9CH953), Chatham County, Georgia; by Chad O. Braley (2003).
No. 13 - The Early Vegetable Tanning Industry in Georgia: Archaeological Testing at the Clinton Tannery (09JO282), Jones County, Georgia; by Daphne L. Owens and Daniel E. Battle (2008).
No. 14 - The Late Archaic to Early Woodland Transition in Northwest Georgia: Evidence for Terminal Archaic (ca. 1100 - 600 B.C.) Period Occupation in the Region; by Jerald Ledbetter, Lisa D. O'Steen, and Scott Jones (2009).

Preface
The following study was inspired by the findings of two recent testing projects, where, under the banner of scientific methodology, observation led to question and question led to investigation. As in many situations, a "discontinuity" led a researcher to look deeper into a particular generally accepted scheme, and to wonder why and how things changed or diverted from that scheme, instead of staying the same. This same exercise has happened again and again in archaeology as a discipline so that we have reduced ourselves to finer and finer temporal and geographic units. This study in particular examines a yet undefined slice of time and space in Northwest Georgia and provides thoughts on the rate and reason of change from the better known cultures of the Late Archaic, as defined, to those of the Early Woodland, through a unit now described as the Terminal Archaic.
These goals are achieved by the authors through an exhaustive study of previous research in and around the time and space defined in this report. They also focus on several research questions relating to changes in material culture and general society and show how previous and current archaeological investigations offer evidence to support a transitional environment. The authors present much fine detail in their discussion; however, their "big picture" mentality further aids the archaeologist in putting together all the small, seemingly disjointed and diverse, material clues to make a coherent, natural narrative of the cultural landscape during a time of technological and social change. This study thus demonstrates the importance of cultural context in analyzing and interpreting archaeological findings.
The Georgia Department of Transportation is happy to publish The Late Archaic to Early Woodland Transition in Northwest Georgia: Evidence for Terminal Archaic (ca. 1100 - 600 B.C.) Period Occupation in the Region as Report Number 14 in its Occasional Papers in Cultural Resource Management series.
Pamela A. Johnson Archaeologist Georgia Department of Transportation
i

ii

Abstract
The development of a cultural context document to examine the time of transition between the Late Archaic to the Early Woodland periods in Northwest Georgia was inspired by the findings of two recent testing projects conducted in Murray County (Gresham 2002) and Gordon County (Patton and Gresham 2003). These projects provided evidence of occupations dating to a time that immediately preceded the introduction, or at least the widespread use, of Early Woodland fabric-marked pottery in the region. While sites of this time period are relatively well studied elsewhere in the Southeast and Midwest, they have generally been unrecognized in Northwest Georgia. For this study, the authors arbitrarily selected a time range of 1,100 to 600 B.C. for this period of transition. The authors feel that this length of time generally represents the interval between the demise of the major Late Archaic cultural spheres of influence in the Southeast (Poverty Point and Stallings Island) and widespread use of Early Woodland pottery in Northwest Georgia. This date range corresponds with that proposed for the period of Late Archaic to Early Woodland transition elsewhere in the Eastern United States (Kidder 2006:195-198). The transition from the Late Archaic to Early Woodland is examined within the context of the essentially aceramic Terminal Archaic period as originally defined at the Westmoreland-Barber site which is located near the northwestern border of Georgia in southeastern Tennessee (Faulkner and Graham 1966a). The geographical extent of Northwest Georgia study area encompasses the Ridge and Valley, the Cumberland Plateau, and the western portion of the Piedmont Physiographic Provinces. The primary study area for Northwest Georgia consists of a ten county area; however, pertinent information from important sites in adjoining counties and in nearby areas of Alabama and Tennessee is included since modern historical and political boundaries are irrelevant to the period addressed. As a means of interpreting the limited data base currently available from Northwest Georgia, comparative data from other regions in Eastern North America will also be utilized in this study.
Interest in the cultural changes that developed between the end of the Late Archaic period and the Early Woodland period in Northwest Georgia began more than half a century ago with Joseph Caldwell's work in Allatoona Reservoir. Caldwell recognized a "discontinuity" within the material culture on sites of the earliest pottery-producing people of the region (Kellog period) and their Archaic predecessors (Caldwell 1957, 1958). Caldwell felt that no culturally intermediate sites had been located in the Allatoona area and, if any such sites are found, they will be rare. Caldwell was probably referring to sites with fiber-tempered pottery (his Stallings period) which he found to be "hardly represented at all in the Allatoona Reservoir" (Caldwell 1950:6). Based on his research in Buford Reservoir on the upper Chattahoochee River, Caldwell (1953:10) referred to this time of discontinuity as "a hiatus of unknown duration." More recently, researchers have argued that Caldwell's idea of a "temporal and cultural hiatus" is not supported by current archaeological data (Bowen 1989:215, Crook 1984:55, Rotenstein n.d.:7). However, others have maintained that some of the data that have been used to refute Caldwell's assertion are flawed (Wood and Ledbetter 1990:19). Caldwell suggested that the people of the Kellog focus moved into Piedmont Georgia and the Carolinas from the north, possibly from eastern Tennessee (Caldwell 1958:23). Bowen (1989) has reexamined much of Caldwell's work and revised some of Caldwell's ideas. Bowen (1989) concluded that a smaller geographical area centered on Allatoona Reservoir represents the Kellog "heartland" representing the area of most intensive Early Woodland occupation in Northwest Georgia.
iii

To date, most of the work relating to the Late Archaic and Early Woodland periods in Northwest Georgia has been conducted in areas along the western edge of the Piedmont province. Relatively little information is available from most of the Ridge and Valley province, an area that is expected to share characteristics with cultural traditions defined for physiographic regions to the north and west. This lack of data for the Late Archaic period in the Ridge and Valley province was previously recognized by Stanyard (2003). Our study of the Late Archaic to Early Woodland transition in Northwest Georgia relies in large measure on information extracted from previous investigations. Previous investigations in Northwest Georgia have been conducted by archaeologists of varied backgrounds who have perceived and interpreted the Late Archaic and Early Woodland cultures of the region in an equally varied fashion. In some instances, these researchers interpreted certain aspects of these sites as transitional or intermediate occupations, and thus provide information that pertains directly to this study. Most frequently these transitional occupations are identified by projectile point types that have been defined elsewhere in the Southeast as dating to the very end of the Late Archaic or the very early part of the Early Woodland periods. Various researchers have referred to these transitional occupations as terminal Late Archaic, Terminal Archaic, or transitional Late Archaic/Early Woodland. These admittedly vague designations result from the fact that fiber-tempered pottery is extremely rare, if not absent, in most of Northwest Georgia. Also, traditional Early Woodland pottery types do not appear until relatively late in the region (ca. 700 - 600 B.C.). In parts of the lower Southeast and in the lower Tennessee Valley where fibertempered pottery does occur, occupations of this time period are usually referred to as Late Archaic or Gulf Formational. The Terminal Archaic period in Northwest Georgia should be viewed as essentially aceramic.
Assemblages consisting of barbed and expanded stemmed projectile points and soapstone vessels are referred to as Terminal Archaic in eastern Tennessee (Faulkner and Graham 1966a; Childress 1999); they have been more formally defined as Wade phase in Middle Tennessee (Keel 1978:160), and Cogswell phase in Kentucky (Ison 1986:212). Very similar projectile points occur in the Terminal Archaic Prairie Lake phase as far away as the American Bottom in Illinois (Emerson 1980). Each of these phases are well dated within the interval of approximately 1100 to 600 B.C. (uncalibrated date range). Previous research shows that similar point types occur most frequently on sites in the northwestern corner of Georgia. This suggests a relationship with cultures living in the Cumberland Plateau region and areas to the north and northwest. The aceramic Terminal Archaic component at the Lick Creek site in Gordon County has been dated within the range of these phases (Benson et al. 2008).
Because so few transitional sites have been previously recognized in Northwest Georgia, data from sites with Late Archaic and Early Woodland occupations have been assessed to better understand this period of transition. Available data suggest that Northwest Georgia was an area of cultural and ethnic diversity during the latter part of the Late Archaic period through the Early Woodland period. This may be explained in large measure by the unique geographic position of Northwest Georgia with respect to the Ridge and Valley province and specifically to the Great Valley (Coosa Valley). It appears that this period of transition in Northwest Georgia was a time of cultural constriction that is visible in the distinctive appearance of the material culture of sites within relatively small geographical areas in the region. Many of the cultural traits that distinguish this transition, and the material culture of these sites in the Southeast, are related to changes in subsistence. The culmination of generations of incipient plant husbandry stabilized the plant resource base, allowing for increased sedentism and more geographically constricted ethnic identities that are reflected in the innovative technologies that accompanied these changes.
iv

Acknowledgements
This project was proposed and funded by the Georgia Department of Transportation (GDOT). GDOT archaeologist Pamela Johnson was the primary contact person for the project and provided enthusiastic support in addition to a thorough review of the document. GDOT archaeologists Rowe Bown and Jim Pomfret also provided valuable assistance. The contract was conducted through the firm of Greenhorne & O'Mara, Inc., and Thomas Tkacs of their staff served as project manager.
For this regional synthesis, the authors relied on the assistance of a number of archaeologists in Georgia and adjoining states. The time and effort provided by our colleagues from Georgia and nearby states is greatly appreciated. Archaeologists from consulting firms, academic institutions, research organizations and museums, and state and federal agencies provided information necessary for the completion of the document. The information provided by these individuals included published and unpublished reports and manuscripts, information related to ongoing research, and comments and criticisms of a number of concepts. An effort was made to examine as many sources of information as possible but some documents could not be located and we must apologize for those omissions.
Rowe Bowen and Dan Elliott provided the peer review for the document and proposed a number of suggestions which helped to clarify the subject matter. In addition to the previously mentioned individuals, appreciation is expressed to the many Georgia archaeologists who shared their research and offered numerous suggestions. Particular thanks is expressed to Steve Webb, Betsy Shirk, Jim Langford, Dennis Blanton, Susan Fishman-Armstrong, Pat Garrow, Chris Espenshade, Jeannine Windham, Stefan Brannan, Dean Wood, Eric Sipes, Scot Keith, Wes Mattox , Bill Stanyard, Dave Davis, Paul Webb, Frankie Snow, Keith Stephenson, and Jeffrey Glover. Mark Williams, director of the Georgia Archaeological Site Files, provided full access to the files and collections and offered a number of helpful suggestions. Additional information was provided by avocational archaeologists Robert Beasley, Steven Beasley, Tommy Jones, and Bill Frazier.
Appreciation is expressed to a number of archaeologists from the adjoining states of Alabama, Tennessee, North Carolina, and Florida. Keith Little, Steven Meredith, and David Anderson provided particularly valuable support by sharing the results of their research and by commenting on specific aspects of this study. Information, comments, and other forms of support were provided by Eugene Futato, Bill Turner, Rebecca Turley Ridley, Van King, Teresa Paglione, Duke Beaseley, Matthew Gage, Ted Wells, Gerald Kline, Jason O'Donoughue, Scott Meeks, Jamie Waggoner, Ken Sassaman, Deborah Mullins, Kris Gramillion, Steve Watts, Cecil Ison and Johnny Faulkner.
John Whatley and Winston Baker provided valuable assistance with respect to projectile point typologies of the region and their knowledge of the local projectile point sequence was relied upon extensively for this document. Our colleagues at Southeastern Archaeological Services, in particular Tom Gresham and Chad Braley, should be acknowledged for their continued support and encouragement during the course of the project. Appreciation is also expressed to Theresa Groover for editing the final report.
v

vi

Table of Contents
Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i
Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
List of Figures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
List of Tables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Chapter 1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Purpose and Goals of Project. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Definition and Geographic Extent of Study Area.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Description of Late Archaic and Early Woodland Transition in the Southeast. . . . . . . . . . . . . . . . . . . 4 The Late Archaic Period. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 The Early Woodland Period. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 The Terminal Archaic Period. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Introduction of Research Issues to be Addressed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Authorship. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Chapter 2. Environmental Setting of Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Physiography and Geology of Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Environment of Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Modern Environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Environmental Conditions During Late Archaic to Early Woodland Transition. . . . . . . . . . 14 Prehistoric Lithic Raw Material Resources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Ridge and Valley Chert Sources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Chert Resources in the Great Valley. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Chert Resources of the Armuchee Ridges.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Chert Resources of the Chickamauga Valley. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Chert Resources of the Cumberland Plateau. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Other Lithic Raw Material Resources.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Quartz. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Quartzite. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Greenstone. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Slate and Shale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Abrasives.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Soapstone. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Chapter 3. A Review of Previous Archaeological Investigations in Northwest Georgia and Adjoining Areas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Early Work in Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Robert Wauchope's North Georgia Survey. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Large Scale Reservoir Salvage Projects of the 1940s through early 1970s. . . . . . . . . . . . . . . . . . . . . 32 The Smithsonian's River Basin Project at Allatoona Reservoir. . . . . . . . . . . . . . . . . . . . . . . 33 Joseph Caldwell's Allatoona Survey and Excavations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Caldwell's Late Archaic Stamp Creek Focus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
vii

Caldwell's Early Woodland Kellog Focus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Carl Miller's Excavations in Allatoona Reservoir. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Implications of Caldwell's and Miller's Work in Allatoona Reservoir

for the Current Study.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Buford Reservoir Project. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

Weiss Reservoir Project. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

Carters Dam Reservoir Project. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

The Georgia Historical Commission's Excavations in Northwest Georgia. . . . . . . . . . . . . . . . . . . . . 73

Mahan Site and Other Sites Investigated by Amateur John Wear. . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

Projects Conducted in the Etowah Valley in Recent Decades. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

Garfield (9BR57) and Wooly's Bend (9BR59) Sites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

Lewis Sam's Bridge Site (WGC 409). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

9CK(DOT)7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

Cagle Site (9CK123).. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

The Rush Site (9FL164)

104

Malbone Power Line Site (9BR891).. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

Route 411 Sewer Line Project (East of Rome). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

Yellow Creek Reservoir Project, Cherokee and Dawson Counties. . . . . . . . . . . . . . . . . . . 124

9PA102 (Soapstone Bowl Quarry). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

Hardin Bridge Site (9BR34). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

Projects Conducted in the Lower Oostanaula/Coosa River Valley in Recent Decades. . . . . . . . . . . 136

The Colonial Pipeline Project at Selman's Bend (9FL56 and 9FL61). . . . . . . . . . . . . . . . . 137

Pumpkin Pile Site 9PO27 (Big Cedar Watershed Project). . . . . . . . . . . . . . . . . . . . . . . . . . 139

Southwest Rome Bypass Project (Sites 9FL203 and 9FL206.. . . . . . . . . . . . . . . . . . . . . . . 160

Plant Hammond Area Sites (9FL3, 9FL174 and 9FL200). . . . . . . . . . . . . . . . . . . . . . . . . . 168

Projects conducted in the Upper Tallapoosa Drainage in Recent Decades. . . . . . . . . . . . . . . . . . . . 172

West Georgia Reservoir Project (Haralson County, Georgia). . . . . . . . . . . . . . . . . . . . . . . 172

Projects conducted in the Chattahoochee River Valley in Northwest Georgia. . . . . . . . . . . . . . . . . 175

The Snake Creek Sites (9CL269 and 9CL270).. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

Projects Conducted in the Upper Coosa River Drainage in Recent Decades. . . . . . . . . . . . . . . . . . . 190

The East Tennessee-Ball Ground Transmission Line Project.. . . . . . . . . . . . . . . . . . . . . . . 191

The Bertha Petty site (9MU8). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

The Swancy site (9GO70). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

Rich Mountain Project (Pickens County). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196

Haig Mill Reservoir Project (Whitfield County).. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198

Site 9MU135 (Murray County Wetlands Project). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202

Lick Creek Site (9GO231). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206

Projects Conducted in the Tennessee River Valley of Northwest Georgia in Recent Decades. . . . . 215

Catoosa Training Center Project (Catoosa County). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216

Assessment of Previous Research on Sites in the Tennessee River Valley in Georgia. . . . 220

Selected Projects in Southeastern Tennessee and Northeastern Alabama. . . . . . . . . . . . . . . . . . . . . 221

Westmoreland-Barber (40MI11).. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222

Russell Cave (1Ja940). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222

Bellefonte Site (1Ja300). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

Wright Farm Site (1Ca18).. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225

Site 1Ra12 in Ruther L. Harris Reservoir. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

Comments on Sites in Nearby Areas of Tennessee and Alabama. . . . . . . . . . . . . . . . . . . . 228

An Examination of Site Files Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229

viii

Chapter 4. The Terminal Archaic in Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 Material Culture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 Fiber-Tempered Pottery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 Soapstone Vessels.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 Projectile Points. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 Other Artifact Types Associated with the Terminal Archaic Period. . . . . . . . . . . . . . . . . . 241 Terminal Archaic Settlement Patterns, Site Structure and Social Organization. . . . . . . . . . . . . . . . . 242 Suggestions for Future Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
Chapter 5. The Early Woodland in Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 Material Culture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248 Early Woodland Projectile Point Types. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248 Early Woodland Pottery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 Dunlap Fabric Marked. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254 Long Branch Fabric Marked. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254 Watts Bar Fabric Marked. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255 Sand Tempered Simple Stamped.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256 Deptford Check Stamped. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256 Plain Wares. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256 Sand Tempered Cord Marked. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 Alexander Pottery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 Other Artifact Types Associated with the Early Woodland Period. . . . . . . . . . . . . . . . . . . 257 Early Woodland Settlement Patterns, Site Structure and Social Organization. . . . . . . . . . . . . . . . . . 258
Chapter 6. Subsistence During the Late Archaic to Early Woodland Transition in Northwest Georgia with Comparative Data from Selected Studies in the Southeast.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 Archaeobotanical Remains from Terminal Archaic and Early Woodland Components in Northwest Georgia and Selected Sites in Northeast Georgia. . . . . . . . . . . . . . . . . . 265 Summary of Archaeobotanical Remains from Terminal Archaic and Early Woodland Components in Northwest and Northeast Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270 Overview of Terminal Archaic and Early Woodland Archaeobotanical Data from Previous Studies in the Southeast. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272 Cold Oak Rockshelter, Lee County, Kentucky. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275 Cloudsplitter Rockshelter, Menifee County, Kentucky. . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 Nutshell from Terminal Archaic and Early Woodland Components in the Southeast. . . . . . . . . . . . 278 Other Plant Remains from Terminal Archaic and Early Woodland Components in the Southeast. . 279 Cultigens and Small Grains.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 Squash and Bottle Gourd. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 Sumpweed/Marshelder. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281 Sunflower. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281 Chenopod/Goosefoot.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282 Knotweed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 Maygrass. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284 Amaranth/Pigweed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 Wild Legumes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 Little Barley.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 Greens. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 Fleshy Fruits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286 Weeds and Herbaceous Plants.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290 Miscellaneous. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
ix

Zooarchaeological Remains from Terminal Archaic and Early Woodland Components in the Southeast.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 Alabama. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 Tennessee. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
Zooarchaeological Remains from the Garfield Site, Cold Oak, and Cloudsplitter Rockshelters. . . . 302 Invertebrates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305 Fish. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 Reptiles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 Birds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 Mammals.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
Summary of Zooarchaeological Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309 Recommendations and Guidelines for Future Research.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
Chapter 7. Research Issues Relating to the Late Archaic to Early Woodland Transition in Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 The Relationship of Biface Production Strategies to the Evolution of Hafted Bifaces. . . . . . . . . . . 315 Function of Hafted Bifaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 Projectile Point Traditions in Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316 Late Archaic and Woodland Biface Production Strategies. . . . . . . . . . . . . . . . . . . . . . . . . . 318 The Relationship of "Archaic" Dart Points and "Woodland" Arrow Points. . . . . . . . . . . . 322 Patterns and Shifts in Lithic Raw Material Preferences Through Time. . . . . . . . . . . . . . . . . . . . . . . 326 Cooking Technology during the Late Archaic to Early Woodland Transition in Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328 Fire-Cracked Rock Cooking Stratigies during the Late Archaic to Early Woodland Transition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328 Hot Rock Technology as Related to Early Vessel Forms. . . . . . . . . . . . . . . . . . . . . . . . . . . 330 Recognition of Late Archaic and Early Woodland Structures and Interpretation of Site Structure and Settlement Patterning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332 The Evolution of Architecture during the Late Archaic and Later Periods. . . . . . . . . . . . . 334 Possible structure Forms for Hunter-Gatherers.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336 Theoretical Concepts for Construction of Late Archaic and Early Woodland Shelters. . . . 337 Building with few or no posts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338 Building for Temporary Permanence:. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338 Evidence for Interaction Between Ridge and Valley and Piedmont Populations during the Late Archaic to Early Woodland Transition in Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . 341 Geomorphological Considerations: Alluvial Burial of Late Archaic/Early Woodland Sites in Northwest Georgia.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343 Soils. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343 Geomorphic Observations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344 Need for Future Work. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346 Conclusions.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
Chapter 8. Summary and Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347 Current Understanding of the Terminal Archaic Period in Northwest Georgia. . . . . . . . . . 354 Site Recognition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355 Recommendations for Future Work. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
References Cited. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
x

List of Figures

Figure 1. Map showing counties and river systems of Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Figure 2. Map showing the Physiographic Provinces in relationship to Northwest Georgia. . . . . . . . . . . . . . 3

Figure 3. Examples of projectile points associated with the Late Archaic through the Early

Woodland periods in Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Figure 4. Physiographic map of Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Figure 5. Map showing the distribution of soapstone deposits in the Eastern U.S.. . . . . . . . . . . . . . . . . . . . 22

Figure 6. Map of a portion of the Etowah River Valley showing Two Run Creek and Walt Jones sites. . . . 29

Figure 7. Map showing the locations of Northwest Georgia reservoirs where large scale

excavations were conducted in the late 1940s through the early 1970s. . . . . . . . . . . . . . . . . . . . . 32

Figure 8. Map showing sites recorded in Allatoona Reservoirs in the 1940s. . . . . . . . . . . . . . . . . . . . . . . . . 34

Figure 9. Illustration from Caldwell's Trend and Tradition showing the "idea of earthenware". . . . . . . . . . 35

Figure 10. Map from Caldwell's Allatoona report showing sites on upper Stamp Creek. . . . . . . . . . . . . . . . . 36

Figure 11. Examples of projectile point types defined for Allatoona Reservoir. . . . . . . . . . . . . . . . . . . . . . . . 37

Figure 12. Illustration of Stamp Creek focus artifacts from a portion of the Stamp Creek site. . . . . . . . . . . . 39

Figure 13. Caldwell's excavation of the Stamp Creek site (9BR60-A) showing all features. . . . . . . . . . . . . . 40

Figure 14. Drawing of a fiber-tempered sherd found on site 9BR60-A in Allatoona Reservoir.. . . . . . . . . . . 41

Figure 15. Map showing Caldwell's work at Boston Creek. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Figure 16. Copy of Figure 35 from Caldwell's Allatoona Reservoir report which

shows the results of stratigraphic testing of the Boston Creek site.. . . . . . . . . . . . . . . . . . . . . . . . . 44

Figure 17. Map showing features in Caldwell's excavation of the Kellog site. . . . . . . . . . . . . . . . . . . . . . . . . 45

Figure 18. Photograph of excavated Early Woodland pits on the Kellog site. . . . . . . . . . . . . . . . . . . . . . . . . . 45

Figure 19. Dunlap Fabric Marked vessels from the Kellog site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

Figure 20. Examples of Kellog and Post-Kellog points from the Kellog site. . . . . . . . . . . . . . . . . . . . . . . . . . 47

Figure 21. Map of Buford Reservoir and locations of investigated sites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

Figure 22. Map of Weiss Reservoir showing excavated sites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

Figure 23. Examples of Coosa points (Upper Valley cluster) from the Weiss Reservoir area. . . . . . . . . . . . . 56

Figure 24. Site map of the Coker Ford site (9Ce200). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Figure 25. Copy of excavation map showing features identified at 1Ce215. . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Figure 26. Photographs showing steatite (soapstone) vessels found on 1Ce215.. . . . . . . . . . . . . . . . . . . . . . . 62

Figure 27. A seriation chart of Early Woodland pottery types found in Weiss Reservoir

63

Figure 28. Maps showing excavations on Carter's Lake site 9MU100. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

Figure 29. Drawing showing the profile of a 20 ft long block excavated in XU-K in 1964. . . . . . . . . . . . . . . 68

Figure 30. Drawing of Late Archaic rock layer from XU-K on 9MU100.. . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

Figure 31. Drawings of Late Archaic points found during

the 1963-1965 excavations of XU-K on 9MU100.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

Figure 32. Map showing sites discussed on the upper Oostanaula and Tennessee River drainages. . . . . . . . . 73

Figure 33. Copy of a schematic cross section from the Williams site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

Figure 34. Map of the Williams site showing locations of artifact

concentrations associated with Archaic and Woodland components. . . . . . . . . . . . . . . . . . . . . . . 75

Figure 35. Examples of chipped and ground stone from the Williams site. . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

Figure 36. Map showing location of the Mahan site

and other Early Woodland sites on a portion of Salacoa Creek. . . . . . . . . . . . . . . . . . . . . . . . . . . 79

Figure 37. Map showing sites to be discussed from the Etowah River drainage. . . . . . . . . . . . . . . . . . . . . . . 83

Figure 38. Sketch map of the Garfield site by James Chapman. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

Figure 39. Map showing Milanich's excavation plan of the Garfield site. . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

xi

Figure 40. Pottery distributions from Unit 480N 565E at Garfield. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Figure 41. A map of the Wooly's Bend site prepared by James Chapman in 1972. . . . . . . . . . . . . . . . . . . . . 89 Figure 42. Examples of possible Terminal Archaic or Early Woodland stemmed
points collected from the surface of Wooly's Bend site by Mary Stover Griffin. . . . . . . . . . . . . . 90 Figure 43. Map showing location of site WGC409.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Figure 44. Map showing distribution of features on the Lewis Sam's Bridge site. . . . . . . . . . . . . . . . . . . . . . 93 Figure 45. Map showing distribution of features on 9CK(DOT)7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Figure 46. Photograph of a possible structure on 9CK(DOT)7.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Figure 47. Early Woodland points from 9CK(DOT)7.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 Figure 48. Profile drawing of lithic workshop excavation trench on the Cagle site. . . . . . . . . . . . . . . . . . . . 101 Figure 49. Plan map of features identified at the base of Stratum D on the Cagle site. . . . . . . . . . . . . . . . . . 102 Figure 50. Two maps showing areas of the Rush site investigated during data recovery. . . . . . . . . . . . . . . . 106 Figure 51. Plots of Early and Middle Woodland pottery types at Rush using
dot density representations from shovel tests excavated during data recovery. . . . . . . . . . . . . . . 107 Figure 52. Map showing distribution of features in the XU1 area at Rush. . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Figure 53. Illustrations of Early/Middle Woodland and Late Archaic types from the Rush site.. . . . . . . . . . 112 Figure 54. Graphs showing relative distribution of point types by level in XU1 at Rush. . . . . . . . . . . . . . . . 113 Figure 55. Illustrations of three bifaces found at Rush that illustrate manufacturing stages
associated with the production of Nolichucky/Copena points.. . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Figure 56. Examples of pottery and groundstone from the Malbone site. . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Figure 57. Examples of projectile points from the Malbone site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 Figure 58. Map showing locations of test pits excavated on the Malbone site. . . . . . . . . . . . . . . . . . . . . . . . 120 Figure 59. Locations of sites investigated for the Route 411 Sewer Line Project.. . . . . . . . . . . . . . . . . . . . . 122 Figure 60. Profile map of Block B on 9FL195. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Figure 61. Illustration of proposed Late Archaic PP/K Lineages at site 9DW77. . . . . . . . . . . . . . . . . . . . . . 130 Figure 62. Drawings of two Otarre-style projectile points found
with Early Woodland pottery in Level 6 on 9DW77. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 Figure 63. Locations of investigated Late Archaic and Early Woodland sites located in
the lower Oostanaula and Coosa River drainages in Northwest Georgia. . . . . . . . . . . . . . . . . . 136 Figure 64. Portion of the excavation map of 9FL56 showing features identified in
two segments of a continuous machine strip. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 Figure 65. Two maps showing distribution of Late Archaic and Early Woodland
sites in the Big Creek Watershed Project area in Polk County. . . . . . . . . . . . . . . . . . . . . . . . . . . 140 Figure 66. Map showing shovel tests and test pits on 9PO27. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 Figure 67. Map showing backhoe trenches and block excavations on 9PO27.. . . . . . . . . . . . . . . . . . . . . . . 142 Figure 68. Two maps showing data from shovel tests and trenches on 9PO27. . . . . . . . . . . . . . . . . . . . . . . 143 Figure 69. Two graphs showing the co-occurrence of various pottery types in Pumpkin Pile
features that also produced simple-stamped pottery.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 Figure 70. Representation of co-occurrence of Dunlap, Long Branch, and Watts Bar
pottery in features on the Pumpkin Pile site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 Figure 71. Early Woodland and Terminal Archaic projectile points from Pumpkin Pile.. . . . . . . . . . . . . . . 149 Figure 72. Examples of probable Early Woodland soapstone objects from Pumpkin Pile. . . . . . . . . . . . . . 152 Figure 73. Illustrations of slate "hoes" found at Pumpkin Pile. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Figure 74. Photograph showing feature excavation in the North Block at Pumpkin Pile. . . . . . . . . . . . . . . 155 Figure 75. Plan maps of North and South Block excavations at Pumpkin Pile. . . . . . . . . . . . . . . . . . . . . . . 156 Figure 76. Map showing the geographic range of the Pumpkin Pile complex in relationship to
the Cedar Bluff and Kellog phases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
xii

Figure 77. Examples of projectile points identified as Late Archaic/Early Woodland from 9FL203. . . . . . 161 Figure 78. Examples of Late Archaic projectile points from 9FL206. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 Figure 79. Examples of stemmed Late Archaic, Terminal Archaic, and Early Woodland points from
9FL47 (UGA's Mary Stover Griffin Collection). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 Figure 80. Examples of projectile points attributed to Late Archaic, transitional Late Archaic to
Early Woodland, and Woodland period occupations on 9FL174.. . . . . . . . . . . . . . . . . . . . . . . . 171 Figure 80. Map showing location of proposed West Georgia Reservoir. . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 Figure 82. Map showing Late Archaic and Early Woodland sites within the
proposed West Georgia Reservoir survey area.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Figure 83. Map showing the locations of two excavated Early Woodland sites in the
Chattahoochee River Valley in Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 Figure 84. Map showing data recovery excavations on 9FL269 and 9FL270. . . . . . . . . . . . . . . . . . . . . . . . 179 Figure 85. Plan map of data recovery excavations on 9CL269. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 Figure 86. Plan map of data recovery excavations on 9CL270. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 Figure 87. Photographs showing the range of Early Woodland and possible Terminal
Archaic projectile points from the Snake Creek sites.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 Figure 88. Comparative figure showing two cache points from Snake Creek
and two Rankin points from East Tennessee. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Figure 89. Map showing sites and survey areas associated with the upper Coosa drainage. . . . . . . . . . . . . 190 Figure 90. Map showing East Tennessee-Ball Ground corridor and locations
of tested sites 9MU8 and 9GO70 on the Coosawattee River. . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 Figure 91. Examples of Late Archaic and Early Woodland artifacts from 9GO70. . . . . . . . . . . . . . . . . . . . 194 Figure 92. Examples of Late Archaic, Terminal Archaic, and Early Woodland points
from the Rich Mountain project area.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 Figure 93. Map showing Haig Mill survey area and locations of sites 9WD42 and 9WD45.. . . . . . . . . . . . 198 Figure 94. Map showing testing plan on 9MU135. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 Figure 95. Projectile points recovered from large pit on 9MU135.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 Figure 96. Testing phase map of site 9GO231. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 Figure 97. Selected artifacts from the testing phase excavations on 9GO231. . . . . . . . . . . . . . . . . . . . . . . . 208 Figure 98. Profile maps of Test Unit 3 and Feature 1 on 9GO231.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 Figure 99. Map of data recovery excavations on 9GO231.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 Figure 100. Examples of possible Terminal Archaic points from 9GO231. . . . . . . . . . . . . . . . . . . . . . . . . . . 213 Figure 101. Map showing Tennessee Valley sites and survey areas referenced in text. . . . . . . . . . . . . . . . . . 215 Figure 102. Map of the Catoosa Training Center survey area with the locations of five
sites interpreted as dating to the Early Woodland period. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 Figure 103. Illustrations of "Early Woodland" point types found in the Catoosa Training Center tract. . . . . 218 Figure 104. Map showing sites to be discussed in nearby areas of Tennessee and Alabama. . . . . . . . . . . . . 221 Figure 105. Examples of tapered-stemmed projectile points found in a ca. 600 B.C.
aceramic zone at site 1Ra12 in Randolph County, Alabama. . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 Figure 106. Map showing the distribution of Late Archaic sites in Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . 229 Figure 107. Map showing the distribution of Early Woodland sites in Georgia. . . . . . . . . . . . . . . . . . . . . . . 230 Figure 108. Graph showing frequencies of Late Archaic and Early Woodland components
recorded in the study area in order from north to south. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 Figure 109. Map showing locations of selected sites and project areas referenced in the text. . . . . . . . . . . . 235 Figure 110. Examples of probable Terminal Archaic projectile points from a private Gordon
County Collection.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 Figure 111. Examples of projectile points attributed to the proposed Flint Creek culture. . . . . . . . . . . . . . . . 240
xiii

Figure 112. Comparison of radiocarbon dates from Early Woodland sites in two areas of Northwest Georgia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
Figure 113. Outline drawings of Early Woodland projectile point types defined in the Phipps Bend area of northeastern Tennessee. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
Figure 114. Illustrations of Caldwell's Middle Eastern Tradition pottery types. . . . . . . . . . . . . . . . . . . . . . . 251 Figure 115. Photographs of selected Early Woodland fabric-marked and simple-stamped sherds. . . . . . . . . 253 Figure 116. Map of Allatoona Reservoir showing locations of Kellog phase sites identified
by Caldwell (1957) and a later survey. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 Figure 117. Large Cartersville "storage vessel" from the Kellog site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261 Figure 118. Map of Northwest Georgia and Northeast Alabama showing the locations of previously
defined Early Woodland (and Late Gulf Formational) phases and complexes. . . . . . . . . . . . . . . 262 Figure 119. Map of the Southeast showing locations of referenced sites and survey areas. . . . . . . . . . . . . . . 266 Figure 120. Map of the lower Tennessee River Valley showing distribution of contemporaneous
Ledbetter and Savannah River phase sites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318 Figure 121. An example of Late Archaic biface production strategy as found in a cache of
projectile points and preforms on site 9GO231 in Northwest Georgia. . . . . . . . . . . . . . . . . . . . 319 Figure 122. A model for Late Archaic biface production strategy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320 Figure 123. Photograph of excavated Middle Woodland structure from site 9FU14 (Kelly 1973)
accompanied by an artist's rendering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336 Figure 124. Examples of Early Woodland structure patterns from four sites in the
Southeast and Mid-South. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340 Figure 125. Photograph of backhoe trench 4 at 9GO286 illustrating a well expressed
buried A horizon in contact with overlying historical sediment. . . . . . . . . . . . . . . . . . . . . . . . . 344
xiv

List of Tables
Table 1. A list of Projects and Sites to be Examined as Previous Research.. . . . . . . . . . . . . . . . . . . . . . . . . . 25 Table 2. Radiocarbon Dates from the Kellog Site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Table 3. Tentative Chart of Woodland Cultures in Northwest Georgia.. . . . . . . . . . . . . . . . . . . . . . . . . . 78 Table 4. Pottery Counts from Early Woodland Features Excavated by John Wear
at the Mahan Site, 9GO22 in Gordon County.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Table 5. Radiocarbon Dates from 9CK(DOT)7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 Table 6. Distribution of Pottery in Test Unit N90 E135 on 9FL164. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Table 7. Artifacts found during Shovel Testing of 9FL164. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Table 8. Radiocarbon Dates from the Rush Site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Table 9. Pottery Distribution (grams per m3 ) in XU1 at Rush.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Table 10. Diagnostic Projectile Points from the Rush site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Table 11. Hafted Bifaces by Raw Material from the Malbone Site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Table 12. Radiocarbon Dates from Sites 9DW64, 9DW77, and 9CK713. . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 Table 13. Diagnostic Projectile Points from 9DW64, 9DW77, and 9CK713. . . . . . . . . . . . . . . . . . . . . . . . . 127 Table 14. Radiocarbon Dates from Site 9PO27. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 Table 15. Identified Pottery Types from the Pumpkin Pile Site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 Table 16. Projectile Point Types from the Pumpkin Pile Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 Table 17. Comparison of Projectile Point Types Found at Two Early Woodland Sites. . . . . . . . . . . . . . . . . 151 Table 18. List of Projectile Points from the Surface of 9FL47. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 Table 19. West Georgia Reservoir Radiocarbon Dates.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 Table 20. Radiocarbon Dates from Sites 9CL269 and 9CL270. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 Table 21. List of all Ceramics from 9CL269. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 Table 22. List of all Ceramics from 9CL270. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 Table 23. List of Hafted Bifaces from 9CL269 and 9CL270 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Table 24. Distribution of Woodland Pottery in Test Unit 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 Table 25. Radiocarbon dates from 9WR42. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 Table 26. Radiocarbon dates from the Lick Creek site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 Table 27. List of Artifacts from Five Early Woodland sites on the Catoosa Training Center.. . . . . . . . . . . . 217 Table 28. List of Component Designations from the State Site Files .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 Table 29. List of Radiocarbon Dates Retrieved from Various Late Archaic, Terminal Archaic,
and Gulf Formational Stage Sites in Northwest Georgia and Northeast Alabama. . . . . . . . . 234 Table 30. List of Early Woodland Radiocarbon Dates from Northern Georgia and Alabama.. . . . . . . . . . . 246 Table 31. Species List, Terminal Archaic and Early Woodland Components in the Southeastern
United States, Non-Commensal (Food) Species.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 Table 32. Species List, Terminal Archaic and Early Woodland Components in the Southeastern
United States, Commensal (Non-Food) Species.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298 Table 33. Vertebrate Zooarchaeological Remains from Russell Cave, Alabama. . . . . . . . . . . . . . . . . . . . . . 299 Table 34. Zooarchaeological Remains from Cloudsplitter Rockshelter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304 Table 35. Zooarchaeological Remains from Cold Oak Rockshelter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304 Table 36. Zooarchaeological Remains from the Garfield Site, Early Woodland Component.. . . . . . . . . . . . 305
xv

xvi

Chapter 1. Introduction
The Georgia Department of Transportation (GDOT) contracted with Southeastern Archaeological Services, Inc. (SAS), to develop a cultural context document for the Ridge and Valley province of Northwest Georgia that would examine the transition from the Late Archaic to the Early Woodland periods. Motivation for the study originated, in part, from two recent testing projects conducted by SAS in Murray County (Gresham 2002) and Gordon County (Patton and Gresham 2003). The projects produced site data that appeared to provide evidence of occupations dating to a time period that immediately preceded the introduction, or at least the widespread use, of Early Woodland fabric-marked pottery in the region. Gresham suggested that the Murray County sites might represent evidence of either a "preceramic phase of the Early Woodland period (or terminal Late Archaic)" or an aceramic occupation from a time when pottery was used only on some sites in the region (Gresham 2002:49). Gresham (2002:73) developed this concept further with the following statement:
If these sites were to date to the later end of that time frame, intriguing questions arise as to why these people maintained an Archaic (non-ceramic) lifestyle, while neighbors to the south (Kellogg) and north (Watts Bar) adopted pottery. It may be more likely that as Early Woodland cultures (such as Kellogg and Watts Bar) coalesced around 800 B.C., that large areas, such as this length of the Conasauga River, were essentially devoid of habitations. By this reckoning, the sites here would date to the Terminal Archaic florescence, around 1,200 to 800 B.C.
From an historical standpoint, the recognition of this intermediate period in Northwest Georgia would do much to resolve the many issues associated with Caldwell's (1957, 1958) suggestion of a discontinuity within the material culture on sites occupied by the earliest pottery-producing people (Kellog focus) and their Archaic predecessors in the region or possibly even a "hiatus" in occupation.
Elsewhere in Eastern North America, archaeologists have developed several models to explain the abrupt changes recognized by Caldwell (1957, 1958) in Northwest Georgia. These include the replacement of Late Archaic populations by immigrants with ceramic technology and perhaps an associated subsistence knowledge (Emerson and McElrath 2001), a collapse of social and political organization related to the demise of Poverty Point (Gibson 1999, 2000), and the effects of long term climate change on hunter-gatherer subsistence strategies and habitation locations (Fiedel 2001; Kidder 2006).
For the purposes of this study, this period of Late Archaic to Early Woodland transition is viewed as beginning at approximately 1,200 to 1,100 B.C. and extending to approximately 700 to 600 B.C. (arbitrarily designated 1,100 to 600 B.C. in this volume). This interval of time generally equates with the demise of the Late Archaic Poverty Point and Stallings Island cultural spheres of influence across the greater Southeast and precedes the widespread use of Early Woodland pottery in Northwest Georgia. The selection of this time interval for a Late Archaic to Early Woodland transition is consistent with previous research conducted in other regions of the Eastern United States (Kidder 2006:195-198). Kidder (2006:196) notes that: "Although chronologies vary, this transition is dated ca. 3200-2400 cal. B.P. Across much of eastern North America there is an abrupt break in regional, local, and site occupations coupled with notable alterations in settlement, economy, and society." The authors of this volume have selected the cultural designation Terminal Archaic period for this time of transition in Northwest Georgia.
1

Purpose and Goals of Project

The primary purpose of this study is the compilation of information that may be used to develop a cultural context for future archaeological investigations of sites of this time period in Northwest Georgia. Traditionally, archaeological investigations in the region have focused on site specific issues which generally consist of artifact identifications and an examination of site structure (stratigraphy, feature distributions, etc.). The quality of the information produced, and the manner in which that information is interpreted, is largely dependent on the background and experience of the supervising archaeologist. Important information has been produced as a result of these studies but much too often the sites have been viewed in cultural isolation or interpreted in such broad terms of little significance. The time period represented by the transition from Late Archaic to Early Woodland is viewed quite differently in various parts of the Southeast (see following discussion), and Northwest Georgia has actually received relatively little intensive study with respect to this time period. A primary goal of this project was to review the existing reports of investigations from Northwest Georgia that relate to this time period and determine how the data may be used to characterize the region. Another goal was to compare the Georgia data to previous studies of the time period in other parts of the Southeast as a means of refining our understanding of this time period in terms of material culture, settlement patterns, and subsistence.

Definition and Geographic Extent of Study Area

The study area will focus on the archaeological investigations conducted primarily within a ten county area of the Ridge and Valley and Cumberland Plateau region in the northwestern corner of Georgia (Figure 1). Because relatively few large scale archaeological investigations have been conducted in many areas of the Ridge and Valley in Northwest Georgia, selected site data is also examined from relevant projects conducted in the counties along the western edge of the Piedmont province. An effort was made to limit this additional study to site investigations located near the outer edge of the Piedmont. However, for practical purposes, the valley of the Chattahoochee River represents the eastern limit of our study area. Supplemental information has also been procured from a few sites and projects in northeastern Alabama and southeastern Tennessee. With respect to a broader context for this study, additional information has been procured from major studies conducted throughout the Southeast.

This area represents a relatively small portion of an extensive physiographic feature that extends from the fall line of Alabama to New England (Figure 2). The major feature of the Ridge and Valley is the Great Valley, also known in Georgia and Alabama as the Coosa Valley. The eastern edge of the Cumberland Plateau is included.
2

Figure 1. Map showing counties and river systems of Northwest Georgia.

Figure 2. Map showing the Physiographic Provinces in relationship to Northwest Georgia.
Description of the Late Archaic and Early Woodland Transition in the Southeast
Because the Southeast is known to be composed of distinct geographical areas of strong cultural influence during the Late Archaic period, and to some extent during the Early Woodland period, simply describing the material, social, and behavioral aspects of Late Archaic and Early Woodland cultural traditions is quite complicated. The following paragraphs provide some basic concepts used in traditional definitions of these two periods and introduces a transitional period that is recognized in certain areas of the Southeast as the Terminal Archaic period.
3

The Late Archaic Period
The Late Archaic period is generally dated to the period of 3,000 to 1,000 B.C. which translates to 4,950 to 2,950 B.P.(radiocarbon years prior to 1950). Throughout most of the twentieth century, any interpretation of the Late Archaic period in Northwest Georgia was linked to Stallings Island culture (Claflin 1931; Caldwell 1952; DePratter 1975; Fairbanks 1942). Although Northwest Georgia lacked Late Archaic shell middens and fiber-tempered pottery was practically absent, Caldwell (1957) defined his Stamp Creek focus for the Allatoona area based primarily on the similarity to the material culture of Stallings Island. More recent studies of Late Archaic sites located along the Savannah River (Elliott et al. 1994; Elliott and Sassaman 1995; Sassaman 1993a, 2006a) have helped to clarify the picture substantially. We now have a series of defined phases and cultural traditions that divide that 2,000 year long period of time into meaningful segments for study. These phases are defined primarily by differences in material culture such as morphological changes in projectile point styles over time, the differential use soapstone cooking technology, and the presence, absence, or differential surface treatment of fibertempered pottery. One particularly critical result of the refinement of the Late Archaic sequence is that the classic Savannah River stemmed point is now firmly established within the time range of the middle third of the period, and it is preceded and followed in time by smaller stemmed forms (Sassaman 2006:56). As a result, the earlier concept that large stemmed points are diagnostic of the entire length of the Late Archaic period in Georgia no longer applies.
While we now have a good understanding of the Late Archaic cultural sequence for the eastern portion of Georgia, that is not the case for Northwest Georgia. Fiber-tempered pottery is essentially absent; Stanyard (2003:58) has suggested that the entire area of Northwest Georgia should be placed within an undifferentiated Late Archaic phase because "a general lack of information impedes our ability to properly assess the nature of Late Archaic developments in those regions."
Part of the problem in interpreting the Late Archaic period of Northwest Georgia is that the projectile point tradition compares more favorably to types found in Tennessee and Alabama than the Piedmont region of Georgia. This tradition consists of notched, expanded stemmed, and straight stemmed point types that contrasts with the predominantly straight-stemmed points found in the Piedmont. The Ledbetter phase in Tennessee is defined by several of these types and it has also been defined as extending over most of the length of the Late Archaic period (Bentz et al. 1986; Bowen 1979; Davis 1974). A report by Webb (1998) is one of the few instances in which parallels to the Ledbetter phase have been proposed for a Late Archaic site in Northwest Georgia. Figure 3 illustrates several examples of projectile point types used in Northwest Georgia during the Late Archaic through Early Woodland periods.
The Gulf Formational Stage, a significant concept developed by Alabama archaeologists (Walthall and Jenkins 1976), removes cultures associated with the introduction of early pottery (predominantly fiber tempered) from both the Late Archaic and Early Woodland periods as traditionally defined. To some extent, the concept expands upon Caldwell's (1958) idea of a Gulf Tradition that witnessed a distinctive pottery development during the Woodland period. The Gulf Formational Stage in the Southeast is divided into Early (2,500-1,200 B.C), Middle (1,200-500 B.C.) and Late (500-100 B.C.) periods. The earliest period was associated with the introduction of fiber-tempered pottery along the Savannah River and the Atlantic Coast. The Middle Gulf Formational period was applied to the spread of fibertempered pottery to the west and north.
4

Figure 3. Examples of projectile points associated with the Late Archaic through the Early Woodland periods in Northwest Georgia.

Geographically, sites defined as Middle Gulf Formational Stage extend northward from the Gulf Coast, into the lower portion of the Tennessee Valley (Wheeler and Pickwick reservoir areas) and along the Tombigbee River of Eastern Mississippi, where the pottery is referred to as Wheeler. The development of the Poverty Point culture that originated in the lower Mississippi valley is an integral aspect of the Middle Gulf Formational period (Jenkins and Krause 1986:33; Walthall and Jenkins 1976:45, Walthall 1980:85-86). The Late Gulf Formational Stage is most frequently defined by sand-tempered Alexander pottery and smaller stemmed projectile points such as Flint Creek. The distribution of Alexander pottery is similar to that of Wheeler on the Tennessee River but it is also found on the middle portions of the Coosa River in Alabama (Meredith 2007). By definition, the Gulf Formational Stage requires the presence of Wheeler and Alexander pottery on sites. For that reason, the designation is used mostly by archaeologists working in Alabama , Mississippi, Louisiana, and parts of Middle and West Tennessee. The concept of a Gulf Formational Stage is seldom recognized elsewhere and in most of Georgia archaeologists still view fiber-tempered pottery as a Late Archaic trait.

5

The latter part of the Late Archaic period in Northwest Georgia remains poorly documented and phase designations remain to be formulated. The highly varied forms of stemmed projectile points typical of this time period, such as Otarre and Iddins (see Figure 3), occur frequently on sites in the western Piedmont and along the eastern edge of the Ridge and Valley, however, evidence of intensive occupation throughout the remainder of the Ridge and Valley province has yet to be adequately documented.
The Early Woodland Period
In Georgia and in most of the Southeast, the date usually applied to the beginning of the Early Woodland period is 1,000 B.C. (Anderson and Joseph 1988:208; Anderson and Mainfort 2002:5; Bense 1994:110). A slightly later beginning date of 700 B.C. may be more appropriate for northern Georgia (Espenshade 2008:123). The 1,000 B.C. date represents the introduction of sand-tempered or crushed quartz-tempered pottery in parts of the Southeast but it is important to note that pottery was not universally adopted at that time. In some areas of the Piedmont, pottery may not appear on sites until the latter part of the Early Woodland or even Middle Woodland times (Knight 1998:190). Anderson and Mainfort (2002:5) end the Early Woodland period at 200 B.C. to correspond with the era of increased mound construction and participation in the Hopewellian exchange network in many areas.
Until quite recently, the Woodland period was thought to begin with the initial appearance of pottery across much of the region, the beginning of burial mound ceremonialism, the emergence of sedentary village lifestyles, and evidence of intensive horticulture (Griffin 1967:180). More current thinking suggest that Early Woodland people continued to maintain many of the Archaic traditions in parts of the Southeast so the division between Late Archaic and Early Woodland is not so clear cut. The description "Archaic with pottery" has been applied to the Early Woodland period in some areas of the Southeast (Anderson and Mainfort 2002:5). However, Sassaman (2002:399) stresses the fact that the widespread use of pottery (as it applies to different points in time during the Early Woodland period) resulted in the "dissolution of Archaic cultural traditions that thwarted change." Referring to the previous work of Smith (1986), Sassaman goes on to suggest that widespread use of pottery resulted in social change and led to important developments in cooking and storage technology that enabled the more effective use of specific starchy and oily seeded plant resources (Sassaman 2002:399).
The regional manifestation of the Early Woodland period in Northwest Georgia is thought to have begun with Caldwell's (1958) Kellog focus which is defined by sand-tempered Dunlap Fabric Marked pottery (Bowen 1989; Garrow 2002:2; Sassaman 2002:418, Wood and Bowen 1995). Wood and Bowen (1995:10) have proposed a beginning date of 760 B.C. (calibrated radiocarbon age) for the Kellog period. Beyond the range of Kellog, Early Woodland culture areas may produce fabric-marked pottery tempered by limestone (Long Branch) or crushed quartz (Watts Bar). At the present time, there is little evidence of an earlier cord-marked pottery tradition known as Swannanoa which represents the first Early Woodland pottery used in northeastern Tennessee (Lafferty 1978) and the Piedmont of North Carolina (Keel 1976). Diagnostic projectile points associated with the Kellog period occupations include varieties of Woodland triangular forms which appear throughout the Southeast around 600 B.C. Earlier types of Woodland points include a variety of rather small types that may be notched or stemmed (Lafferty 1981). The predominance of stemmed points over triangular points is viewed as one means of distinguishing earlier occupations of the Early Woodland period in the region (Futato 1998:216). Chapter 5 should be consulted for further information for pottery types and point types found in the region.
6

The Terminal Archaic Period
The previous discussion demonstrates that the ca. 1,000 B.C. date for separating the Late Archaic from the Early Woodland periods in areas such as Northwest Georgia is arbitrary and of limited significance. A number of years ago, following the excavations at the Westmoreland-Barber site near Chattanooga, Faulkner and Graham (1966:122) proposed the concept of a "terminal Archaic" period to bridge the "shadow zone" (ca. 1,200-600 B.C.) between the Archaic and Woodland periods. According to Chapman (1985:151-152) the Terminal Archaic lithic assemblage included biface forms that continued to be used into the Early Woodland period (thus forming a link between "aceramic" and ceramic period occupations). Over the ensuing years, the idea of a Terminal Archaic period has been embraced by archaeologists working in the Cumberland Plateau region and areas to the north and west where pottery does not appear in any numbers until late in the Early Woodland period. Formally defined Terminal Archaic phases include Wade in Middle Tennessee (Keel 1978), Cogswell in eastern Kentucky (Ison 1986:212), and Prairie Lake in the American Bottoms area in Illinois (Emerson 1980). The Terminal Archaic period is generally not recognized in areas where Early Woodland pottery appears earlier.
In many areas of the Southeast, the Terminal Archaic period is now defined by a number of radiocarbon dates in direct association with barbed projectile point types such as Wade and Motley and bulbous base forms such as Flint Creek (Childress 1999). Soapstone vessels are frequently found on Terminal Archaic sites and interactions with Poverty Point have been suggested (Childress 1999:75). While the Terminal Archaic period is normally defined by lithic artifacts, DeJean and Benthall (1994:130) note that the period is also marked by the occurrence of early cultigens such as gourds and squash. Terminal Archaic floral remains have been well studied in the dry rock shelters of Kentucky, and that information forms the basis for much of the subsistence data discussed in Chapter 6 of this volume. Terminal Archaic projectile point types such as Wade (Cambron and Hulse 1964), Delhi (Ford and Webb 1956), Motley (Ford et al. 1955), and other point types included by Justice (1987:179-184) in his "Terminal Archaic Barbed Cluster," have now been identified on several sites in Northwest Georgia. A radiocarbon date of 760 B.C. (uncalibrated) was recently procured from a soapstone sherd on the Lick Creek site in Gordon County (Benson 2008). For now, the available data allows Terminal Archaic period occupations to be isolated on a very few sites in portions of extreme northwestern Georgia.
Previous investigators in Northwest Georgia have referred to potentially comparable "aceramic" components by several names that include "Terminal Archaic" (Beasley 1972; Ledbetter 1992; Ledbetter et al. 1987; Smith et al. 1988,), "terminal Late Archaic" (Byrne 1987; Elliott et al. 1993b; O'Steen 1987; Patton 2003:42) and the "Late Archaic/Early Woodland transitional period" (Deter-Wolf and Karpynec 2005; Southerlin et al. 1996). A review of these Georgia reports suggest that occupations during this time period are most often defined by a variety of stemmed projectile points of medium size, a lack of pottery, and frequently by the presence of soapstone vessel fragments. Sites located along the boundary of the Piedmont and Ridge and Valley (Cartersville Fault area) frequently produce points with straight or contracting stems of the Piedmont tradition rather than the barbed types that traditionally define the Terminal Archaic period elsewhere in the Southeast. Because some of these sites may simply represent local manifestations of the latest part of the Late Archaic period, caution must be exercised in distinguishing Late Archaic from Terminal Archaic occupations. For that reason, the use of the designation "Terminal Archaic" is best applied to sites that display artifact types (primarily distinctive projectile point types) considered diagnostic of the period in Tennessee, Kentucky, and elsewhere.
7

Introduction of Research Issues to be Addressed
This volume represents a concerted effort to address a series of issues that will hopefully lead to further research for a poorly understood time period in prehistory. A substantial portion of the report examines previous research relating to the Late Archaic and Early Woodland periods in Northwest Georgia. This was considered important as a means of establishing a foundation for the concept of a Late Archaic/Early Woodland transition. Basically, the interval has received little attention from Georgia archaeologists and for that reason is inadequately researched in the region The review of previous research has allowed the authors of this volume the luxury of procuring relevant information from a highly diverse range of archaeological perspectives based on nearly seven decades of survey and excavation of Late Archaic and Early Woodland sites in Northwest Georgia. A more limited review of previous research was conducted for nearby areas of Alabama and Tennessee.
Using data from previous research conducted in Georgia and elsewhere in the Southeast, several topics will be examined that relate to cultures of the latter part of the Late Archaic period and the Early Woodland periods in Northwest Georgia (Chapters 4 and 5). Because a substantial amount of research has been conducted on sites of this general time period elsewhere in the Southeast, Chapter 6 has been allotted to a discussion of subsistence during the period of Late Archaic to Early Woodland transition. The subsistence data from sites investigated in the Southeast, and in some areas of the Mid-South allows a fuller interpretation of the rather limited subsistence data recovered previously in Northwest Georgia.
A series of hopefully relevant research issues will be addressed in Chapter 7 of this volume. Some of the research issues originated from recent work conducted in Murray and Gordon Counties which provided evidence of occupations during this transitional period (Gresham 2002; Patton and Gresham 2003; Benson et al. 2008). These research issues relate to the changes in hafted biface styles from the Late Archaic to Early Woodland periods, changes in lithic raw material procurement during this interval of time, aspects of cooking technology that preceded the introduction of ceramic vessels (with a focus upon the use of fire-cracked rock), the nature of Late Archaic through Early Woodland period domestic architecture, and evidence for interaction between people occupying the Ridge and Valley and Piedmont provinces, and the scale of alluvial deposit on floodplain sites in the region. These research issues are addressed using information from previous work conducted in Northwest Georgia, interpretations of other researchers elsewhere in the Southeast, and some experimental data. In all cases, our efforts to address these research issues should be viewed as a preliminary effort to inspire further research.
Authorship
This report is a joint effort of the three authors. With the exception of Chapter 3 which was compiled by the senior author, each contributed to the various chapters of the volume. Primary authorship of two chapters should be noted. Chapter 6, which reviews subsistence data from throughout the Southeast was compiled and written by Lisa O'Steen. Chapter 7, which addresses a number of research issues that required a background in replication and experimentation in a number of areas of primitive technology, was written primarily by Scott Jones with some contributions by the other authors. David Leigh addressed specific issues relating to geomorphology in Chapter 7. Mr. Jones was also the primary author for the geologic and physiographic background sections of the report (Chapter 2). The artifact drawings found throughout the report were prepared by Lisa O'Steen.
8

Chapter 2. Environmental Setting of Northwest Georgia
The northwestern corner of Georgia is characterized by geologic conditions that are quite distinct from any other region in the state. The Ridge and Valley Province consists of Paleozoic age sedimentary rocks that have been altered through time to form a series of parallel ridges and valleys. Linearity of major landforms and extremes of topography increase towards the northwest. Wharton (1978:108) refers to Northwest Georgia as the Sedimentary Region. The environmental setting of the province will be discussed with respect to physiography, geology, and present and past environmental conditions. The lithic resources considered most important to the prehistoric inhabitants of the region during the this transitional period will be discussed in greater detail.
Physiography and Geology of Northwest Georgia
It is generally accepted that the Ridge and Valley Physiographic Province in Georgia is made up of four distinct regions (Figure 4). These are the Cumberland Plateau, Chickamauga Valley, Armuchee Ridges, and the Great Valley. Although the Cumberland Plateau (the Appalachian Plateau) region is typically considered a separate physiographic region, it is continuous with the Ridge and Valley, and comprises only a small portion of the extreme northwestern corner of the state. In consideration of the small area occupied by the Cumberland Plateau within the state, it is included in this study for purposes of cultural and environmental continuity.
The Ridge and Valley province is comprised of folded and eroded Paleozoic sediments. In Georgia, this province consists of all sedimentary formations located to the west and north of the Cartersville Fault which is sometimes referred to as the Great Fault or the Great Smoky Fault (Georgia Geologic Survey Map 1976). Along the northern part of the Cartersville fault, the rolling topography of the Great Valley contrasts sharply with the Blue Ridge, which stands as much as 450 m above. The present form in the province is the result of late Paleozoic tectonic activity and subsequent erosion (Hodler and Schretter 1986). The Ridge and Valley in Georgia is unique in that it sits at the southern and eastern terminus of the Appalachian and Blue Ridge Mountain chain and lies in close proximity to, and is readily accessible from, the Piedmont province.
The present structure of the Ridge and Valley (including the Cumberland Plateau) in Georgia results from the weathering of these folded sediments. Erosion over the last 200 million years has worn these folded sediments down to the original synclinal floors. With continued erosion, the resistant Pennsylvanian sandstone that once formed valley floors became elevated as surrounding rocks weathered away. This sandstone now caps the ridges, forming "synclinal ridges" (Wharton 19787:108).
The topography of the Ridge and Valley becomes more pronounced and inaccessible from southeast to northwest. These distinct changes in topography have led to the division of the Ridge and Valley into four main sub-regions or sections. In order of treatment here, these are the Great Valley, the Armuchee Ridges, the Chickamauga Valley, and the Cumberland Plateau.
9

Figure 4. Physiographic map of Northwest Georgia.
The Cartersville Fault separates the Ridge and Valley and Piedmont Physiographic Provinces. In Northwest Georgia, the fault forms the eastern and southern boundary of the Ridge and Valley province and parallels the region known as the Great Valley. In Georgia, this fault runs from the area near Tenga, southward through Chatsworth, to Emerson, and just to below Cartersville. From Emerson, the fault winds in a west-southwesterly direction into the southern edge of Polk County, just below the community of Esom Hill.
10

For the purposes of this study, Charles Wharton has succinctly summarized this region in The Natural Environments of Georgia (1978:108-112). The following descriptions cite liberally from his work, with minor omissions of figure references and less relevant portions of his text. His introductory paragraph sets the tone for any natural history consideration of the region:
"Nowhere in Georgia is the bedrock geology of the land more visible, exciting, or easier to relate to land form than in the Sedimentary Region. Enclosed by the great Cartersville Fault and composed of sedimentary deposits of Paleozoic age, many rocks contain the fossils of early life on earth, life that inhabited shallow, inland seas and, at death, was covered by the sediments of these areas. These sediments are now limestones, shales, sandstones, and cherts. Even some coal beds containing fossil plant life are present. Driving across these long, linear ridges and intervening level valleys can be a fascinating lesson in geological history. Natural environments in this sector can often be tied into the bedrock geology with more confidence than in the Blue Ridge or Piedmont provinces." [Wharton 1978:108].
The Great Valley is bounded to the east and south by the Cartersville Fault, and it abuts the Armuchee Ridges section to the west and north. This region extends in a southward and westward arc. It averages around 15 km wide at the northern border with Tennessee, widening to an average of about 26 km in the southern and western portions. The Coosa River and tributaries form the major watershed region through the Great Valley. Numerous rolling hills are contained in the Great Valley and the Coosa River has encroached upon some of these hills, creating bluff forests. The bulk of Great Valley soils are underlain by rocks of the Knox group, composed of limestone and dolomites underlying flatter terrain, and the Conasauga group, composed of shales, sandstones, and limestones underlying gently rolling hills (Wharton 1978:108). Wharton continues: "Since it possesses some remarkable Coastal Plain plants and animals, the valley of the Coosa and tributaries is considered a separate area....This valley sweeps down from Tennessee diagonally across northwest Georgia and extends nearly halfway down the state of Alabama" (Wharton 1978:112). Underlain by rocks of Ordovician and Cambrian age, the Great Valley is characterized by a landscape of generally low topography. At an elevation of 200 to 250 m, the Great Valley has a few ridges and rolling hills with local relief of 15 to 30 m.
To the north and west of the Great Valley lies the Armuchee Ridges section. This section is characterized by an elevation of 250 to 500 m, with ridges rising 200 m above the valleys. The tops of the linear hills that form these ridges are composed of Red Mountain sandstone of Silurian age. Downslope on the east side and often forming cliffs are cherts, slates, and shales of Mississippian age. The west flanks may be underlain by limestones or siltstones, both of which support deciduous forests (Wharton 1978:108).
The Armuchee Ridges section is bounded by Taylor's Ridge to the west and southeastward by an outlier called Horseleg Mountain, almost in the city limits of Rome. The valleys between Horseleg and Taylor's Ridge are underlain by Floyd shale (with some sandstone and limestone), thus differing from the Great and Chickamauga Valleys which are largely floored by limestones and dolomites.
The Chickamauga Valley lies to the west of the Armuchee Ridges, dividing this section from the Cumberland Plateau. Basically, it is a linear valley comprised of Cambrian/Ordovician limestones and dolomites. The northern portion consists of two valleys divided by the Pigeon Mountain extension of the
11

Cumberland Plateau. West Chickamauga Creek drains the northwestern-most of these. Flowing to the north-northeast, Little Chickamauga Creek drains the central part. Near LaFayette, the Chattooga River takes over from Little Chickamauga Creek, flowing to the south-southwest. The combined Little Chickamauga/Chattooga systems cut an almost continuous linear swath across the northwestern corner of the state. The elevations of the ridges range from 200 to 300 m and stand about 60 to 90 m above the adjacent valley floors (Wharton 1978:108).
The Cumberland Plateau ranges in elevation from 250 to 600 m, with ridges rising 250 to 350 m above valley floors. The Cumberland Plateau section is represented by broad, sheer-sided mountains (Lookout, Pigeon, and Sand) with broad plateaus on top. This is quite unlike any other area in the state. The plateau top is capped by Castle Rock sandstone. The rim-rock is weathered into unusual-appearing formations and below are a series of shales and sandstones forming a thousand foot eastern escarpment. The way in which the rock strata are inclined along present day Lookout and Sand mountains is evidence of former valleys (synclines), thus earning them the title of "synclinal ridges." These former ridges, or anticlines, have eroded away and today are now lower than the valley floor (Wharton 1978:108).
Environment of Northwest Georgia
The Ridge and Valley Province in Georgia represents a relatively small segment of a belt of Paleozoic age sedimentary rocks that extends from New York state to the Fall Line of Alabama (see Figure 4). Millions of years of geologic processes have resulted in faulting, folding, and erosion which have formed the prominent parallel ridges and valleys that define the region today. The types of faunal and floral communities of the region are dictated in large manner by geologic features such as the type of bedrock (limestone or sandstone) and elevation (Wharton 1978; Johnson and Sellman 1977, 1979).
Within the province, elevations of valley floors range from 200 to 250 m above sea level with ridge crests of the Armuchee District rising to as much as 500 m above sea level. The flat-topped crests of the Cumberland Plateau attain maximum elevations of as much as 600 m, and are faced with escarpments that may be as much as 300 m high (Clark and Zisa 1976; Wharton 1978). The variation in topography within the Ridge and Valley province has resulted in the establishment of a number of distinctive environmental areas as forest regions. Wharton (1978) suggests that as many as ten distinctive forest communities may be represented. These are described as bluff and ravine forests, forests of colluvial flats, the Coosa flatwoods, submesic ridge and slope forests, the forests of the Chickamauga Valley, deciduous forests of the Great Valley, oak-pine forests of the Great Valley, Armuchee Ridges forests, ravine, gorge and cove forests, and cedar glades (Wharton 1978:112-120).
Modern Environment
The climate of Northwest Georgia is temperate with long and hot summers, and relatively short and mild winters. Across the area, the frost-free season ranges from 200 to 215 days. Rainfall is moderate, averaging four to five inches per month with December through March being the wettest months (Bramlett 1965, Tate 1978). Throughout Georgia, Tennessee, and Alabama, the native flora and fauna of the Ridge and Valley are generally viewed as being similar to those found in the adjoining areas of the Appalachians and Piedmont. Wetmore (2002:251) makes the interesting observation that species diver-
12

sity in the Ridge and Valley is somewhat lower because there are fewer microenvironments and the terrain is more homogeneous. Wharton (1978:123) also makes some interesting observations with respect to several plant and animal species found in the Great Valley that are closely related to types found in the Coastal Plain. Wharton (1978:123) views the Great Valley as a wide and important corridor linking the Coastal Plain and areas to the north.
An excellent description of the river systems and hydrology of the Ridge and Valley region is provided by Hally and Langford (1988:6-11). That volume may be referenced for more detailed information concerning this subject. As would be expected, the major streams of the Ridge and Valley generally flow in the same northeast/southwest pattern as the ridge systems. Exceptions are the Coosawattee and Etowah Rivers that flow in a nearly due west direction. Most of the streams of the province are drained by the Coosa River and its tributaries (Etowah, Conasauga, Coosawattee, Oostanaula,) which are a part of the Alabama River system. Because the Coosa River represents the dominant river system in the region, the Great Valley is frequently referred to locally as the Coosa Valley (Hally and Langford 1988:2). The exception to this drainage pattern is found in the extreme northwestern corner of the state where Chickamauga Creek and Lookout Creek flow north into the Tennessee River. In general, the stream valleys are wide enough to allow meandering. Hally and Langford (1988:9) note that because of fine sediments, levees within the Ridge and Valley are generally poorly developed except in areas along the Cartersville Fault where coarse sediments create larger levees. Within the Ridge and Valley Province, the streams in the Great Valley are large but, because of the smaller watersheds in the valleys to the west, the valleys are smaller. Within the province, large rocky shoals are limited to a 15 km stretch of the Etowah River east of Rome and are nearly non-existent along the remainder of the major rivers. Extensive shoals do occur along the major rivers where they cross the Cartersville Fault (Hally and Langford 1988:9).
The soils of the province are derived from the folded and faulted sedimentary rocks of the Paleozoic era which consists of shales, limestones, and sandstones (Bramlett 1965). Based on a review of available soil surveys for the region, Hally and Langford (1988:9) recognized eight major soil groups or associations (Bramlett 1965; Perkins and Shaffer 1977; Tate 1978). With respect to productivity, the upland and valley soils of the Chickamauga and Great Valley are moderate to good for farming while those of the Armuchee Ridges are poorly suited for agriculture. The floodplains of the major streams are generally quite productive and are extensively cultivated. Hally and Langford (1988:10) support the concept that the most productive floodplain soils occur within 10 to 15 km of the Cartersville Fault where fresh deposits of nutrient-rich alluvium are periodically deposited. This would suggest that the most productive floodplain soils would be present near the fault on the Etowah, Conasauga, and Coosawattee Rivers. This observation is significant in terms of both modern and prehistoric agriculture.
The most extensive floodplain deposits are associated with the Coosa River and the major tributary streams (Etowah, Coosawattee, and Conasauga). Except along portions of the Etowah and Conasauga Rivers where narrow valleys exist, the floodplains are relatively broad and uniform in size. However, west of Rome in the Weiss Reservoir area, these floodplains expand greatly and this trend continues into Alabama (Hally and Langford 1988:10-11, Figure 6). By contrast, the streams of the Chickamauga Valley and Armuchee Ridges have only limited amounts of floodplain deposits due to their small size and limited drainage areas.
13

Environmental Conditions During the Late Archaic to Early Woodland Transition
The interval of geologic time that corresponds to the Late Archaic through the Early Woodland periods extends over portions of the middle and late Holocene epoch. Beginning about 5,000 years ago, the region's climate became slightly cooler and possibly wetter when compared to the preceding climatic interval. Paleo-environmental studies suggest that essentially modern climate and forest conditions were established by 4,000 years ago in Northwest Georgia (Delcourt and Delcourt 1985; Watts 1975) with oakchestnut forests covering much of the area during Woodland times (Braun 1950).
As prehistoric populations increased and settlements became more permanent, the surrounding environment was changed. Research conducted in the North American Midwest (Smith 1987, 1992) suggests that some potentially important changes occurred as early as 6,500 years ago when prehistoric peoples began to favor specific locations for settlement, thus creating disturbed habitats which favored the spread of certain weedy species of plants such as lambsquarter, marshelder, and sunflower. A similar level of environmental disturbance may have also occurred in upland settings through efforts to increase the yields of mast trees (Gardner 1996). In both cases, the aboriginal use of fire would have resulted in significant alterations (Williams 2003). While horticultural traditions in the Southeast appear to have lagged behind those of the Midwest (Gremillion 1996a:103), changes in the natural environments caused by these initial steps at plant domestication in the Eastern Woodlands would have gradually increased through time. By the end of the Woodland period, widespread changes in forest composition caused by aboriginal land clearing would have significantly affected the region's environment (Cridlebaugh 1984).
Multidisciplinary research in the Little Tennessee Valley area has documented a 10,000 year record of human impacts on the ecosystems of the region (Chapman et al. 1982). The data that is applicable to the Northwest Georgia region suggest that during the early and mid Holocene (approximately 9,500 to 3,000 B.C.), the landscape was covered by closed canopy deciduous forests. Landscapes changed during the Late Holocene as a mosaic gradually developed composed of crop lands near permanent settlements, successional forests invading abandoned fields and areas of timber extraction, and deciduous forest remnants on high terraces and bedrock interfluves (Chapman et al. 1982:118). Changes of this magnitude would have occurred well after the Early Woodland period in Northwest Georgia but the establishment of more permanently occupied settlements during the Late Archaic and Early Woodland periods would have resulted in changes to the local environmental at some level.
Short term global climate changes on prehistoric populations represents another issue to be considered (Anderson 2001, Anderson et al. 2007; Kidder 2006; Little 2003; Tanner 1992). Tanner (1992:732) suggested that during the late Holocene, periods of warmer global climate would be accompanied by higher sea levels and wetter conditions inland. During the warmer periods, the climate in areas such as Northwest Georgia would have been dominated by moisture rich air flows that spread westward from the Atlantic Ocean. By contrast, during the intervals when sea levels dropped, the climate of the region would have been drier than today. These climate changes would have affected stream flows and consequently the availability of certain plant and animal species. Little (2003:21) suggests that the periods of increased precipitation, perhaps accompanied by increased sediment stream-loads, may have reduced the availability of certain mollusk species. Early Woodland shell middens have received little attention in Northwest Georgia, but some have been identified in the Coosa River drainage in eastern Alabama (Little 2003:22). The Garfield site (Milanich 1975a) may be an example in Northwest Georgia.
14

Little's model of late Holocene climate fluctuations and cultural change is based on connections between sea-level oscillations, shellfish utilization by Woodland populations in the Tennessee Valley of North Alabama and East Tennessee, and global climate change (Little 2003:10). Archaeologists working in the Tennessee Valley have long noted the differential rates of accumulation of shell middens which are high during the Early Woodland, low during the Middle Woodland and high again during the Late Woodland (Faulkner and Graham 1966b:130; Lewis and Kneberg 1959:166; Walthall 1980:219). Lewis and Kneberg (1966:166) suggested that declines in shellfish use were related to elevated stream levels accompanied by extended periods of increased rainfall. In essence, during periods of cooler and drier climatic conditions, shellfish could be procured more easily and shell middens would accumulate. Based in part on the numerous large subterranean pits found on Kellog phase sites in the Coosa Valley, Little (2003:25) suggests that the Early Woodland period coincides to a cooler and drier period which resulted in a shift to more of a collector-based economy that included changes requiring an increased dependence on storage and an increased use of shellfish. Little (2003:27) further suggests that the effects of a more favorable climate change may be related to the rise of Povery Point during the Late Archaic and Hopewell during the Middle Woodland. The emergence of these sociopolitical organizations coincided with late Holocene global warming, which created extended growing seasons and increased precipitation in the Southeast. In Northwest Georgia, these conditions should have been quite beneficial to the huntergatherers of the Late Archaic period with respect to a foraging-based economy.
Kidder (2006) examined the results of climate change during a period of time referred to as the Archaic to Woodland transition (ca. 3000-2500 cal B.P.). While the focus of Kidder's research was the lower Mississippi River Basin, the entire drainage was considered which includes the extreme northwestern corner of Georgia. Kidder concluded that throughout the basin, this transition is marked by an occupation hiatus caused by an abandonment of river bottom sites accompanied by changes in settlement and material culture organization (Kidder 2006:195). Kidder suggests that global climate changes led to increased flooding and greater magnitudes of flooding during the shift from the Late Archaic to the Early Woodland. As a result of this increased flooding in the Lower Mississippi basin, floodplain settlements were abandoned and that led to the demise of the Poverty Point culture (Kidder 2006:195). For reference, Anderson (2001:164) discusses the implications of a severe global temperature downturn known as the "1159 B.C. event" that may have contributed to the collapse of the Late Archaic cultures.
The manner in which Kidder's model may be applied to most of Northwest Georgia should be pursued in future research relating to the Late Archaic to Early Woodland transition. The limited data available at present for the region suggests that Terminal Archaic settlement is characterized by sites located in floodplain settings that seem to display widely dispersed artifact scatters and have few subsoilintruding features. This very tentatively recognized pattern could be interpreted as repeated, short term, occupations on landforms not always conducive to occupation because of flooding.
To summarize, modern environmental conditions were in place in Northwest Georgia by the beginning of the Late Archaic period. Major tributary floodplains would have become stabilized by this time and would have been conducive to settlement, at least on a seasonal basis. Minor changes in climate would have occurred that extended over many decades or a few centuries and subsequent environmental changes would have affected the availability of some food sources. Prehistoric people would have adapted to these environmental changes. These adaptations should be reflected in the archaeological record as differential intensity of occupation of specific habitats, and changes in material culture.
15

Prehistoric Lithic Raw Material Resources
A variety of lithic raw materials are found on archaeological sites in Northwest Georgia. Ridge and Valley chert represents a lithic resource that was heavily utilized by prehistoric people. Other lithic materials include quartz, quartzite, greenstone, slate, shale, sandstone, and soapstone. Sources of many of these raw materials were located and described in a recent study conducted as a part of the data recovery excavations of site 9GO231 in Gordon County (Jones 2008). Comparable studies had been conducted earlier in the region. Goad (1979) conducted a rather general study of the entire Ridge and Valley province as a part of a much larger study. Ebright (1986) conducted an intensive study of a specific project area in Floyd County. The more important raw materials of Northwest Georgia will be briefly examined with respect to occurrence and prehistoric utilization.
Ridge and Valley Chert Sources
With respect to the subject of Ridge and Valley chert, Goad's (1979) Chert Resources in Georgia, remains the standard reference for lithic studies in the state. Goad's treatment of Ridge and Valley chert sources draws heavily from the work of Cressler (1963; 1964a; 1964b; 1970) and others. This durability speaks well for her work, but the lack of intervening research reflects poorly on the interest in lithic materials as ongoing research.
Similarities in the different chert types in Northwest Georgia makes distinguishing them difficult and the material is often lumped into groups based on geologic source. The cherts of the Ridge and Valley Province are of Paleozoic Era deposition, primarily the Upper Cambrian, Ordovician and Mississippian Systems. They occur sporadically near the southeastern edge of this province, increasing in abundance in the northwestern portion. Ridge and Valley chert occurs as discontinuous beds within the Conasauga, Knox Group, the Newala Limestone, the Armuchee, and the Fort Payne chert formations. They also occur as nodules within these formations and as redeposited nodules, blocks, and shatter, especially in the Conasauga formation. Ridge and Valley chert is predominantly black, gray, or blue-gray and red. Yellow, green, blue-white, white, and purple may occur as secondary colors. Heat treatment darkens the black or gray chert and increases the intensity of the other colors and makes the luster glossy. Chert from the Conasauga, Knox, Newala, and Fort Payne formations was available in sufficient quantities for raw material exploitation by aboriginal populations. Chert of the Knox Group and the Fort Payne formations is most abundant and seems to have been most extensively utilized by prehistoric populations (Goad 1979:18).
Chert Resources in the Great Valley. Goad (1979:18) suggested that chert was distributed sporadically near the southeastern edge of the Ridge and Valley Province and increased in abundance in the northwestern portion. This assessment relegates the Great Valley to minor status as a chert source and may not be entirely accurate. Goad's observation area deserves careful evaluation, and must be interpreted within the context of lithic usage, not just in the Ridge and Valley, but in a broader context that includes adjacent portions of the Piedmont and Blue Ridge. When viewed as an isolated entity, the Ridge and Valley in Georgia does appear to have a central area of mixed, chert-bearing Paleozoic rocks corresponding approximately to the Armuchee Ridges section. The abrupt appearance of less fractured, goodquality Mississippian-age nodular chert in the Armuchee Ridges compares rather starkly with the fragmented chert of the Great Valley.
16

Apart from Goad's (1979) investigation, little is known about chert resources in the Great Valley section of Georgia. A study of chert characteristics from investigated sites in the Haig Mill Reservoir in Catoosa County (Elliott et al. 1993a, 1993b) illustrates the range of variation of material that would have been available along the western edge of the Great Valley. The Haig Mill study focused upon surface characteristics of the material (color, texture, luster) which allowed sorting of chert into as many as 35 varieties (Elliott et a. 1993b:94). A major finding of the study was that local chert sources were used more frequently by Late Archaic people, while better quality cherts used by the Late Woodland/Early Mississippian people were being brought to the sites from more distant sources (Elliott et al. 1993b:147).
Perceptions about increased availability of chert to the north and west of the Great Valley are influenced by two circumstances. First, owing to less intense tectonic action near the Cumberland Plateau, these cherts are more intact, although like most cherts within the region they are fractured to some degree (the Mississippian system in particular yields more intact nodules). Larger, intact nodules allow for a much more systematic and formal tool production scheme. Second, increasing elevation and an area of sandstone ridges to the northwest creates an increasingly inaccessible landscape and the lithic resources are concentrated in drainages and valleys. Combined with the incidence of better raw material package form, this creates an impression of greater abundance in areas to the west of the Great Valley.
The perception that chert was being procured in prehistoric times from areas to the west must be weighed carefully against the actual occurrence and utilization of chert in other portions of the Ridge and Valley, specifically the Great Valley. Although alluvium blankets much of the eastern part of the Great Valley, chert residuum is very common in soils derived from the Knox and Conasauga groups (Bramlett 1965). It is true that much of the early/middle Paleozoic chert in the eastern portion of the Ridge and Valley occurs as highly fragmented, disk-like nodules which would be difficult to use in all forms of tool making. The chert ranges in color from vitreous black and blue-black to opaque gray, with a thin, gritty cortex. Some is oolitic, while other varieties are banded or mottled with gray or white. Thick beds of Knox chert do occur in other areas (Jones 2008:68).
The proximity of the Great Valley to the Piedmont and Blue Ridge makes this the most likely source for many of the chert artifacts found in these regions. The highly fractured nature of much of this chert would predispose it to a deformalized lithic strategy. It should also be noted that, while Ridge and Valley cherts are very different from the vein quartz commonly used in the crystalline provinces, the reduction strategy for these fractured cherts is similar to that for quartz. To summarize from a previous work, this strategy involves reducing the fractured raw material to the most solid constituent pieces, and then grading these for suitability for further reduction (Jones 2006a:34). Ridge and Valley chert artifacts from sites in the Great Valley, Piedmont, and Blue Ridge consistently display this pattern.
Chert Resources of the Armuchee Ridges. Upon exiting the Great Valley, the first major expanses of Mississippian age rocks are encountered in this region. The synclinal structure results in ridges of Silurian (Red Mountain) sandstone, with formations of Devonian and Mississippian age immediately to the east. This results in numerous chert deposits containing various Mississippian cherts, as well as Armuchee (Devonian age) chert. These cherts are typically nodular with a chalky cortex. Chert of the Armuchee Ridges section varies in color, but black and gray varieties are common, sometimes with a brownish hue. Chert deposits located in the south end of Houston Valley in Whitfield County are dark
17

and translucent, with intricately lacy patterns visible in thin pieces (Jones 2008:68). In the northern part of this section, Mississippian rocks diminish in amount, with rocks of the Chickamauga and Knox groups in greater evidence. Little is known about chert deposits in this part of the Armuchee Ridges.
The Armuchee Ridges section displays a northeast to southwest ridge structure that characterizes the rest of the Ridge and Valley province, but distortions along the ridge axis interrupt this linear structure. This is especially evident in the southern portion. The gaps that break up the ridges create a more accessible upland landscape than is seen in areas with long, continuous ridge tops. This characteristic, combined with abundant, good quality chert resources, a diverse range of habitats, and ready access to the Coosa River drainage in the Great Valley makes this section a highly habitable and utilitarian landscape. An earlier study by Ebright (1986) from the Little Texas Valley in the southern section provides a good overview of chert resources from the Floyd County area. Archaeological investigations conducted near the western edge of the Armuchee Ridges District in Catoosa County (Deter-Wolf and Karpynec 2005) provide additional information relating to prehistoric utilization of chert. Within the Catoosa Training Center project area, locally available sources of chert were recognized as being extensively utilized. Based on recovered diagnostics from these sites, the primary occupation of the area was during the Early Woodland time period (Deter-Wolf and Karpynec 2005:130).
Chert Resources of the Chickamauga Valley. With respect to the major length of the Valley (the Little Chickamauga/Chattooga River valleys), little can be said with certainty about chert in the Chickamauga Valley section. The main swath of the valley in the vicinity of LaFayette appears to contain only poor-quality, massive white chert of the Knox group (possibly the Chepultepec variety). A thin band of Newala limestone occurs at the edges of the valley throughout much of the length of this section, but no chert sources are reported. To the west of Lafayette, small pebbles of brown and gray chert occur occasionally in the drainages of Straight Gut Valley, but these may be residual Mississippian-age material from the slopes of Pigeon Mountain to the west (Jones 2008:68).
Pigeon Mountain and the eastern flank of Lookout Mountain figure prominently into the chert resources of the northwestern extension of that portion of the Chickamauga Valley created by West Chickamauga Creek. It is therefore difficult to separate the chert resources of the eastern Cumberland Plateau section from those of the northwestern Chickamauga Valley. Extensive exposures of Mississippian age rocks (many of them potentially chert-bearing) surround the long, spur-like northern end of Pigeon Mountain. This would make chert available as residual deposits along the western edge of the Little Chickamauga drainage. These rocks continue southward along its western slopes to the head of McLemore Cove and turn to the north along Lookout Mountain. Although bedrock exposures are often difficult to access, drainages in this area often produce large concentrations of good quality chert. Much of this chert shows an affinity with the Fort Payne variety, yet the range of colors, patterns, and textures suggest others such as St. Genevieve and St. Louis. The Newala chert sources listed by Goad (1979: Appendix I) for Walker and Catoosa are from the West Chickamauga Creek valley.
Chert Resources of the Cumberland Plateau. With respect to the Cumberland Plateau that is exclusive of that portion adjacent to the Chickamauga Valley, chert resources (in Georgia) are somewhat difficult to assess. Geologically, chert-bearing formations are widely exposed, and chert is abundant in specific areas. Massive sandstone-capped ridges create large expanses that would likely contain little
18

chert, and extreme topography creates special problems for the study of settlement patterns and resource usage. It is generally believed that chert sources would be concentrated in lower elevations near drainages. Long, nearly impassable ridges complicate human (and animal) movement from drainage to drainage. Extremes of topography and the relatively small amount of accessible and habitable terrain give the impression that chert occurs here in greater quantity than elsewhere in the Ridge and Valley. Goad (1979: Appendix I) lists Ordovician (possibly Newala) and Mississippian (Fort Payne) chert in this area as occurring principally within the narrow confines of Lookout Valley.
Past archaeological surveys in the Lookout Mountain Valley provide relevant information about chert availability in the region. That work has focused upon interpretations of lithics with respect to settlement patterns rather than descriptive analysis or raw materials (Jefferies and Smith 1975; Jefferies 1982, Smith et al. 1986b). Jefferies (1982:102) noted that 99 percent of the flaked stone artifacts collected during the University of Georgia's Lookout Valley survey in Dade County was made from local chert sources. The local chert was attributed to the Fort Payne and St. Louis limestone formations which are closely associated and display a linear distribution along the valley floor and lowest ridges. Jefferies concluded the chert deposits used by the prehistoric inhabitants of the valley were located on or near the valley floor (Jefferies 1982:103).
Other Lithic Raw Material Resources
Several other lithic raw materials deserve mention because of their importance in prehistoric tool production. Certain raw materials such as quartz and quartzite occur primarily within the Piedmont but may appear on sites within the Ridge and Valley in moderate amounts during various periods in prehistory. Also, some quartz and quartzite may occur in stream gravels within the Ridge and Valley. Materials such as greenstone and soapstone also originate in the Piedmont and Blue Ridge but regularly appear on sites in the Ridge and Valley province. Materials that originate in the sedimentary deposits of the Ridge and Valley include shale, slate, and abrasive materials such as sandstone. With respect to archaeological studies, there has been only a limited amount of previous investigation of the sources of most of these raw materials. Sources of some of these materials were recorded as part of the data recovery investigations at site 9GO231 (Jones 2008). Specific locations referenced in the following paragraphs may be attributed to that study.
Quartz is particularly common on sites located near outcrops along the eastern boundary of the Piedmont and the eastern edge of the Great Valley. Studies in Georgia suggest that quartz used for the production of chipped stone tools may be separated into a series of categories that range from a pure glass-like crystal form to a grainy material that may be indistinguishable from quartzite (Jones 2006a:2930). As a mineral originating largely as a result of igneous and metamorphic processes, primary sources of quartz in Georgia are limited to the Piedmont and Blue Ridge.
To date, no large veins or exposures of particularly high-quality quartz have been documented in the immediate vicinity of the Cartersville Fault, but cursory examination of residual quartz float at scattered localities reveals that small pieces are common. It is also possible that quartz material is available in the Great Valley, in the form of gravels from streams and rivers that drain the Piedmont and Blue Ridge.
19

Quartzite as used here, and in much archaeological literature, represents shorthand for metaquartzite, a rock type that may be broadly defined as quartz sandstone within which the individual grains are fused through metamorphism. Quartzites are often massive, a distinction that is helpful in separating them from sugary varieties of quartz. Quartzite is available in close proximity to the Great Valley portion of the Ridge and Valley in Georgia. While smaller outcrops may be present, the Georgia Geological Map (1976) shows two large outcrops in the Piedmont, both to the northeast of Cartersville. Closest to the Cartersville Fault is the outcrop that comprises Pine Log Mountain. Two smaller extensions of this outcrop extend southward from Pine Log Mountain into the vicinity of Lake Allatoona. Northeast from Pine Log Mountain is the Oakey Mountain outcrop locality. Although it lies at a greater distance from the Cartersville Fault than the Pine Mountain outcrop, Oakey Mountain is nonetheless a possible source for quartzite during the Late Archaic.
Drainages that enter the Great Valley from the adjacent Piedmont are possible sources of quartzite, as well as other crystalline lithic materials, in the Ridge and Valley. Within much of the Appalachian region, quartzite is closely associated with the production of large stemmed, broad blade projectile points of the Late Archaic period (Whyte 2007:138). Relatively few examples of quartzite Savannah River Stemmed points have also been found within Great Valley (Benson et al. 2008:44).
Greenstone is a generic word, which, while widely used by archaeologists and collectors alike, has little intrinsic meaning. However, it generally denotes a tough, greenish rock containing little quartz that is soft enough to be worked with relative ease by a process of flaking, pecking, grinding and/or polishing. Specifically, it includes metabasalts, fine-grained amphibolites, and other metamafic rocks. These rocks are the product of varying degrees of metamorphism of iron- and magnesium-rich rocks comprised chiefly of amphibole. Thus, greenstone covers a variety of metamorphic rocks, all of which originate in the Piedmont or, further to the north, in the Blue Ridge. Though few large outcrops of such material are indicated on the Georgia Geologic map (1976), metamafic rocks are scattered throughout the crystalline provinces within reasonable distances from the Ridge and Valley. As with other crystalline rocks, streams and rivers may be sources of material originating in the Piedmont.
Slate and shale are sedimentary rock composed of clay-like particles. Slate is metamorphosed shale. Both rocks are characterized by strongly foliated (layered) structure. Although slate is derived from shale and shares the layered structure, the layers exhibited by slate often are not the same as the bedding planes of the parent shale. This juxtaposition of layers is the result of metamorphic pressure during the formation of slate. It also gives rise to banded slates, in which the original bedding of the parent shale deviates from the orientation of layers in the slate.
Within the Ridge and Valley itself, shales are quite common. With a few exceptions, shales are generally soft and heavily weathered, making them unsuitable for most cultural uses. To the northwest along the edges of the Cumberland Plateau in Georgia is found a compact, black shale tentatively identified as Floyd Shale. This shale occurs in Mississippian-age rocks, with in situ deposits containing abundant thinly bedded and friable material. In the drainages investigated in Walker County, however, thick slabs (upwards of 3 cm) of solid material may be found. These residual slabs apparently consist of single layers with no intervening foliation. Though soft, it may be chipped into suitable forms, which can then be ground to the desired shape.
20

For Northwest Georgia, many other slate/shale resources are likely available along the periphery of the Ridge and Valley. It is also likely that certain varieties were sought after for particularly useful properties. The production of slate digging implements or "hoes" are commonplace on Early Woodland sites in Northwest Georgia and have been associated with Terminal Archaic occupations in Tennessee (Keel 1978:154).
Abrasives such as sandstones are common in the Ridge and Valley of Georgia, although in the Great Valley they are relatively scarce. In the Cumberland Plateau, sandstones of Pennsylvanian and Mississippian age are common. Though massive, these sandstones can be obtained in a wide range of sizes. Outside of the Cumberland Plateau, Silurian sandstone of the Red Mountain formation is common and is widespread in the Armuchee Ridges section. This sandstone is available in thinly bedded pieces with durable grit. In the Great Valley area, high-quality sedimentary sandstone is scarce. In this area, as in the adjacent Piedmont and Blue Ridge, crystalline rocks are the most likely source for abrasive materials. Slabs of a variety of rock types may be used in conjunction with wet sand to create stationary grinding surfaces.
Soapstone (steatite) will be examined in somewhat greater detail because, like chert, it represents a highly significant lithic resource during the Late Archaic period and to a lesser, but still important, extent during the Early Woodland period. Much of the following information is taken from a study of soapstone resources conducted as part of the data recovery investigations of site 9GO231 in Gordon County (Jones 2008:91-95), and the results of that study may be consulted for a more detailed examination of the subject. Site 9GO231 is significant because it represents one of the few Ridge and Valley sites with relatively high numbers of soapstone vessel fragments that have been assigned to the Terminal Archaic period (Benson et al. 2008:31-33).
Soapstone is massive, impure talc that was utilized extensively throughout the Southeast for the production of bowls, atlatl weights, gorgets, and perforated slabs. Historically, it has been used for laboratory counters and sinks, wood stove parts, and heating devices. Though steatite and soapstone are approximately synonymous, steatite specifically refers to massive talc, while soapstone is applied somewhat less precisely to a variety of massive talc-like rocks. While steatite gives the impression of technicality, soapstone is perhaps a better word for archaeological use because it better describes the range of soft rocks used prehistorically.
Soapstone belongs to a group of rocks consisting predominantly of magnesium and iron silicates. These rocks, along with chlorite and serpentine, are collectively referred to as ultramafic rocks. The various ultramafic rocks commonly encountered in mineral resource maps and reports are best defined by Sorrell (1973) and Whitten and Brooks (1972). Besides soapstone, these utramafic rocks include talc, talc schist, chlorite shist, and serpentine (Jones 2008:91-92).
The distribution of soapstone outcrops along the Appalachian Mountain chain and within the Piedmont has been widely studied by both geologists and archaeologists. Soapstone occurs at widely scattered localities throughout the state. Outcrops occur in clusters in some areas, while other outcrops appear to be isolated occurrences (Figure 5).
21

The map depicted in Figure 5 shows the locations of known commercial soapstone deposits of the Eastern U.S. as depicted in an archaeological study of a prehistoric bowl quarry at Soapstone Ridge near Atlanta (Elliott 1986: Figure 2). The map of soapstone deposits provides an excellent picture of the distribution of source material and confirms its absence in the northwestern corner of Georgia. The map is based on a larger map that shows the distribution of soapstone throughout the Eastern U.S. (Chidester et al.1964).

According to

Elliott (1986:9), soap-

stone bowl quarries are

known from Carroll,

Clayton, Cobb, Co-

lumbia, DeKalb,

Greene, Elbert, Fulton,

Figure 5. Map showing the distribution of soapstone deposits in the Eastern U.S. (adapted from Elliott 1986: Figure 2).

Jackson, Lumpkin, Union, White, and

Wilkes Counties in

Georgia. Quarries are also reported from Banks (Jordan and Webb 2003) and Paulding counties (Cantley

1998; Cantley et al. 1998). Soapstone bowl quarries in Alabama have been reported from Tallapoosa and

Talladega Counties (Wright 1974:17; O'Donoughue and Meeks 2007a). The small soapstone bowl

quarry recorded in Paulding County (Cantley 1998:56) represents a rare instance of confirmed aboriginal

exploitation of an outcrop located along the western edge of the Piedmont in Northwest Georgia. Byrne

(1987:31) has stated that outcrops of soapstone occur farther north in Murray County but prehistoric

quarry sites have yet to be recorded.

22

The study of soapstone in Georgia is greatly aided by two maps produced by the Georgia Geological Survey. The Geologic Map of Georgia (Georgia Geological Survey 1976) shows ultramafic rock outcrops, but does not differentiate between the various types. The Mineral Resource Map of Georgia (Georgia Department of Natural Resources 1969) distinguishes between talc, soapstone, and asbestos. This map should be used cautiously when investigating potential prehistoric soapstone sources, since it is based largely on commercially exploited resources and all soapstone/steatite sources have not been quarried historically. An earlier geologic map produced specifically for Georgia also shows the distribution of soapstone and related materials within the state (Hopkins 1914).
Hopkins (1914) shows a discontinuous band of ultramafic deposits (including talc, soapstone, and chloritic schist) extending from Rabun County southwestward to Carroll County. This band is largely confined to an area that corresponds closely to the Gainesville Ridges, a physiographic district that is dominated by the Brevard Shear Zone. Outliers occur up to approximately 30 km from the main band. This accounts for deposits in Banks County, Lumpkin County, Cherokee County, and Paulding County (Cantley et al. 1998, Jordan and Webb 2003). The soapstone quarry sites previously excavated within the Soapstone Ridge complex of DeKalb County (Dickens and Carnes 1983, Elliott 1986, Webb 1993a) may also be related to ultramafic rocks of the Gainesville Ridges. This band probably gives rise to soapstone outcrops and quarries (1TA15 and 1TA21) known to exist in the Piedmont of Alabama (Wright 1974, O'Donoughue and Meeks 2007a). This is of interest to the present study since soapstone of the Brevard Zone/Gainesville Ridges association yield abundant bowl quarries.
Ultramafic rocks occur commonly in an area east of the north-south trending portion of the Cartersville Fault (see Figure 4 for referenced area). This is confirmed by the state geologic map, the state mineral resources map, and locally procured research data. The state geologic map shows a string of ultramafic outcrops in this area and many of these are known talc mines. Although this area was initially believed to be a probable source of soapstone, and unsubstantiated references to soapstone outcrops do exist (Byrne 1987:31), it appears that these outcrops consist mostly, if not entirely, of foliated talc (Jones 2008:93).
The possibility that soapstone deposits suitable for vessel production exist in the areas from Cherokee County southward and westward into Paulding County is suggested by the Georgia mineral resource map which shows one soapstone deposit near Holly Springs and by limited survey data. In a previous archaeological study, Webb (1998:371) reported that possible bowl-quality soapstone was known from the Ball Ground area. Ledbetter et al. (1987:214) reported soapstone vessel fragments from 38 sites in Allatoona Reservoir, although no prospective quarry sites were identified (one site from Allatoona, 9CK638, did produce a number of small chunks of fine green soapstone showing no sign of modification). Based upon published information in Hopkins (1914:281) and unpublished data provided by Dan Elliott (personal communication 1986) the authors suggested that outcrops could be found 15 to 20 km south of Allatoona (Ledbetter et al. 1987:214). Recent work in northwestern Paulding County has now confirmed the presence of at least one soapstone bowl quarry within an area of historic talc mining that is even closer to Allatoona Reservoir (Cantley 1998, Cantley et al. 1998). While the possibility must be considered that many vessels originated at Soapstone Ridge or some other deposits in DeKalb County, research in the Allatoona area provides insight into the possibility that at least some vessels found in Northwest Georgia were manufactured from nearby source material.
23

Just west of Lake Allatoona, the Cartersville Fault turns sharply to the west. As it cuts across portions of Paulding, Polk, and Haralson Counties, the fault abuts the northern edge of the Piedmont, which provides a rich array of metamorphic rocks. With minor exceptions (Cantley et al. 1998), soapstone resources along the southern edge of the Cartersville Fault are poorly researched. For instance, the map by Hopkins (1914) shows a soapstone deposit in north central Paulding County, yet the Georgia Geological Map shows no ultramafic outcrops in that area. Although considerably south of the area under discussion, soapstone in Carroll County presents a similar set of problems. Elliott (1986:9) notes that soapstone was quarried prehistorically there, yet neither of the Georgia maps shows ultramafic rocks in the county. The Carroll County prehistoric soapstone quarry is recorded as 9CL18 (Sheldon 1973).
Because soapstone does not occur naturally in the Ridge and Valley Province of Northwest Georgia, the movement of vessels into the region presents a unique opportunity to study Late Archaic, Terminal Archaic, and Early Woodland resource exploitation. And while most of the ultramafic material that lie in close proximity to the Cartersville Fault is probably unsuitable for bowl manufacture, bits of scattered information suggest that some suitable soapstone does occur among these rocks. The sources of the material discussed by Webb (1998:370-371) from Cherokee County and by Cantley and colleagues (1998) from Paulding County should be further assessed with respect to potential usage by aboriginal peoples. This assessment should also be extended to include (or specifically exclude) other sources along the Cartersville and Great Smoky Mountain Faults. Based on the one example provided by Cantley et al. (1998) there is the possibility that other minor soapstone bowl quarries once existed near the edge of the Ridge and Valley but have been obliterated by historic mining. Although the relationship between widely separated outcrops of the material is largely unknown, the occurrence of soapstone within well-defined geological belts is potentially informative, and functions as an organizational tool for discussions of soapstone exploitation.
24

Chapter 3. A Review of Previous Archaeological Investigations in Northwest Georgia and Adjoining Areas

A review of previous research in Northwest Georgia and selected sites from nearby portions of Alabama and Tennessee will primarily focus on testing and data recovery projects that produced information pertaining to the Late Archaic, Early Woodland, and transitional Late Archaic to Early Woodland time periods. A more limited discussion will be provided for a few large scale surveys and the compilation of information provided in the Georgia Archaeological Site Files database. The earlier work of Robert Wauchope, Joseph Caldwell, Dan Morse, David DeJarnette, and a few others will be examined in detail because their work laid the foundation for this study. The discussion of the more recent work will be grouped according to geographic areas, primarily by river drainage. By discussing sites in this manner, the cultural complexity of Northwest Georgia will be highlighted. This review of investigations will begin with the Allatoona Basin in the Etowah River valley because the data recovered from that area has shaped much of our understanding of the Early Woodland period. The chapter will conclude with an examination of site data. Table 1 provides a list of the projects or specific sites that were examined in order of presentation, along with a few brief comments relating to investigated components that are relevant to the current study.

Table 1. A list of Projects and Sites to be Examined as Previous Research.

Project/Site

Relevance to the Late Archaic to Early Woodland Transition

Early projects conducted in Northwest Georgia (the foundation for modern research)

Wauchope's 1930s Survey in North Georgia (Wauchope 1966)

Extensive survey and excavation of sites of all time periods resulted in initial efforts to define a Woodland period ceramic sequence for Northwest Georgia. (Includes numerous sites from the Ridge & Valley and western Piedmont).

Smithsonian Instituition's Allatoona Reservoir project (Caldwell 1957)

Excavations at intensively occupied Early Woodland sites and a limited number of Late Archaic sites. Caldwell's interpretations form the basis of much of our current understanding of Northwest Georgia archaeology (Etowah drainage, Piedmont).

Smithsonian Institution's Buford Reservoir Project (Caldwell 1953)

Survey and excavation data related primarily to Woodland period and later occupations provided support for the idea of a Late Archaic/Early Woodland hiatus (Chattahoochee River drainage, Piedmont).

Weiss Reservoir Project (DeJarnette et al. 1973)

Extensive excavation data associated with Early Woodland occupations of the Cedar Bluff phase (several sites located in the Coosa River floodplain, R. & V. province)

Carters Dam , 9MU100 (Beasley 1972)

Excavation data related to the Late Archaic period with some inferences for Terminal Archaic occupation in the western Piedmont (Coosawatee River floodplain)

Georgia Historic Commission's excavations at 9GO31, 9CT2, and 9WD1 (Morse and Morse 2004).

Extensive excavation data that include Late Archaic and Early Woodland data. Work resulted in a series of Woodland period phase designations (the most relevant site, 9GO31, is located in the Oostanaula River floodplain, R. & V. province).

Mahan 9GO22 (Morse 1960)

Excavated feature data provide some of the oldest Early Woodland (Kellog-related) radiocarbon dates in region (Salacoa Creek/western edge of Piedmont)

Research conducted in the Etowah Valley (Ridge & Valley and western Piedmont)

Garfield 9BR57 and Wooly's Bend Excavation data from two major Early Woodland sites located in the floodplain of the

9BR59 (Milanich 1975)

Etowah River as possible rare examples of shell midden sites in region (R. & V.)

25

Lewis Sam's Bridge WGC 409 (McCluskey 1976)

Excavation data from a Woodland occupation site on a levee of the Etowah River. (excellently described example of a hunter-gatherer site pattern in region (R. & V)

9CK(DOT)7 (Bowen 1982)

Data recovery data associated with an Early Woodland habitation site on a tributary of the Etowah River in the western Piedmont.

Cagle 9CK123 (Crook 1984)

Data recovery data associated with Late Archaic and Early Woodland occupations from a floodplain site in the Etowah River valley in the western Piedmont.

Rush site 9FL164 (Wood and Ledbetter 1990)

Data recovery data from a major Early Woodland (minor Middle Woodland) habitation site in the floodplain of the Etowah River valley in Ridge & Valley.

Malbone Power Line site (Blanton 1986; Stanyard 1999)

Survey and testing data associated with a significant Early Woodland occupation site on the Etowah River (floodplain of Etowah River, Ridge & Valley province).

Route 411 Sewer Line, Rome Area Testing data associated with Late Archaic and Early Woodland occupations

Stanyard and Baker 1992)

(floodplain of Etowah River in Ridge & Valley province).

Yellow Creek Reservoir 9CK713, 9DW64, 9DW77 (Webb 1998)

Significant excavation data from rockshelters and open air site relating to Late Archaic and Early Woodland occupations in the western Piedmont (Yellow Creek).

9PA102 (Cantley et al. 1998)

Rare report of soapstone bowl quarry in the Piedmont region (Pumpkinvine Creek).

Hardin Bridge 9BR34 (Pomfret 2005; Windham et al. 2008)

Testing and data recovery data related to Late Archaic and Middle Woodland occupations (floodplain of the Etowah River in Ridge & Valley province).

Research conducted on the Lower Oostanaula/Coosa River Drainage in Northwest Georgia (Ridge & Valley)

Selman's Bend 9FL56 and 9FL61 (Alexander 1979)

Testing data related to Late Archaic/Early Woodland occupations (floodplain of the Oostanaula River).

Pumpkin Pile 9PO27 (Ledbetter 1992)

Survey and testing data associated with an Early Woodland Village with limited information relating to a Terminal Archaic occupation (tributary of Coosa River)

SW Rome bypass FL203-206 (Southerlin et al. 1996)

Survey, testing and data recovery excavations associated with Late Archaic, Early Woodland, and probable transitional period occupations (Coosa River floodplain).

Plant Hammond 9FL3, 9FL174, 9FL200 (Patton 2003; Sipes 2008)

Survey, testing and data recovery excavations provided data for Late Archaic, Early Woodland, and probable transitional period occupations (Coosa River floodplain)

Research conducted in the Upper Tallapoosa Drainage in Northwest Georgia (Piedmont).

West Georgia Reservoir (Stanyard and Pietak 1997)

Extensive survey and testing data for Late Archaic and Early Woodland occupations in the upper Tallapoosa drainage in the western Piedmont region.

Research conducted in the Chattahoochee Valley in Northwest Georgia (Piedmont)

Snake Creek sites 9CL269 and

Major excavations of two well dated Early Woodland sites with parallels to Kellog

9CL270 (Benyshek and Wild 2003) phase and Pumpkin Pile complex as defined for Northwest Georgia.

Research conducted in the Upper Coosa River Drainage in Northwest Georgia (Piedmont and R.& V.)

East Tennessee-Ball Ground Survey General survey data relating to Late Archaic and Woodland settlement in the

(Blanton et al. 1987)

Piedmont and Ridge and Valley.

Bertha Perry 9MU8 (Byrne 1987)

Testing data related to "terminal Late Archaic" and Early Woodland occupation in the Ridge and Valley in Northwest Georgia.

Swancy 9GO70 (O'Steen 1987

Excavation data related primarily to Early Woodland (Watts Bar pottery) and "terminal Late Archaic" occupation in the Ridge and Valley in Northwest Georgia.

Rich Mountain Survey (Smith et al. 1988)

Survey data that include a discussion of Terminal Archaic occupation from an upland area located at the western edge of the Piedmont in Northwest Georgia.

Haig Mill Reservoir (Elliott et al. 1993a, 1993b).

Large scale hand excavation of two sites with radiocarbon dates and other evidence of transitional Late Archaic/Early Woodland occupation in the Ridge and Valley.

26

Murray County Wetlands 9MU135 Information related to transitional Late Archaic/Early Woodland period projectile

(Gresham 2002)

point types and general site data (Ridge and Valley region).

Lick Creek 9GO231 (Benson et al. 2008)

Information related to Terminal Archaic occupation includes projectile point types, subsistence remains, and radiocarbon date (Ridge and Valley region).

Research conducted in the Tennessee River Valley in Northwest Georgia (Ridge & Valley)

Catoosa Training Center (Deter-Wolf and Karpynec 2005)

Excavated data from several transitional Late Archaic/Early Woodland and Early Woodland period (essentially aceramic) sites in northwestern corner of Georgia.

Selected Excavations from Southeastern Tennessee and Northeastern Alabama

Westmoreland-Barber (Faulkner and Graham 1965, 1966a)

Work at this site near the Tennessee/Northwest Georgia line defined the concept of a Terminal Archaic period (Tennessee River drainage in Ridge & Valley province).

Russell Cave (Griffin 1969)

Occupation and subsistence data related to the Late Archaic and Early Woodland periods in Cumberland Plateau of northeastern Alabama.

Bellefonte site (Futato 1977)

Site data related to Middle Gulf Formational stage and Early Woodland Colbert phase occupations from the Cumberland Plateau region of northeastern Alabama.

Wright Farm (Little et al. 1997)

Kellog phase occupation in Coosa drainage northeastern Alabama (Ridge & Valley)

1Ra12 (Knight 1980)

Aceramic occupation zone with artifact types characteristic of the Late Archaic period dated to 600 B.C. in the eastern Alabama Piedmont (Tallapoosa drainage).

The majority of the previous excavations in Northwest Georgia have focused on sites of the Early Woodland and subsequent periods. Until quite recently, relatively few investigations have been specifically directed to Late Archaic sites and little effort has been expended on sites of the transitional period. An effort is made to synthesize the work for each project so that Late Archaic and Early Woodland occupation in the Northwest Georiga region may be better understood and, at the same time, retrieve any information that may benefit the interpretation of the transitional time period between the Late Archaic to Early Woodland periods. The information procured from the previous research will be used to characterize the Terminal Archaic and Early Woodland periods in Northwest Georgia (Chapters 4 and 5) and to address a variety of research issues that include subsistence, raw material utilization, architecture, and site patterning. (Chapters 6 through Chapter 8).
Early Work in Northwest Georgia
The earliest archaeological work in Northwest Georgia focused on Mississippian period mound sites (Jones 1873, Stevenson 1873, Thomas 1894, Whittlesey 1883). Later researchers such as Warren K. Moorehead and Margaret Ashley continued work at Etowah Mounds and investigated other mound sites such as Little Egypt in Murray County (Moorehead 1929, 1932). According to Hally and Langford (1988:13), Ashley also investigated the smaller mounds at Plant Hammond (9FL3) and Two Run Creek (9BR3). A thorough review of the early mound research in the Etowah Valley may be found in King (2003:33-49). The first real effort to examine Archaic or Woodland components was directed by Robert Wauchope in the late 1930s (Wauchope 1966).

27

Robert Wauchope's North Georgia Survey
By the time of Robert Wauchope's North Georgia survey, a general ceramic sequence for the Woodland and Mississippian periods in the Southeast was already in place. Major federally funded archaeological projects had been conducted in Georgia, Tennessee, and Alabama during the WPA period of the 1930s and the analysis of these large collections provided the basis for ceramic classification for the region. Beginning with the first southeastern pottery conference in 1938, and continuing under the name of the Southeastern Archaeological Conference, archaeologists developed rigid classification procedures for the region's aboriginal pottery (Lyon 1996:192). Ceramic type descriptions were being published by the late 1930s for many of the known prehistoric pottery types. These type descriptions provided field archaeologists such as Wauchope with an extremely important tool for interpreting sites. For that reason, Wauchope's identification of Woodland and Mississippian period components may be accepted with a high degree of certainty. Also, because Wauchope's report was completed more than two decades after the field work, he had an opportunity to include some of the refinements proposed by Joseph Caldwell (1958) and others to the older type descriptions. However, because projectile point sequences had not been defined in a similar manner, Wauchope's identification (or lack of identification) of Archaic period sites is much more problematical. As a result, we are able to utilize much of Wauchope's survey data relating to Woodland and Mississippian period occupations, but information for earlier components is not readily interpretable.
Robert Wauchope, while stationed at the University of Georgia in Athens, began a two-year archaeological survey of northern Georgia in 1938. The survey work was conducted as part of the Works Progress Administation's (WPA) public works program with little direct field contribution from University of Georgia students. The contributions made by Wauchope to our understanding of the culture history of North Georgia remain significant today. Wauchope worked extensively in the Etowah Valley and he reported on site types ranging from the major Mississippian mound complex at Etowah to small Archaic period lithic scatters. A substantial part of his time was spent in Bartow County where he recorded information on 41 sites and directed excavations at five. Two of the excavated sites, Two Run Creek (9BR3) and Walt Jones (9BR9), contained important Early Woodland deposits that allowed him to define the first pottery sequence for Northwest Georgia (Wauchope 1948, 1950). Wauchope noted that the Two Run Creek site had been visited previously by Margaret Ashley, who was working for Moorehead at the time, and that her work was restricted to the Mississippian mound (Wauchope 1966:224).
Wauchope's collections from Bartow County and other northwest Georgia sites produced Early to Middle Woodland ceramics and a number also included projectile points that may be assumed to date to the Late Archaic and Early Woodland periods. Wauchope's field crews also recorded information from 32 sites in Cherokee County, of which four were excavated. Like Bartow County, many of the Cherokee County collections produced Early Woodland and Late Archaic artifacts, but the major excavations were directed to later occupations. Wauchope recorded fewer sites in other northwest Georgia counties. Among the counties listed in his report, five sites were recorded in Floyd, two in Murray, one in Paulding, and eight in Cobb County (Wauchope 1966:480-482). While artifacts associated with Early Woodland or Archaic components were found by Wauchope's crews in Floyd and Cobb Counties, excavation data is not reported. Of the sites recorded by Wauchope in Northwest Georgia, the excavation data from Two Run Creek and Walt Jones are most pertinent to the present study (Figure 6).
28

The Two Run Creek site (9BR3) is located between Kingston and Cassville, about 5 km upstream from the confluence of the stream with the Etowah River (Wauchope 1966:223). A mound located on the site is attributed to the Mississippian period, but the village excavations produced significant information related to Woodland occupation. Excavations in the village area Figure 6. M ap of a portion of the Etowah River Valley showing Two Run yielded some of the first eviCreek and Walt Jones sites (adapted from Wauchope 1966: Figure 154). dence for the temporal relationship of several Early and Middle Woodland pottery types. Wauchope's report does not provide pottery counts for the site, but it does provide percentages. Wauchope concluded that the lowest deposits contained the earliest pottery type in the region (Wauchope 1966:224).
The early pottery zone, which contained Dunlap Fabric Marked and Mossy Oak Simple Stamped, was seen as evidence of a "lower Early Woodland" occupation (Wauchope 1966:8). Wauchope's lower Early Woodland subdivision included Dunlap Fabric Marked, Mossy Oak Simple Stamped, and fibertempered pottery, which were the earliest pottery types thought to occur in the region at that time. A later "upper Early Woodland" was defined by the Deptford series that locally included Cartersville Check Stamped and Deptford Simple Stamped pottery (Wauchope 1966:8). Dunlap Fabric Marked pottery from the lowest levels of the site was tempered entirely with sand or grit. A very few limestonetempered, fabric-marked sherds were recovered from the plowzone (Wauchope 1966:226).
At Two Run, Wauchope distinguished the "fine scratch-like lines" of Mossy Oak Simple Stamped from the bolder execution of Deptford Simple Stamped (Wauchope 1966:47). Wauchope noted that Mossy Oak Simple Stamped only occurred in small amounts in the lowest levels, thus suggesting a co-occurrence with Dunlap Fabric Marked pottery at Two Run. Wauchope noted that on most Northwest Georgia sites the Mossy Oak and Deptford Simple Stamped types occurred together, but at Two Run there was stratigraphic separation between levels containing Mossy Oak Simple Stamped and Deptford Simple Stamped (Wauchope 1966:47). Referring to later work in Allatoona Reservoir, Wauchope noted that in the Etowah River area, the two old types were being referred to under the name "Cartersville Simple Stamped" (Caldwell 1958:45). Wauchope found more than one hundred projectile points on the Two Run site. These included significant numbers of medium-sized, triangular projectile points, many with indented bases. The triangular points were found in the lowest levels of the village excavation, but some probable Archaic stemmed points were found in the mound fill (Wauchope 1966:228). A variety of groundstone tools were found at Two Run along with three soapstone vessel fragments (Wauchope 1966:230).
29

Wauchope concluded that the Early Woodland people at Two Run built round houses at ground level, threw trash into nearby refuse pits and buried their dead in flexed positions in apsidal graves (Wauchope 1966:224). A flexed burial of an adult female was associated with the lower Early Woodland occupation. Wauchope (1966:223) also identified and illustrated the post pattern of an oval structure (House 6) that he associated with the lower Early Woodland occupation.
Wauchope interpreted the pattern as a round house 7.9 m (26 ft) in diameter. The structure contained a number of large interior posts but lacked a central hearth. The post stains averaged 0.4 ft in diameter and 1.2 ft deep and were spaced at intervals of about 1.2 ft. Wauchope noted that four of the large interior posts formed the corners of a roughly square area (Wauchope 1966:450). One burial pit, that of a flexed infant, was located within the structure. Later researchers have questioned the Early Woodland association of the structure. Garrow's (1975:20) assessment that the structure dates to the Early Woodland period has now been changed to Middle Woodland based on the presence of the pottery type now called Cartersville Simple Stamped (Garrow 2002:4). Anderson (1985:36) also viewed the structure as Middle Woodland period, but it should be noted he uses a beginning date of 500 B.C. for that period. A beginning date for the Middle Woodland period of about 200 B.C. is generally used for northwest Georgia (Garrow 2002:2). Whether the structure dates to the Early Woodland or "early" Middle Woodland period is of secondary importance to the fact that Wauchope's early work led to the recognition of one of the best defined structure patterns in Northwest Georgia.
The Walt Jones site lies on the banks of the Etowah River adjacent to a small, unnamed tributary stream (Wauchope 1966:231-233). Wauchope excavated 20 test pits and recovered a large surface collection of artifacts from this Early Woodland site. Because the site was being cultivated at the time, the test pits were placed on the slopes and all cultural material was recovered from the plowzone. The central portion of the site was not investigated. Wauchope's inventory of the artifacts includes 1,601 Dunlap Fabric Marked, 3 Mossy Oak Simple Stamped, 9 Deptford Simple Stamped and 15 Deptford Check Stamped sherds as his early Woodland pottery types. A few hundred Late Woodland and Mississippian sherds were also recovered. A very small minority of the fabric-marked sherds (n = 14) from Walt Jones were limestone tempered (Wauchope 1966:231).
In a quarterly report prepared for work conducted through September 30, 1939 (page 8) Wauchope noted that two refuse pits at Walt Jones yielded a great number of shells. In those features, 83 percent of the pottery was fabric marked, 5 percent was simple stamped, 8 percent was complicated stamped (Swift Creek), and less than 2 percent was check stamped. The surface collection also seemed to corroborate the findings from Two Run Creek where fabric-marked pottery was associated with stemless, concave-base projectile points. According to his list, 147 of 172 project points were triangular forms. Based on illustrations, the remainder appear to represent primarily Late Archaic types (Wauchope 1966: Figures 239-244). Three soapstone bowl fragments were recovered.
Wauchope made an interesting discovery some time prior to the completion of his report. A single fiber-tempered sherd was discovered in the pottery collections from the Walt Jones site that had previously gone unnoticed. In Wauchope's assessment of the site, he noted that, "This was the only fiber-tempered sherd we found in northern Georgia" (Wauchope 1966:231). In an earlier paper Wau-
30

chope had commented that fiber-tempered pottery had not been found in the Northwest Georgia collections "so far examined and I do not remember any at the time of the excavations" (Wauchope 1948:201). The recovery of a single sherd appears consistent with the extreme scarcity of the pottery type in the region.
Wauchope discovered a substantial number of Early Woodland sites along the Etowah River and smaller numbers along the Chattahoochee River and its major tributaries. Sites with fabric-marked pottery, as well as check stamping and simple stamping, were consistently found throughout his survey areas of Northwest Georgia, and Wauchope demonstrated for the first time that a significantly large aboriginal population existed well before the advent of the Mississippi period. It has been suggested that Wauchope's survey identified as many as 118 sites with components comparable to the Early Woodland Kellog phase as currently defined (Bowen 1980:41).
Wauchope was not able to interpret the settlement of the region for the period that we now call the Late Archaic. The primary limitation was a lack of understanding of projectile point types. As late as the 1960s, when Wauchope was attempting to describe the projectile points recovered during his survey, there were few reports available with data from stratigraphic excavations relating to the Archaic period. The detailed typological information found in the reports by Coe (1964) and Cambron and Hulse (1964) only became available as his book was reaching completion (Wauchope 1966:92).
Wauchope did recognize the type Savannah River Stemmed (Coe 1964) as the hallmark of the Late Archaic period and followed Phelps (1964:81) in placing the large, broad-bladed point within the Stallings Island sub-phase of the South Coast Savannah River Phase (Wauchope 1966:6). Because most of the projectile points had been recovered from surface collections, Wauchope devised a series of categories or groups based on attributes of shape and size. Wauchope reviewed the existing archaeological literature in an effort to associate the individual groups with time periods and previously named point types. Based on information from "pure" sites in his survey, 27 of the 32 projectile point groups seemed to have appeared in the Archaic period, or earlier but most continued to be used in later times (Wauchope 1966:96). Possible information relating to the Late Archaic period was also obscured because all early pottery types were viewed as being associated with the Early Woodland period. His "lower Early Woodland" period was introduced around 2,000 B.C. with fiber-tempered pottery that were gradually replaced by fabric-marked ceramics. His "upper Early Woodland" (Deptford) began about 1,000 B.C. (Wauchope 1966:8).
Wauchope's report provides a good discussion of soapstone vessels. In commenting on Coe's (1952:305) ideas that these vessels were a prototype for pottery, Wauchope noted that most of the vessel fragments found in his survey were recovered from Woodland period sites. Wauchope concluded that soapstone (steatite) vessels were probably used throughout Georgia's prehistory (Wauchope 1966:194). Wauchope's survey data from Northwest Georgia shows that less than 20 soapstone sherds were found on six sites in Bartow County and a roughly equal number were found on six sites in Cherokee County. Most of these sites produced stemmed points that probably date to the Late Archaic period. Only two of the sites produced high counts of fabric-marked pottery and eight sites produced small amounts of fabricmarked pottery. There was no reference to soapstone vessel fragments being found in feature contexts.
31

As a result of his extensive WPA-era work in northern Georgia, Robert Wauchope was able to introduce a ceramic sequence that is still used today (Wauchope 1948, 1950, 1966). By contrast, there are numerous shortcomings related to his projectile point typology and that data is quite difficult to interpret. However, Wauchope's (1966:228) observation that medium-sized triangular projectile points are consistently found on sites with the earliest fabric-marked pottery in Northwest Georgia still seems valid. This observation concerning triangular points was further supported by Joseph Caldwell's subsequent research in the region (Caldwell 1957, 1958). The association of these triangular points with fabric-marked pottery is now viewed as characteristic of the Early Woodland Kellog phase in Northwest Georgia (Wood and Bowen 1995:8). As far as identifying projectile points that might bridge the gap between the Archaic and Early Woodland period; Wauchope's survey data is of limited value. While he discussed the concept of chronology based on projectile point categories, his data was limited by the fact that most of the material was taken from surface collections and most of the larger samples were taken from large sites in the Piedmont province that displayed evidence of multiple occupations (Wauchope 1966:95).
Large Scale Reservoir Salvage Projects of the 1940s through early 1970s.
A series of major archaeological salvage projects associated with reservoir construction were conducted in Northwest Georgia and northeastern Alabama beginning in the late 1940s. The work began with the Smithsonian Institution's River Basin projects at Allatoona and Buford Reservoirs and continued with work sponsored by the U.S. National Park Service at Carters Lake and the investigations paid for by the Alabama Power Company at Weiss Basin (Figure 7). The majority of the large scale excavations at these sites focused on components dating to the Woodland and Mississippian periods. A few sites were investigated that contained primary occupations dating to earlier and later periods.
Figure 7. Map showing the locations of Northwest Georgia reservoirs where large scale excavations were conducted in the late 1940s through the early 1970s.
32

The Smithsonian's River Basin Project at Allatoona Reservoir
A few years after Wauchope completed his field investigations, additional survey and excavation was conducted in Bartow, Cherokee, and Cobb counties, in an area that was to become the Allatoona Reservoir (Caldwell 1947a, 1950:5, 1957; Miller n.d.; Roberts 1951). That work greatly increased our understanding of the prehistory of Northwest Georgia and the Woodland period in particular (Caldwell 1958).
In late 1946, Joseph Caldwell, working with the Smithsonian Institution's River Basin Survey Program (Roberts 1951:351; Lyon 1996:203), began a six-month survey of the proposed Allatoona Lake. Adding to the 26 sites previously found by Wauchope, Caldwell's survey of cultivated lands within and around the reservoir basin identified another 180 sites (Figure 8). Many of the sites produced artifacts associated with Late Archaic and Early Woodland components. The Smithsonian, with the assistance of the University of Georgia, finally undertook a program of salvage archaeology in 1949. Eight sites were extensively excavated and another 13 were tested. The 1949 excavations were initially supervised by Carl Miller, who was later joined by Caldwell. William Sears participated in excavations at 9CK5 (Sears 1958). A preliminary report of Caldwell's work appeared in a 1950 volume of Early Georgia (Caldwell 1950:5-21). Many of Caldwell's more important concepts developed from this work were published in Trend and Tradition in the Prehistory of the Eastern United States (Caldwell 1958). The results of Miller's work exist only in manuscript form (Miller n.d.).
A preliminary summary of the archaeological investigations in Allatoona Reservoir was prepared for the 1951 Annual Report of the Smithsonian Institution (Roberts 1951). The report appears to contain information contributed by both Caldwell and Miller, and seems to include ideas expressed by both individuals. According to Robert's report, the "Kellogg period" represented a semi-sedentary hunting and gathering culture that made considerable use of storage pits. An earlier period was represented only by "a certain type of projectile point" and scattered finds of fiber-tempered pottery. Roberts called that period Stallings because the points and potsherds were similar to those occurring on Stallings Island in the Savannah River valley (Roberts 1951:369). The oldest of the sequences identified at Allatoona Reservoir was designated "Prepottery." Roberts noted that the pre-pottery stage preceded Stallings, possibly represented several periods, and covered a long duration of time (Roberts 1951:370).
Caldwell submitted a manuscript of his survey and excavations to the Smithsonian Institution in 1955 but his Allatoona report was not published due to funding issues. After a period of about ten years, the manuscript was returned to the University of Georgia for publication as a Laboratory of Archaeology Series Report. The report was edited but never completed and remains today as an unfinished manuscript. Today, Caldwell's Allatoona manuscript consists of his original report to the Smithsonian with some minor introductory additions and numerous hand written editorial comments. The UGA Laboratory of Archaeology files also contain earlier drafts of the report that include some information omitted from the final report. Approximately two decades ago, archaeologists began citing the report as "Caldwell 1957" (Hally and Rudolph 1986). While the justification for adopting that date is unclear, it will be retained for this paper because of general acceptance. The Allatoona report has frequently been cited as "Caldwell n.d."
33

Figure 8. Map showing sites recorded in Allatoona Reservoir in the 1940s (from Caldwell 1957: Figure 1).
Carl Miller's report was to be included in Caldwell's report as an appendix. Miller's completed manuscript is not dated but cited references suggest a finishing date in the mid-1960s. At the present time, the two reports are filed as separate manuscripts at the University of Georgia's Laboratory of Archaeology in Athens. Caldwell's report includes a full text and graphics but the existing copy of Miller's report lacks figures. The Smithsonian Institution's archives apparently no longer have copies of the two manuscripts but Miller's field notebooks and site maps are there and are available for study (Lorain Wang, Reference Archivist, National Anthropological Archives, personal communication 2007).
Most of the findings from Carl Miller's work at Allatoona have never been widely disseminated and his interpretations have been relegated to obscurity. For that reason, the information found in Miller's manuscript is seldom cited in the archaeological literature of the region. On the other hand, Caldwell's interpretations were published in his 1958 Trend and Tradition volume and these concepts were widely accepted by the archaeological community. Figure 9 shows one of Caldwell's illustrations relating to the manner in which Early Woodland ceramics (his Middle Eastern Tradition) evolved in the region. The belief at that time was that soapstone vessels preceded earthenware and the earliest earthenware originated in the Northeast (Vinette). An earlier period of development of fiber-tempered pottery was not yet recognized.
34

Figure 9. Illustration from Caldwell's Trend and Tradition (1958: Figure 2) showing the "idea of earthenware" (reproduced courtesy of the American Anthropological Association).
In reading through the two manuscripts, it is clearly evident that the two authors had very different ideas concerning the interpretation of their findings and the overall interpretation of culture history for Northwest Georgia. A case in point, that relates specifically to the present study, concerns the period when fiber-tempered pottery was introduced to the region. The scarcity of fiber-tempered pottery in the Allatoona area suggested to Caldwell that either the region was largely unoccupied at that time or the inhabitants of the region never adopted pottery (Caldwell 1957:283). Miller concluded that they had probably not looked at the right sites or dug deeply enough (Miller n.d.:281).
Note to the reader. It is important to keep in mind that many of the site designations used in the original Allatoona manuscripts have been replaced by new site numbers. This has resulted from the fact that county site numbers were being assigned independently by both the University of Georgia and Georgia State University. When the site files from the two institutions were merged in the 1980s, the Georgia State numbers took precedence in Bartow, Cherokee and Cobb counties. Because of the confusion that exists with site numbers, the more important Allatoona sites are usually designated by name in the archaeological literature. When considered appropriate for this discussion, both the old (Caldwell's original numbers as used in the Smithsonian report) and the new site numbers will be noted. Also, Caldwell's original spelling of "Kellog" will be used in this volume except in cases where the alternative spelling (Kellogg) is used as a direct quotation.
35

Joseph Caldwell's Allatoona Survey and Excavations

Caldwell's Allatoona report provides brief site descriptions and artifact counts for most of the sites recorded during his survey. A substantial portion of the report is devoted to detailed discussion of a few major excavations. In general, the information presented for the pottery-producing sites may be readily used today and much of that data has been incorporated into more recent settlement studies (Bowen 1989:110-133; Ledbetter et al. 1987:170-285). With respect to the subject of the present study, Caldwell's report provides detailed excavation data for three relevant sites. The Stamp Creek site (9BR60, now 9BR139) produced evidence of a pre-pottery component (Stamp Creek focus) as well as Woodland and Mississippian occupations (Figure 10). The nearby Boston Creek site (9BR73, now 9BR141) produced evidence of intensive Early and Middle Woodland occupation. The Kellog site, formerly 9CK62 now 9CK102, was located on the Etowah River (see Figure 8). Artifact assemblages from that site are still used today to define local Early and Middle Woodland phases. A portion of the Kellog site also contained an area described by Caldwell as a pre-pottery lithic workshop that he associated with the Stamp Creek focus.

A fourth site, located on Clark Creek and designated 9CK57 (now 9CK112), represented a minor excavation, but it produced Archaic period artifacts that are pertinent to our study (see Figure 8). Caldwell's pre-pottery Stamp Creek focus was intended to represent the final stage of the Archaic period (Caldwell 1957:280). The primary diagnostics of this focus were large Savannah River or Appalachian Stemmed points and soapstone vessels but a variety of other projectile point types were also included (Caldwell 1957:279). Caldwell defined an Early Woodland Kellog focus, which is still utilized today under the name Kellog phase. Caldwell's Kellog focus was defined by the almost exclusive use of Dunlap Fabric Marked pottery on sites that also produced medium-sized triangular points. This association had been recognized earlier at the Two Run Creek site (Wauchope 1939:12).
36

Figure 10. Map from Caldwell's Allatoona report showing sites on upper Stamp Creek (adapted from Caldwell 1957: Figure 1).

It is important to stress the fact that in the mid-1940s Caldwell was working in a portion of Georgia for which material culture was poorly documented. This was particularly true during the first couple of years of the Allatoona project. The collections, notes, and records from Wauchope's earlier work were utilized (Caldwell 1947b:224); however, Wauchope's (1948:201) published pottery sequence for the region would have been critical to Caldwell's work. Caldwell's early use of names such as "cordwrapped stick decoration" (Caldwell 1950:17) is one such example. The following note from Caldwell's Allatoona files (UGA), relates to a pottery type now known as Dunlap Fabric Marked.
"After consultation with Mr. Fairbanks of Frederica, we have substituted the term "cordwrapped stick" to refer to the decoration which we formerly called "dowel-marked." This is the same decoration as occurs on the so-called dowel-marked pottery of eastern Tennessee and the same as the so-called "fabric-marked pottery of central Georgia" [Caldwell 1947c:3].
There was even less of an understanding of projectile point types at the time of Caldwell's survey. In an effort to remedy this problem, Caldwell created his own point classification scheme. His point "typology" was in place by the early part of 1947 (Caldwell 1947c:2 ). As such, it represents one of the first efforts to classify projectile points in Georgia. The surviving records of his classification consists of outlines with notations relating to various attributes. Site locations, represented by temporary field numbers (in Roman numerals) are attached to the outlines. A copy of his classification scheme was submitted to the Smithsonian Institution as a part of a progress report but was not included in his final report. Figure 11 illustrates examples from this "Projectile Point Classification" (Caldwell 1947c).
Figure 11. Examples of projectile point types defined for Allatoona Reservoir (Caldwell 1947c).
37

Caldwell's Late Archaic Stamp Creek Focus. Caldwell defined the period of prehistory prior to the introduction of the earliest sand-tempered pottery as the Stamp Creek focus (Caldwell n.d.; 1957:280). The text of Caldwell's report implies that the Stamp Creek focus was meant to correspond to what we now call the Late Archaic period, but an examination of the illustrated artifacts suggests that he probably included some earlier and later material. His projectile points include typical Savannah River/Appalachian Stemmed projectile points, a variety of small stemmed and notched points, and soapstone (steatite) sherds (Figure 12). The artifacts associated with Caldwell's Stamp Creek focus were considered similar to material found at other Eastern U.S. sites of the "Archaic pattern" as defined at that time, but Caldwell felt the Stamp Creek assemblage differed in some respects from the related Savannah River focus of eastern Georgia (Fairbanks 1942:223-231) and the Lauderdale focus of northern Alabama (Webb and DeJarnette 1942:19).
With respect to the traits used to define Stamp Creek, Caldwell noted that of the various stemmed points found on Stamp Creek focus sites, the medium to large, simple-tang (stemmed) points, "probably javelin tips," were the most characteristic and also showed the closest resemblance to materials from other Southeastern pre-ceramic foci (Caldwell 1957:279, 1958:13). Such points are usually relatively large and heavy, the stem is square, and the shoulders broad and well defined (Caldwell 1957: Figures 8-9). Caldwell also included hemispherical soapstone bowls and other groundstone artifacts such as the cylindrical object shown in Figure 12. Caldwell recognized that perforated tablets made from soapstone, "the so-call net sinkers," that were so numerous at Stallings Island and other Savannah River focus sites, were practically absent in the Allatoona area. The excavated Stamp Creek focus sites produced no axes, atlatl weights, bone or shell artifacts (Caldwell 1957:280).
Caldwell noted that at Allatoona, quartzite was usually employed for large, simple-tang points, but quartz was little used. Flint (chert) was used to produce smaller points that were highly variable in shape and that included slight (expanded) tang, simple (straight) tang, bifurcated tang, corner notched, side notched and stemless (isosceles) bases. At the Stamp Creek site and other Allatoona Reservoir sites, most of these "flints" belonged to two general categories that Caldwell called "slight tang points" and "simple tang" points. Caldwell noted that the slight tang points were produced over an extended time period, with some examples found on pottery-producing sites (Caldwell 1957:9).
Caldwell singled out only a few sites as having Stamp Creek focus components. These were two areas of the Stamp Creek site (9BR60-A and 9BR60-C ) that represent the type site, a second site (9BR85) that was also located on Stamp Creek, the pre-pottery workshop area on the Kellog site (9CK62), and site 9CO52 (Caldwell 1957:278). Our examination of the artifact descriptions in Caldwell's report indicate that site 9BR60-A and 9BR60-C were the only two sites within this group to produce significant numbers of large Savannah River Stemmed points made from quartzite. According to artifact data in his report, 9CK62, 9BR85, and 9CO52 produced soapstone sherds but no chipped stone quartzite artifacts (Caldwell 1957:55, 163, and 241). Caldwell's 1947 appraisal report also illustrates two quartzite Savannah River points and a soapstone sherd from site 9CO75 in Cobb County (Caldwell 1947:17). A portion of that site, which is now designated 9CO4, was excavated in 1990 but there was little evidence of intensive Late Archaic occupation (Cable and Raymer 1991:90).
38

Figure 12. Illustration of Stamp Creek focus artifacts from a portion of the Stamp Creek site, 9BR60-A (adapted from Caldwell 1957: Figure 8).
39

The excavations on 9BR60-A produced a number of features dating to several time periods (Figure 13). The palisade lines and most of the post molds are associated with later components. Caldwell concluded that 18 pits on the Stamp Creek site (9BR60-A) could be attributed to the pre-pottery occupation. Most appeared to be used for storage of food, but one contained red ochre and some traces of human bone. The possibility existed that some of the other pits were burials from which the bone had disappeared. The possible burials contained relatively purer redeposited clay with fewer broken rocks and other inclusions.
Figure 13. Caldwell's excavation of the Stamp Creek Site (9BR60-A) showing all features (adapted from Caldwell 1957: Figure 6).
Caldwell described his pre-pottery pits as similar in appearance, usually with straight sides and flat bottoms. Dimensions were given as 2.5 to 5 ft in diameter and 1.5 to 3 ft deep. A few were oval or oblong and in two or three instances sides were sloping. Some features found by Caldwell in site area 9BR60-A displayed evidence of "hot rock" cooking, but features that may have represented hearths were not found (some pits contained fire-cracked rock but not in hearth arrangements). Caldwell noted that later cultures did use cooking pits or "fireless cookers," with an upper layer of rocks over a fire which had been built in the bottom, but none of these were found in the Stamp Creek component. Caldwell suggested that most of the pits associated with the Stamp Creek focus on 9BR60-A were for food storage, in much the same manner as the pits of the later Kellog period (Caldwell 1957:280).
40

Caldwell's report provides artifact counts for features thought to be pre-ceramic. The Stamp Creek focus pits were identified by certain types of projectile points and soapstone sherds. Five of the pits have single examples of large quartzite Savannah River type projectile points, one contains a large quartzite blade, and four contain soapstone sherds. Nine of the pits produced smaller stemmed points made from chert or quartz. Four features contained intrusive sherds and three contained triangular projectile points (Caldwell 1957:24). Based on our current understanding of diagnostic artifact types, the recovery of artifacts that post-date Savannah River Stemmed projectile points suggests that some of the features identified by Caldwell are probably associated with later occupations (Terminal Archaic or Woodland).

Caldwell's site 9CK57 on Clark Creek produced artifacts that are probably pertinent to our study. Collections were made from the surface and an unspecified number of test pits. Caldwell interpreted the site as a possible early pottery period lithic workshop with a pre-pottery occupation (Caldwell 1957:256). The list of 81 artifacts includes a single plain sherd, three triangular points, a variety of small to medium sized stemmed and contracting stemmed points, chipped stone debris, and seven soapstone vessel fragments. Plate 50 in Caldwell's report illustrates several bifaces that probably date to the Middle and Late Archaic periods. His evidence for a pre-pottery occupation was largely based on the recovery of several soapstone sherds. The absence of large stemmed points (Savannah River Stemmed) led Caldwell to conclude that the Archaic occupation was something other than Stamp Creek focus, which meant that the assemblage belonged "to a yet undefined prepottery period" (Caldwell 1957:257). The site was revisited during a later survey but the only artifacts found were chipped stone debris and fragmentary bifaces that appeared to be Archaic (Ledbetter et al. 1987:94).

Caldwell regarded the Stamp Creek focus as a relatively late pre-ceramic culture. By repeatedly

stating that fiber-tempered pottery was

practically absent in northern Georgia,

Caldwell appears to imply that the

ceramic type was not associated with the

Stamp Creek. His report (Caldwell 1957:

Figure 10) actually illustrates a fiber-tem-

pered sherd under a brief section on early

pottery from 9BR60-A (Figure 14), but the

report text contains no specific reference

to the sherd. An examination of an earlier

draft of the report produced a single para-

graph that was omitted from the final re-

port that states: "In the Allatoona Reser-

voir, three sites, including 9BR60-A, each

yielded one sherd of this type apiece. We

know of two ways to account for the ab-

sence of fiber tempered pottery in this

area. Either the region was deserted at the time it was being made elsewhere, or ce-

Figure 14. Drawing of a fiber-tempered sherd found on site 9BR60-A in Allatoona Reservoir (adapted from Caldwell 1957: Figure 10).

41

ramics simply did not come into use by the people who were living here until a much later date" (Allatoona Reservoir River Basin Survey notes and manuscript file, UGA Laboratory of Archaeology, Athens). His report does describe the one fiber-tempered sherd recovered from site 9CK101 during the original survey. The sherd was identified as the type "Stallings Island Incised and Punctate" (Caldwell 1957:207). That fiber-tempered sherd was considered to be an isolate on an otherwise Woodland period site. Caldwell's report refers to Carl Miller's appendix for the results of additional investigations on 9CK101, but that appendix contains no reference to the sherd or the recovery of additional fiber-tempered pottery. The illustrated sherd from site 9BR60-A would appear to be quite similar to the example reportedly found on site 9CK101. Caldwell's report and field notes provide no further information on the third pottery site.
Caldwell's report contains a few additional references to the early fiber-tempered earthenware that is represented on the shell middens of the Savannah River and coastal Georgia by the types Stallings Incised and Punctated and Stallings Plain (after Griffin 1943). He noted that, "Such pottery was practically absent in Allatoona, and indeed has a limited distribution in Georgia, few sites of any size having been found outside areas where the shell heaps occur. This observation might in fact hold for the entire Southeastern area, that no consequential site with fiber tempered pottery has been found at any distance from a shell heap" (Caldwell 1957:283). Caldwell had earlier emphasized the scarcity of the pottery type in the region by noting that Stallings Island, the period of earliest fiber-tempered pottery, was hardly represented at all in the Allatoona Reservoir (Caldwell 1950:6).
The extreme rarity of fiber-tempered pottery in the Allatoona Reservoir area required an explanation. Caldwell suggested that the absence of sites during the interval between the pre-ceramic Stamp Creek focus and the time when the first fabric-marked pottery was introduced could be explained either as abandonment or the failure of people to adopt earthenware vessels.
"With fiber tempered pottery practically absent in interior northern Georgia, the earliest ware appearing
in quantity at Allatoona was Dunlap Fabric Marked. This sand or grit tempered type, decorated by impressions of a plain plaited fabric, was found to be associated with trait assemblages classifiable as "Early Woodland," and apparently culturally discontinuous with the preceding non-ceramic sites of the Stamp Creek Focus. Possibly the Etowah Valley had been deserted while fiber tempered pottery was in vogue on the shell heaps, but having noted the apparent connection between fiber tempered pottery and the shell middens, it is equally possible that the Etowah Valley was inhabited, and the people simply failed to adopt earthenware at that time" [Caldwell 1957:283].
In Trend and Tradition, Caldwell further examined the discontinuity between the Late Archaic and the subsequent Kellog period. He emphasized the concept by noting "No culturally intermediate site has been located, and there has been sufficient survey of the area to suggest that should any such sites be found, they will be rare. In view of this and the continuous distribution of similar sites northward, we infer that the Kellog Focus moved southward into Piedmont Georgia and Carolina, possibly from eastern Tennessee..." (Caldwell 1958:23). Some researchers have argued that Caldwell's idea of a "temporal and cultural hiatus" cannot be supported based on current archaeological data (Bowen 1989:215, Crook 1984:55, Rotenstein n.d.:7). However, other excavation data suggest that Caldwell's concept of discontinuity of artifact types is valid (Wood and and Ledbetter 1990:19).
42

Caldwell's Stamp Creek focus is seldom mentioned in the ongoing archaeological research of Northwest Georgia (Ledbetter 2007). The term has appeared infrequently in archaeological reports of the past few decades (Cable and Raymer 1991, Blanton et al. 1987). In their discussion of the Late Archaic period, Blanton and his colleagues (1987:19) simply note that Caldwell's Stamp Creek focus does exhibit many of the traits associated with the Late Archaic.

Caldwell's Early Woodland Kellog Focus. Caldwell's investigation of two Allatoona Reservoir sites provided much of the information used for defining the series of traits for his Early Woodland Kellog focus. One site, which he designated 9BR73, is presently known as 9BR141or the Boston Creek site. The Boston Creek site was identified by Caldwell as a large (ca. 5 ha) Woodland period habitation site (Caldwell 1957:56). The site is separated from the Stamp Creek site (9BR60) by the narrow channel of Stamp Creek (see Figure 10). Caldwell's limited excavations on 9BR73 exposed a portion of an Early Woodland feature cluster composed of pits, post molds, and rock hearths in association with Dunlap Fabric Marked pottery and medium-sized triangular points. Separate stratigraphic excavations provided pottery counts used to define much of the local Early and Middle Woodland ceramic sequence for the Allatoona area. The second site, which he called the Kellog site, covered approximately one hectare and was located on a narrow ridge parallel to the Etowah River (see Figure 8). Caldwell's excavations encountered cultural deposits that were seldom more than 0.3 m deep; however, underlying pit features were abundant and most contained Dunlap and Cartersville pottery along with triangular points. The information recovered from the large artifact collection from these features was used to define many of the traits associated with the Kellog focus.

Caldwell's excavations on the Boston Creek site (9BR73) confirmed the intensive Early Wood-

land occupation with midden deposits and preserved features (Figure 15) and allowed the means to

interpret the stratigraphic sequence for most of the Woodland period (Figure 16). Caldwell's strati-

graphic pit, which measured 20 ft (6 m) long by 15 ft (4.6 m) wide, was placed near the confluence of

Stamp Creek and Boston Creek where the cultural material extended to 6 ft below the surface. The unit

was excavated in 3-inch levels. The upper depositional layers were identified as light brown silt of

recent origin that covered a thin layer of darker silt "wash." These soil layers covered the main occupa-

tion deposit which was a

black sandy soil that ex-

tended from 2 to 6 ft below

surface. The great depth of

the cultural zone was attrib-

uted to midden filling an

old depression that may

have been an old tributary

stream of Stamp Creek.

Caldwell (1957:57) noted

that a few additional arti-

facts continued to be found

in the underlying zones

within the depression.

Figure 15. Map showing Caldwell's work at Boston Creek (redrawn from Caldwell 1957: Figure 39).

43

Figure 16. Copy of Figure 35 from Caldwell's Allatoona Reservoir report which shows the results of stratigraphic testing on the Boston Creek site (adapted from Caldwell 1957: Figure 35).
Caldwell defined the lower levels that contained only Dunlap Fabric Marked and plain pottery as belonging to the Kellog period and the middle levels that contained a mixture of Dunlap and Cartersville Check Stamped as belonging to a "Post-Kellog Period" (Caldwell 1957:58). Caldwell's sequence generally confirmed Wauchope's earlier work at Two Run Creek, except that an early form of simplestamped ware (Mossy Oak) was lacking. Within the stratigraphic unit, "Cartersville Simple Stamped" occurred only in the uppermost layers. Caldwell later revised his interpretation of this sequence to include the Forsyth phase (a pure check-stamped assemblage) that occurred stratigraphically below cultural deposits with a mixture of check-stamped and simple-stamped pottery of the Cartersville period (Caldwell et al. 1952:326).
The Kellog site (9CK62), located on a narrow ridge parallel to the Etowah River, was Caldwell's most significant Early and Middle Woodland period excavation in the reservoir. The features on the site included storage pits, a smaller number of cooking-related pits, one cremation pit, and post molds (Figure 17). Figure 18 shows typical pits, most of which were cylindrical and about 3 ft deep (Caldwell 1958:25). One circular pattern of post molds defined a small structure 3.3 m in diameter (Caldwell 1957:142-175). That structure was probably associated with a Middle Woodland Cartersville phase occupation.
44

Figure 17. Map showing features in Caldwell's excavation of the Kellog site (Caldwell 1957: Figure 34).

Caldwell noted that the fill of most of the pits contained Dunlap Fabric Marked pottery (Figure 19), along with small amounts of Cartersville Checked Stamped. A few post-Kellog pits were identified and while the site contained evidence of "a heavy Cartersville settlement, it is of interest that we were unable to identify a single storage pit as definitely belonging to the Cartersville horizon" (Caldwell 1957:153).

Figure 18. Photograph of excavated Early Woodland pits on the Kellog site. A.R. Kelly and
an unidentified man are shown in the background near the rising waters of Allatoona
Reservoir (courtesy UGA photo files).

Chipped and groundstone artifacts were also used to define the Kellog period. The presence of medium-sized, triangular points at the Kellog site was viewed as representing a radical change from Late Archaic stemmed points. The most common point type at the Kellog site was a stemless isosceles type usually made from quartz (Figure 20). Some of the chert points displayed "ears" at the base of the haft area. Small to medium-size, stemmed points, made from chert and quartz ovates, were represented by four occurrences (Caldwell 1957:175).

45

Figure 19. Dunlap Fabric Marked vessels from the Kellog site (adapted from Caldwell 1957: Figures 39-40).
The recovery of these triangular points along with bar gorgets, which Caldwell believed to be arm guards, was viewed as evidence that the inhabitants of the Kellog site possessed the bow and arrow (Caldwell 1958:26). Groundstone artifacts associated with the Kellog period included soapstone pipes, boatstones, and two-hole bar gorgets. A number of soapstone vessel fragments were found in the excavations and in several instances in the fill of Kellog period storage pits. Only one fragment was recovered from the pre-pottery workshop. Noting that vessels of this period are widely reported from eastern pre-pottery sites and that they occurred locally on Stamp Creek focus, Kellog, and Cartersville sites, he suggested that soapstone vessels continued to be used for some time after the introduction of earthenware (Caldwell 1957:171).
Caldwell recognized a strong reliance on seasonally available forest resources. The storage pits at Kellog were filled with charred acorns, hickory nuts, walnut shell and charred seeds. Caldwell used the data from the Allatoona sites to help refine his idea of primary forest efficiency. The exploitation of varied food sources, along with the ability to predict the time and place of their availability, allowed the people to use certain technological innovations to efficiently harvest the resources of the eastern deciduous forests. These technological innovations, such as durable containers for processing and storing foods and more efficient hunting techniques, were extensively utilized for the first time. Caldwell postulated that people settled at locations such as the Kellog site because a variety of resources would be available throughout a significant portion of the year (Caldwell 1958:23-26).
46

Figure 20. Examples of Kellog and Post-Kellog points from the Kellog site (adapted with modifications from Caldwell 1957: Plate 41).
47

Caldwell recognized some similarities in material culture between the Kellog people and the earlier inhabitants of the area but he consistently stressed the differences. That attitude is displayed in the following sentences.

"In summary, the Kellog Focus of the Kellog Period in the Etowah Valley may be viewed as representing the settlements of a pre-agricultural, hunting-gathering people who placed great reliance on nuts gathered in the autumn and stored in subterranean pits. The Kellog site in particular is riddled with storage pits, which were often dug into one another. There is some slight evidence that these people were in possession of the bow and arrow. Most of their minor artifacts differed from those used by the preceding Stamp Creek Focus, and a few similarities with earlier periods: the occasional manufacture of steatite vessels, and the continuing use of shallow biconcave grinding slabs and biscuit-shaped hand stones, are widespread and can hardly imply any direct connection" [Caldwell 1957:292].

The excavated data from the Kellog site was re-examined by Rowe Bowen (1989). A detailed examination of features from the Kellog site provided Bowen with a means of quantifying feature types and examining feature patterning. Bowen noted that certain feature types did tend to cluster and that semi-circular alignments of postholes often appeared to be associated with certain features. These post alignments may be the remains of structures such as racks or windbreaks (Bowen 1989:46). Bowen processed Caldwell's soil and charcoal samples for analysis and radiocarbon dating (Table 2). With respect to Early Woodland pottery, Bowen found that in addition to Dunlap Fabric Marked, Watts Bar Fabric Marked sherds constituted an important minority type (n = 105). Watts Bar sherds were found in about half of the excavated pits (28 of 60), but there was no limestone-tempered pottery on the site (Bowen 1989:167).

Table 2. Radiocarbon Dates from the Kellog Site (data from Bowen 1989:146, Tables 1 and 11).

Provenience (Feature No.)

Lab Number

Conventional Radiocarbon

Uncorrected Date

Diagnostic Artifacts

Feature 47 UGA-5545 2405 80 B.P.

455 B.C.

Feature 39/45 UGA-5544 2320 100 B.P.

370 B.C.

Feature 25 UGA-5543 2245 200 B.P.

295 B.C.

Feature 55 UGA-5546 2150 180 B.P.

200 B.C.

Dunlap Fabric Marked, *Cartersville Check Stamped (N = 1),
plain pottery, mano, grinding stone, lithics
Dunlap Fabric Marked, simple stamped (N = 2), plain pottery,
fired clay, lithics
Dunlap Fabric Marked, *Cartersville Check Stamped (N = 1),
plain pottery, lithics
Dunlap Fabric Marked, lithics

* interpreted as intrusive or incidental material in a Kellog phase feature (Bowen 1989:44).

48

Four of five samples submitted by Bowen for dating were viewed as suitable for the Kellog phase. A fifth sample from Feature 34 produced a date of A.D. 507 and was rejected as being too recent (Bowen 1989:146). The samples were taken from silo-shaped pits found from two of the pit clusters. Feature 25 was in Pit Cluster 3, while Features 39/45, 47, and 55 were from Pit Cluster 4 (Bowen 1989: Table 1). Bowen noted that the samples taken from Pit Cluster 3 were a part of the site that contained the heaviest Cartersville influence and likely represented the latter part of the Kellog occupation, while the earliest occupation area (Pit Cluster 1), located at the southern end of the site, produced no datable samples (Bowen 1989:146). This comment implies that the radiocarbon dates from Kellog do not define the full extent of the Kellog phase occupation on the site.
Carl Miller's Excavations in Allatoona Reservoir
Miller's excavations in Allatoona Reservoir were conducted from July through December, 1949. Miller's major excavations were conducted on sites 9BR52, 9BR60, 9CK72, 9CK101, 9CK103, 9CK23, and 9CO82. Limited testing took place at sites 9BR78, 9BR79, 9BR80, and 9BR81, "as well as a number of unlisted sites lying between 9CK23 and 9CK7" (Miller n.d.:17). Miller began the excavations of the Stamp Creek site (9BR60) prior to being sent to Louisiana for another project, and the site was turned over to Caldwell to finish (Miller n.d.:23). While several of the sites investigated by Miller produced significant amounts of Archaic or Early Woodland material, the Stamp Creek site (9BR60-A), Fields Bridge site (9CK72), and the Guess site (9CO82) appear to have contained the best preserved Archaic deposits. Evidence of less intensive occupation during the Late Archaic or Early Woodland periods was found during excavation of the Knox Creek site (9CK101) and 9CO82.
Miller's voluminous report of his work in the Allatoona Reservoir consists largely of site interpretations of the major excavations. The presentation of specific details such as artifact counts varies from site to site. Substantial portions of his report are concerned with trait lists and discussions of pottery types. Miller's sites are sometimes referenced in the main text of Caldwell's report, but most frequently Caldwell refers the reader to his appendix for information on Miller's sites (Miller's report was included as an appendix to Caldwell's report). Unfortunately, Miller's report includes little or no information on the sites that he tested throughout the reservoir.
In the opening pages of his manuscript, Miller comments on the shortcomings of his profession's understanding of Georgia archaeology. His comment on perceptions relating to the Archaic period seems particularly relevant to the present study.
"As the tentative chronological sequence of Georgia's prehistory now stands the earliest well documented cultural horizon is the Archaic Period which has been subdivided into an Early and Late Phase. These are characterized by a group of long headed people who were ignorant of agriculture and pottery-making; they used the atlatl and javelin in place of the bow and arrow, possessed no known type of architecture, and were confined to a hunting-food gathering economy. These people were supposed to have occupied Georgia until about 1,000 B.C. when the art of pottery-making, along with the bow and arrow, was first introduced into the Southeast. This pottery was a molded, vegetal-tempered ware embellished with very simple punctate-incised designs on bowl shapes [Miller n.d.:3].
49

Miller's manuscript provides some interesting interpretations of prehistoric settlement in the Allatoona Reservoir. Miller viewed the earliest communities of northwest Georgia as temporary stations located along the banks of the main streams and their tributaries. He noted that these sites contain chipped stone, whole or fragmentary stone artifacts, and fragments of soaptone or sandstone vessels, "all of which was nonperishable and served as indicators of a culture prior to the introduction of pottery and the refinements that accompanied it" (Miller n.d.:242). Miller suggested that very crude temporary shelters were constructed of small branches and brush around hearths or cooking areas and they left little or no imprint on the site except for "a hardpacked floor area and its accompanying hearth" (Miller n.d.:242).
Miller concluded that some of the pit features exposed during his excavations were attributable to the Archaic occupations.
"Numerous pits were dug into the sites which functioned as temporary storage areas. Later these same pits were used as depositories for accumulated midden material. In other pits fires had been built, as shown by the mass of ashes and fire cracked and broken stones, attesting to the use of hot-rock cooking in either fiber or hide bags. In still other pits were found the concentrated remains of fires, fire cracked or broken stones and burned animal bones which may indicate that these people practiced barbecuing or used them as roasting pits. In some cases mixed with the broken rocks, ash, and animal bones were an occasional projectile point, either whole or fragmentary. Whether this projectile was used to kill the animal that was cooked in this particular pit or it was discarded after the meal was over was immaterial" [Miller n.d.:243].
Miller elaborated on these observations with respect to one specific site, 9CK72. That site apparently contained good evidence of pre-pottery period occupation. Miller noted that,
"Here a number of scattered pits were found and investigated. Evidence points toward an early lithic cultural group of the Late Archaic who settled on this spot quite some time. Their cultural remains in the form of projectile points, a crude notched ax as well as steatite vessel sherds were recovered from a number of pits as well as the lowest level within the midden deposit. Associated with this artifact assemblage were large hearth areas composed of masses of cracked and broken stones intermixed with ash and burned earth which indicated that at one time a shelter of some sort had been built over and around the area. These shelters were of such a flimsy nature that they left no trace within the site outside of the hearth areas" [Miller n.d.:260].
Relying upon the prior works of "Claflin, Miller, Fairbanks, and others," he concluded that the pre-ceramic levels on sites within the Allatoona Reservoir area represented the first real indication of the presence of man in Northwest Georgia. He felt that artifacts from these sites conformed in shape and pattern with those found and described from comparable sites throughout the Southeast and thus assigned the sites to the Stallings Island Aspect (Miller n.d.:281).
In reviewing Miller's report, it is apparent that he viewed the scarcity of fiber-tempered pottery in Allatoona Reservoir as an important issue that was not being adequately addressed during the project. Based on several references throughout his report, Miller believed that the early pottery was probably present but in low numbers. He noted that,
50

"Following the occupancy of these sites of the pre-pottery period, there was a slight cultural gap that was not filled with the present investigations in the Allatoona Reservoir basin area. This gap occurred at the start of the pottery-making art [and] was introduced into the Southeast in the form of fibertempered ware. Proof positive that such a ware existed within the immediate vicinity of the reservoir area was determined by fiber-tempered sherds occurring on a site on Slough Creek a few miles west of Atco, Georgia, on the limits of the Allatoona Reservoir proper. Here the lithic assemblage conformed with the earlier cultural pattern in which were intermixed with the sherds of a slightly later group" [Miller n.d.:243].
He repeatedly emphasized the importance of his Slough Creek discovery in his Allatoona Reservoir report. Miller added additional information with each reference. He noted that he found a number of vegetal-tempered sherds on a site alongside Slough Creek, 7 miles west of Atco (Miller n.d.: 47) and later that he had recovered a number of Stallings Island Punctate sherds from the site on Slough Creek " which would indicate that this ware was present in limited quantities and that the Stallings Island series possibly existed even though it was not found within the Allatoona Reservoir area" (Miller n.d.:188). Miller clearly felt that fiber-tempered pottery comparable to that found at Stallings Island and Lake Springs sites in eastern Georgia should occur in the Allatoona Reservoir. He noted that "comparable examples" have been found in northern Alabama, eastern Tennessee, and northern Florida which suggested to him that the type of pottery should not be excluded from occurring within this reservoir basin. He concluded his assessment by noting that, "We just did not find any site that contained this particular pottery type probably because we either did not locate those sites or dug deeply enough to locate it" (Miller n.d.:281).
The approximate location of Miller's fiber-tempered pottery site appears to have been in the vicinity of the Garfield site (9BR57, formerly 9BR99), which is a well-known Early Woodland, Kellog phase habitation site located at the confluence of Macedonia Slough and the Etowah River (Milanich 1975). Macedonia Slough is located about 7 miles west of the community of Atco. A previously cited map prepared during Robert Wauchope's earlier survey shows the stream, but he recorded no sites near the confluence (see Figure 6). It is interesting that Wauchope recovered his only fiber-tempered sherd from the Walt Jones site (9BR9), which is located in the same general area of the Etowah River valley (Wauchope 1966:231).
Miller found soapstone vessel fragments on most, if not all, of his excavated sites. Most of these sites contained evidence of Late Archaic occupation and about half produced fabric-marked pottery. Miller's report generally discusses soapstone vessels within the context of a pre-pottery tradition. Miller's report suggests that one site, 9CK72 (the Fields Bridge site), contained an intensive pre-pottery period occupation with little evidence of Early Woodland occupation. The excavations at 9CK72 produced a number of fragments of "hemispherical, thick-walled steatite pots"(Miller n.d.:73). Miller's report provides no context for these "pots" so we do not know if they were recovered from features or midden deposits. His report contains no specific references to ceremonial or mortuary significance.
Carl Miller's excavations in Allatoona were mainly Middle Woodland and Mississippian period occupations, but some Early Woodland material was recovered. A substantial portion of his discussion of the Early Woodland period relates to pottery types. Miller considered the possibility that the earliest sand-tempered fabric-marked pottery was equivalent in age with some of the fiber-tempered wares.
51

"A type that appeared to be coeval with the late vegetal-tempered ware was a coiled, fine grained sand-tempered ware bearing the impressions of a "cord wrapped stick or dowel" and called Dunlap Fabric Impressed (Fairbanks 1940). Wauchope (1948) noted that this was the earliest ware he recovered in any strength during the Early Woodland Period. Caldwell (1950) reported that this same type appeared in quantities at an early level and that it persisted over quite a time span in a number of sites along Stamp and Kellog Creeks. He assigned this ware to his Kellog Period in the present report" [Miller n.d.:47].
Miller also discussed pottery that he preferred to call Deptford. He stated that, "Following Dunlap Fabric Marked came Mossy Oak Simple Stamped and a check stamped which the present writer has identified as Deptford Check Stamped of the type description and which Caldwell has termed "Cartersville" check stamped in contradistinction to the type of the Georgia Coast" (Miller n.d.:282). Miller saw no need for a new pottery name because Deptford was entirely appropriate for Allatoona.
He made relatively few references to Early Woodland projectile points. In one instance he noted that the Deptford projectile points were "tanged of varying degrees and ranged from small to large in size with some isosceles triangular forms to supplement them" (Miller n.d.:263). Miller noted that sites with the earliest of the sherd types from the Allatoona sites (Deptford, Mossy Oak Simple Stamped, and Dunlap Fabric Impressed) produced "time accepted projectile forms" (Miller n.d.:243).
Miller concluded that these early sites were small riverine settlements in which pits were used as depositories for food storage, for midden material, for cooking purposes, and even for burial purposes" (Miller n.d.:243). He suggested that the living areas associated with the early pottery types could be recognized by roughly circular ground-packed areas containing either hearth areas or fire basins. He viewed the absence of post molds as indicating the structure supports were not inserted deeply into the ground which meant that after abandonment, all indications of the shelters were lost. He further noted that "These shelters probably resembled the earlier pre-ceramic structures" (Miller n.d.:244).
Implications of Caldwell's and Miller's Work in Allatoona Reservoir for the Current Study
The Allatoona Reservoir excavations were conducted in much the same manner as the older WPA-era work carried out throughout the southeast (Lyon 1996) which means that the quality of the excavations and artifact collecting techniques (lack of screening) limits the degree to which their work can be interpreted today. However, Bowen's (1989) reexamination of the material remains from the Stamp Creek and Kellog sites has demonstrated that meaningful information may be retrieved from the collections. Caldwell used detailed analysis as a means of interpreting sites within the context of culture history and culture change (Johnson 2004:19) while Miller seemed content with the traditional presentation of "trait lists." The excavations in Allatoona Reservoir provided information that refined the chronology of the Woodland and Mississippian periods in Northwest Georgia but contributed relatively little to our understanding of what may have transpired during the Late Archaic to Early Woodland transition. Their excavations focused on pottery-producing sites and relatively little investigation was directed to earlier components. Caldwell's recognition of a Late Archaic Stamp Creek focus is meaningful, but his work was limited by a lack of understanding of Archaic projectile point types. However, Caldwell's concept of a cultural or temporal discontinuity between the Late Archaic and Early Woodland has remained a viable research topic.
52