<%BANNER%>

A Complex Web of History and Artifact Types in the Early Archaic Southeast

MISSING IMAGE

Material Information

Title:
A Complex Web of History and Artifact Types in the Early Archaic Southeast
Physical Description:
1 online resource (385 p.)
Language:
english
Creator:
Bridgman, Kara A
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Anthropology
Committee Chair:
Sassaman, Kenneth Edward, Jr
Committee Members:
Heckenberger, Michael Joseph
Wallis, Neill Jansen
Binford, Michael William
Goodyear Iii, Albert C

Subjects

Subjects / Keywords:
archaic -- hunter-gatherers -- side-notched -- southeast
Anthropology -- Dissertations, Academic -- UF
Genre:
Anthropology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
This research project was designed to answer the question: how are social boundaries, and the processes that shape and transform them, manifested in patterned variation in material culture? I attempted to answer this question with respect to the Early Side-Notched Horizon of the American Southeast, and analyzed numerous (n=2,020) artifacts from two classes of flaked stone tools: side-notched hafted bifaces and side-notched unifaces. Artifacts in these tool classes have been given specific type names (including Taylor, Bolen, and Big Sandy for hafted bifaces, and Edgefield Scraper for unifaces), and have been treated as if they are isomorphic with distinct geographic regions. This research was conducted using artifact collections from throughout seven drainage systems in the Coastal Plain Southeast. These areas include the following: the Santee-Cooper, Savannah-Ogeechee, Ocmulgee, Flint,Chattahoochee, Aucilla-Suwannee, and Tampa Bay drainage systems. I identified evidence for subregional technological traditions within the Early Side-Notched Horizon, and suggest that the origin and persistence of these traditions can be traced to similar Paleoindian traditions. Also, I identified specific areas of regional differentiation, demarcated by a centrally located boundary zone. These areas may correspond to the geographic ranges of distinct social groups. With the presumption of approximate contemporaneity for artifacts from the Early Side-Notched Horizon, I identified a high degree of variation for artifacts from the boundary zone, contrasting with the centers of zones where hafted biface types were defined. One key result of this research project has been that basal configuration of side-notched artifacts appears to be directly correlated to geography throughout the Coastal Plain. I propose that the Coastal Plain Southeast likely was home to three macroband territories; the Santee-Cooper/Savannah-Ogeechee, the Flint/Chattahoochee, and the Aucilla-Suwannee/Tampa Bay. Also, I propose that macrobands were characterized by regular cross-drainage mobility, and that macroband group members aggregated periodically at the Ocmulgee River, for marriage alliances, sharing, and trade.  I provide evidence for large-scale sharing networks, and suggest that a multiscalar and macroregional approach is essential to studies of social reproduction for the Early Archaic Southeast.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Kara A Bridgman.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Sassaman, Kenneth Edward, Jr.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045361:00001

MISSING IMAGE

Material Information

Title:
A Complex Web of History and Artifact Types in the Early Archaic Southeast
Physical Description:
1 online resource (385 p.)
Language:
english
Creator:
Bridgman, Kara A
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Anthropology
Committee Chair:
Sassaman, Kenneth Edward, Jr
Committee Members:
Heckenberger, Michael Joseph
Wallis, Neill Jansen
Binford, Michael William
Goodyear Iii, Albert C

Subjects

Subjects / Keywords:
archaic -- hunter-gatherers -- side-notched -- southeast
Anthropology -- Dissertations, Academic -- UF
Genre:
Anthropology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
This research project was designed to answer the question: how are social boundaries, and the processes that shape and transform them, manifested in patterned variation in material culture? I attempted to answer this question with respect to the Early Side-Notched Horizon of the American Southeast, and analyzed numerous (n=2,020) artifacts from two classes of flaked stone tools: side-notched hafted bifaces and side-notched unifaces. Artifacts in these tool classes have been given specific type names (including Taylor, Bolen, and Big Sandy for hafted bifaces, and Edgefield Scraper for unifaces), and have been treated as if they are isomorphic with distinct geographic regions. This research was conducted using artifact collections from throughout seven drainage systems in the Coastal Plain Southeast. These areas include the following: the Santee-Cooper, Savannah-Ogeechee, Ocmulgee, Flint,Chattahoochee, Aucilla-Suwannee, and Tampa Bay drainage systems. I identified evidence for subregional technological traditions within the Early Side-Notched Horizon, and suggest that the origin and persistence of these traditions can be traced to similar Paleoindian traditions. Also, I identified specific areas of regional differentiation, demarcated by a centrally located boundary zone. These areas may correspond to the geographic ranges of distinct social groups. With the presumption of approximate contemporaneity for artifacts from the Early Side-Notched Horizon, I identified a high degree of variation for artifacts from the boundary zone, contrasting with the centers of zones where hafted biface types were defined. One key result of this research project has been that basal configuration of side-notched artifacts appears to be directly correlated to geography throughout the Coastal Plain. I propose that the Coastal Plain Southeast likely was home to three macroband territories; the Santee-Cooper/Savannah-Ogeechee, the Flint/Chattahoochee, and the Aucilla-Suwannee/Tampa Bay. Also, I propose that macrobands were characterized by regular cross-drainage mobility, and that macroband group members aggregated periodically at the Ocmulgee River, for marriage alliances, sharing, and trade.  I provide evidence for large-scale sharing networks, and suggest that a multiscalar and macroregional approach is essential to studies of social reproduction for the Early Archaic Southeast.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Kara A Bridgman.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Sassaman, Kenneth Edward, Jr.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045361:00001


This item has the following downloads:


Full Text

PAGE 1

1 A COMPLEX WEB OF HISTORY AND ARTIFACT TYPES IN THE EARLY ARCHAIC SOUTHEAST By KARA BRIDGMAN SWEENEY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Kara Bridgman Sweeney

PAGE 3

3 ACKNOWLEDGMENTS Most deserving of my acknowledgment are the indigenous American groups who created the archaeological record that forms the basis of this research. I hope that this document will serve as a partial history of the Early Archaic, which so long has been releg ated to prehistory. I would very much like to thank the numerous individuals who shared their artifact collections and local knowledge with me. I especially want to express appreciation to those artifact collectors who make a concerted effort to share t he information they have with archaeologists, and who assist in the recording of archaeological sites. In Alabama, I thank Ed and Richard Kilborn. I would also like to thank Don Marley and Steve Lamb for their contributions, as well as one individual who asked that he not be named in my dissertation (in this document, I have called him X, as requested). In South Carolina, I thank Doug Boehme, Darby, Kevin, and Ted Hebert, Gary LeCroy, Earline Mitchum, George Neil, Jimmy Skinner, and Larry Strong. Becky S huler of the Elloree Heritage Museum and Barbara Butler of the South Carolina Bank and Trust of Elloree deserve thanks for allowing access to artifact collections at their repositories. In Georgia, I thank Danny Greenway, Seaborn Roberts, Danny Ely, Emery Fennell, John Arena, Larry Meadows, Marvin Singletary, and John Whatley. I also thank Bud Carter, Mike Warren, Ann Parker Gilmore, and Mary and Tommy Bailey. In Florida, I thank Brian Evensen, Jerry Spencer, Jim Drawdy, Alvin Hendrix, and Tom Pertierra. I also thank Becky and Bill Dean. I would also like to thank several people from the numerous institutions at which I either performed research, or else more generally consulted people regarding my project. In North

PAGE 4

4 Carolina, I thank Randy Daniel of East Car olina University, and Joel Gunn of the University of North Carolina at Greensboro. In South Carolina, I thank Tommy Charles, Chris Gillam, Al Goodyear, Daryl Miller, and Kenn Steffy, all of whom are or were affiliated with the South Carolina Institute of A rchaeology and Anthropology (SCIAA) at the University of South Carolina in Columbia, during my periods of research. In Alabama, I thank Eugene Futato and Robert Clouse of the David DeJarnette Res earch Center in Moundville, Philip Carr of the University of South Alabama and Mac Brooms of Troy University In Georgia, I thank Mark Williams at the Georgia Archaeological Site Files and David Leigh of the University of Georgia, as well as Frankie Snow of South Georgia College. I also thank Scott Jones and Jerald Ledbetter of Southern Research, and Lisa OSteen of New South Associates, as well as Dan Elliott of the Lamar Institute. In Florida, I thank Bob Austin and Ann Stokes of Southeastern Archaeological Research, Inc., Brinnen Carter of the National Park Servi ce, Dave Dickel and Jim Dunbar of the Florida Bureau of Archaeological Research, and Harley Means of the Florida Geological Survey. I also thank Barbara Purdy and Jamie Waggoner at the University of Florida, as well as Asa Randall of the Laboratory of Sout heastern Archaeology at the University of Florida. These individuals all assisted me in tracking down collections of interest, or made time to discuss my research throughout this project. Bob Austin generously identified the raw material sources of a larg e sample of the collections I analyzed from the northern and Tampa Bay regions of Florida. I also learned a great deal from informal discussions with numerous Southeastern archaeologists since I began this project. I sincerely apologize for not naming all of these kind people here. I be nefit ed enormously from the contributions of all of these professionals to my research.

PAGE 5

5 Todd Braje, Diana Byrne, and Neil Jacobs deserve thanks for sharing their time and their statistics knowledge. Todd Braje deserves specia l thanks for his continued friendship and support as a valued colleague, and Neil Jacobs generously wrote script to enhance my statistical analyses. I also want to recognize those generous individuals who provided me with a place to stay during my research trips. This support was greatly appreciated. These individuals include Mary Askew, Jane Anne Blakney Bailey, Brendon Bailey, Meggan Blessing, Alysia Bridgman, Jen Coody, Eugene Futato, Asa Randall, Jane Sharp, and Andrea and John Whatley. Also, the brothe rs at Mepkin Abbey in Moncks Corner, South Carolina generously welcomed me as a lecturer/retreatant for a portion of my research efforts. My dissertation committee members deserve special thanks for their participation and suggestions. While this work bene fitted greatly from discussions with numerous researchers, Ken Sassaman and Al Goodyear have been instrumental mentors in my development as an archaeologist and scholar. Much of my research in South Carolina and eastern Georgia was made possible by generou s grants from the Archaeological Research Trust (2003), and the Archaeological Society of South Carolina (2004). I also am truly grateful for my loving and supportive family, especially my parents Mary Ann and Howard, my sister s Laura and Alysia, and my P apa. Finally, and most especially, I would like to thank my husband, Alex for his invaluable assistance, humor, love, and support throughout this research effort. His countless contributions and insights propelled me toward the completion of this project.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................3 LIST OF TABLES .........................................................................................................................10 LIST OF FIGURES .......................................................................................................................13 ABSTRACT ...................................................................................................................................17 CHAPTER 1 INTRODUCTION AND PROBLEM ORIENTATION ........................................................19 Patterned Variation in the Early SideNotched Horizon ........................................................21 Structure of the Dissertation ...................................................................................................25 Summary ................................................................................................................................27 2 HUNTER GATHERER SOCIAL NETWORKS AND PLACE MAKING .........................36 Hunter Gatherers and Social Networks ..................................................................................36 PlaceMaking ..........................................................................................................................42 Summary ................................................................................................................................47 3 GEOGRAPHY AND CULTURE HISTORY ........................................................................49 The Pl eistoceneHolocene Transition ....................................................................................50 The Late Pleistocene Environmental Setting and Culture History .................................50 The Early Holocene En vironmental Setting and Culture History ..................................56 Early Holocene environmental setting ....................................................................57 Early Side Notched Horizon artifact types .............................................................62 Early Archaic sites and settlement models ..............................................................71 Raw Material Sources and Selection ......................................................................................80 Chert ................................................................................................................................81 So uth Carolina ................................................................................................................83 Georgia ............................................................................................................................84 Florida .............................................................................................................................86 Alabama ..........................................................................................................................87 A Regional View of Coastal Plain Chert Distribution ...................................................88 Summary ................................................................................................................................89 4 RE SEAR CH METHODS AND DATA SETS .......................................................................95

PAGE 7

7 Attributes for Analysis ...........................................................................................................95 Individual Objects ...........................................................................................................96 In dividual Artifact Attributes ..........................................................................................97 Hafted bifaces ..........................................................................................................98 Unifaces .................................................................................................................100 Statistical Techniques ..........................................................................................................101 Descriptive Statistics .....................................................................................................101 Inferential Statistics ......................................................................................................103 Tests of Associ ation ......................................................................................................104 Sample Selection and t he Use of Collected Material ...........................................................104 Research Area and Data Sets Used .....................................................................................106 South Carolina ..............................................................................................................107 Lakes Marion and Moultrie region artifacts ..........................................................110 Allendale County artifacts ....................................................................................111 Georgia ..........................................................................................................................112 Burke County artifacts ...........................................................................................114 Coffee C ounty artifacts ........................................................................................115 Dooly County artifacts ..........................................................................................117 Worth County artifacts ..........................................................................................118 Early County artifacts ............................................................................................118 Miller County artifacts ..........................................................................................119 Alabama ........................................................................................................................119 Florida ........................................................................................................................... 120 Jackson County artifacts ........................................................................................120 Jefferson County artifacts ......................................................................................121 Taylor County artifacts ..........................................................................................121 L afayette County artifacts .....................................................................................122 Santa Fe and Suwannee Rivers region artifacts ....................................................125 Tampa Bay region artifacts ...................................................................................126 Mapping Artifac ts to Drainage and Watershed ....................................................................129 Drainage Systems of the Atlantic Watershed ...............................................................129 Drainage S ystems of the Gulf Watershed .....................................................................131 Analyses of Patterned Variation ...........................................................................................133 Summary ..............................................................................................................................134 5 RESULTS .............................................................................................................................148 Side Notched Haf ted Bifaces across Drainages ...................................................................148 Mapping Nominal Data .................................................................................................148 SanteeCooper hafted bifaces ................................................................................149 SavannahOgeechee hafted bifaces .......................................................................149 Ocmulgee hafted bifaces .......................................................................................150 Flint hafted bifa ces ................................................................................................150 Chattahoochee hafted bifaces ................................................................................150

PAGE 8

8 Auc illa Suwannee hafted bifaces ..........................................................................151 Tampa Bay hafted bifaces .....................................................................................151 Compar isons across Drainages ..............................................................................151 Mapping RatioS cale Data ............................................................................................156 SanteeCooper hafted bifaces ................................................................................158 SavannahOgeechee hafted bifaces .......................................................................159 Ocmu lgee hafted bifaces .......................................................................................159 Flint hafted bifaces ................................................................................................160 Chattahoochee hafted bifaces ................................................................................160 Aucilla Suwannee hafted bi faces ..........................................................................161 Tampa Bay hafted bifaces .....................................................................................162 Discussion for Bla de Portion of Hafted Bifaces ...........................................................162 Discussion for Ha ft Element of Hafted Bifaces ............................................................174 Side Notch ed Unifaces across Drainages .............................................................................186 Mapping Nominal Data .................................................................................................187 SanteeCooper unifaces .........................................................................................187 Savannah Ogeechee unifaces ................................................................................188 Ocmulgee unifaces ................................................................................................188 Aucilla Suwannee unifaces ...................................................................................189 Tampa Bay unifaces ..............................................................................................189 Comparisons across Drainages ..............................................................................189 Mapping RatioS cale Data ............................................................................................193 SanteeCooper unifaces .......................................................................................194 Savannah Ogeechee unifaces ...............................................................................195 Ocmulgee unifaces ...............................................................................................196 Aucilla Suwannee unifaces ..................................................................................197 Tampa Bay unifaces .............................................................................................198 Discussion f or Blade Portion of Unifaces .....................................................................199 Discus sion for Haft Element of Unifaces .....................................................................209 Comparing Side Notched Hafted Bifaces to Edgefield Scrapers across Drainages .............220 Zones of Patterned Variation across Drainages ...................................................................234 Pilot Study for Quarry Cluster Sourcing of Florida Bifaces ................................................241 Lake Bird ......................................................................................................................243 Tampa Bay ....................................................................................................................244 Tampa Bay area .....................................................................................................245 St. Petersburg area .................................................................................................246 Pinellas County ......................................................................................................247 Hillsborough County, Harris Grove Collection ....................................................249 Hillsboroug h County, Harney Flats .......................................................................250 Summary of Pilot Study ................................................................................................250 Summary ..............................................................................................................................253 6 CONCLUSIONS AND RECOMMENDATIONS ...............................................................290

PAGE 9

9 Hunter Gather ers during the Early Archaic .........................................................................290 Subregional Trends in th e Early SideNotched Horizon ......................................................293 San teeCooper/Savannah Ogeechee .............................................................................296 Flint/Chattahoochee ......................................................................................................297 Aucilla Suwannee/Tampa Bay .....................................................................................297 Early Archaic Southeast Macroband Territory Characteristics ....................................298 Large S cale Social Interaction in the Early Archaic ............................................................302 Summary and Recomme ndations for Future Research ........................................................305 APPENDIX A BASAL CONFIGURATION FOR HAFTED BIFACES ....................................................313 B METRIC ATTRIBUTES OF HAFTED BIFACES BY SITE AND COLLECTION AREA ...................................................................................................................................315 C METRIC ATTRIBUTES OF HAFTED BIFACES BY DRAINAGE SYSTEM ................332 D BASAL CONFIGURATION FOR HAFTED UNIFACES .................................................339 E METRIC ATTRIBUTES OF HAFTED UNIFACES BY SITE AND COLLECTION AREA ...................................................................................................................................341 F METRIC ATTRIBUTES OF HAFTED UNIFACES BY DRAINAGE SYSTEM .............354 REFERENCES CITED ................................................................................................................359 BIOGRAPHICAL SKETCH .......................................................................................................385

PAGE 10

10 LIST OF TABLES Table page 41 Data sets used in analysis .................................................................................................108 42 Frequencies of artifacts from assemblages throughout the Coastal Plain ........................109 43 Quarry cluster sourcing for side notched hafted bifaces from 8LF54 .............................124 51 Absolute and relative frequencies of hafted bifaces by drainage and basal configuration ....................................................................................................................152 52 Hafted biface basal configuration (Observed) .................................................................155 53 Hafted biface basal configuration (Expected) ..................................................................155 54 Summary of blade length (mm) data for hafted bifaces ..................................................163 55 Results of Leven es tests conducted for ratio scale data .................................................166 56 Summary of shoulder width (mm) data for hafted bifaces ..............................................167 57 Ratios of mean blade length to mean shoulder width (mm) for side notched hafted bifaces ...................................................................................................................169 58 Summary of maximum thickness (mm) data for hafted bifaces ......................................170 59 Summary of shoulder length (mm) data for hafted bifaces .............................................175 510 Summary of between notch width (mm) data for hafted bifaces ....................................178 511 Summary of notch heig ht (mm) data for hafted bifaces .................................................181 512 Summary of base width (mm) data for hafted bifaces .....................................................183 513 Absolute and relative frequencies of hafted unifaces by drainage and basal configuration ....................................................................................................................191 514 Hafted uniface basal configuration (Observed) ...............................................................193 515 Hafted uniface basal configuration (Expected) ................................................................193 516 Summary of left blade length (mm) data for hafted unifaces ..........................................199

PAGE 11

11 517 Summary of right blade length (mm) data for hafted unifaces ........................................201 518 Sum mary of shoulder width (mm) data for hafted unifaces ............................................203 519 Summary of primary edge angle (degrees) for hafted unifaces .......................................205 520 Summary of primary edge angle (degrees) for left handed haft ed unifaces ....................206 521 Summary of maximum thickness (mm) data for hafted unifaces ....................................207 522 Summary of left shoulder length (mm) data for hafted unifaces .....................................210 523 Summary of right shoulder length (mm) data for hafted unifaces ...................................212 524 Summary of between notch width (mm) data for hafted unifaces .................................214 525 Summary of left notch height (mm) data for hafted unifaces ..........................................215 526 Summary of right notch height (mm) data for hafted unifaces ........................................217 527 Summary of base width (mm) data for hafted unifaces ...................................................218 528 Trends in basal configuration for hafted bifaces and unifaces ........................................221 529 Trends in basal configuration for hafted bifaces and unifaces by multi drainage subregion ...........................................................................................................225 530 Ratios of hafted bifaces to uni faces throughout the Coastal Plain...................................234 531 Descriptive statistics for the center of the Taylor Zone ...................................................236 532 Descriptive statistics for the center of the Bolen Zone ....................................................236 533 Descriptive statistics for the Ocmulgee boundary area ....................................................237 534 Results of Levenes tests conducted f or ratio scale data (zoneboundary test case) ...........................................................................................................................239 535 Absolute and relative frequencies of hafted bifaces by sample and basal configuration ....................................................................................................................240 536 Quarry cluster sourcing for side notched hafted bifaces from Lake Bird .......................243 537 Quarry cluster sourcing for side notched hafted bifaces from the Tampa Bay area .......246

PAGE 12

12 538 Quarry cluster sourcing for side notched hafted bifaces from the St. Petersburg area .................................................................................................................247 539 Quarry cluster sourcing for side notched hafted bifaces from Pinellas County ..............................................................................................................................248 540 Quarry cluster sourcing for side notched hafted bifaces from the Harris Grove Collection ..............................................................................................................250 541 Estimated maximum mobility using quarry cluster sourcing data ...................................252 61 Percentages of basal configuration by macroband territory ............................................307

PAGE 13

13 LIST OF FIGURES Figure page 11 Side notched hafted bifac es from central South Carolina ................................................ 29 12 Side notched hafted bifac es from southcentral Georgia. ..................................................30 13 Side notched hafted bifa ces from north central Florida ................................................... 31 14 Side notched haf ted bifaces from Tampa Bay, Florid a ................................................... 32 15 Side notched hafted bifa ces from the Florida panhandle ................................................. 33 16 A sample of Edgefield Scrapers from southwestern South Carolina (top) and southcentr al Georgia (bottom) ....................................................................................................34 17 Major rivers in the Coastal Plain Southeast .......................................................................35 31 Band ranges for the South Atlantic Macroband in the Early Archaic ...............................90 32 Project research area and the Floridan aquifer ...................................................................91 33 Coastal Plain chert dis tributions throughout Georgia ........................................................92 34 The locations of Florida chert quarry clusters ...................................................................93 35 The distribution of high quality Coastal Plain chert throughout the study region ............94 41 Measurements taken for side notched hafted bifaces ......................................................136 42 Measuremen ts taken for side notched hafted unifaces ....................................................137 43 Research area in the Coastal Plain Southeast ..................................................................138 44 Locations of Lakes Marion and Moultrie, and Allendale County, South Carolina; and Burke County, Georgia sites and collection areas ...................................................139 45 Locations of Lakes Marion and Moultrie, and Allendale County, South Carolina; and Burke County, Georgia sites and collection areas, in the context of Coastal Plai n cherts .......................................................................................................................140 46 Locations of Coffee, Dooly, and Worth counties, Georgia sites and collection areas .................................................................................................................................141

PAGE 14

14 47 Location of Feronia sites in the context of Coastal Plain chert distribution ....................142 48 Locations of Early and Miller Counties, Georgia; Barbour, Geneva, Henry, and Houston counties, Alabama; and Jackson County, Florida, sites and collection areas .................................................................................................................................143 49 Locations of J efferson, Lafayette, and Taylor counties, Florida sites and collection areas .................................................................................................................................144 410 Locations of Florida archaeological sites in the context of chert quarry clusters ............145 411 Locations of Hernando, Hillsborough, Pasco, and Pinellas counties, Florida sites and collection areas ..........................................................................................................146 412 The Atlantic/Gulf watershed boundary in the context of Coastal Plain drainages ..........147 51 Hafted biface basal configuration by percentage ( n=1,723) ............................................255 52 Differentiation in basal configuration for hafted bifaces throughout the Coastal Plain .................................................................................................................................256 53 Major drainage systems and the distribution of Coastal Plain chert throughout the study region ......................................................................................................................257 54 Possibly h ypertrophic hafted bifaces recovered from the Waccasassa and Santa Fe Rivers (Artifact IDs 505 [top] and 506 [bottom]) ...........................................................258 55 Normal distribution p ercentile plot for hafted biface blade length ..................................259 56 Normal distribution p ercentile plot for hafted biface shoulder width ..............................260 57 Normal d istribution p ercentile plot for hafted biface thickness .......................................261 58 Trends for blade length, shoulder width, and maximum thickness for hafted bifaces throughout the Coastal Plain ............................................................................................262 59 Variation in the blade portion of hafted bifaces throughout the Coastal Plain ................263 510 Normal distribution p ercentile plot for hafted biface shoulder length .............................264 511 Normal distribution p ercentile plot for hafted biface between notch width ....................265 512 Normal distribution p ercentile plot for hafted biface notch height ..................................266

PAGE 15

15 513 Normal distribution p ercentile plot for hafted bifac e base width ....................................267 514 Variation in the haft element of hafted bifaces throughout the Coastal Plain .................268 515 Hafted uniface basal configuration by percentage ( n=241) ............................................268 516 Differentiation in basal configuration for hafted unifaces throughout the Coastal Plain .................................................................................................................................269 517 Normal distribution p ercentile plot for uniface left blade length ....................................270 518 Normal distribution p ercentile plot for uniface right blade length ..................................271 519 Normal distribution p ercentile plot for uniface shoulder width.......................................272 520 Normal distribution p ercentile plot f or uniface thickness ................................................273 521 Trends for blade length, shoulder width, and maximum thickness for hafted unifaces throughout the Coastal Plain ............................................................................................274 522 Variation in the blade portion of ha fted unifaces throughout the Coastal Plain ..............275 523 Normal distribution p ercentile plot for uniface left shoulder length ...............................276 524 Normal distribution p ercentile plot for uniface right shoulder l ength .............................277 525 Normal distribution p erc entile plot for uniface between notch width .............................278 526 Normal distribution p ercentile plot for uniface left notch height ....................................279 527 Normal distribution p ercentile plot for uniface right notch height ..................................280 528 Normal distribution p ercentile plot for uniface base width .............................................281 529 Variation in the haft element of hafted unifaces throughout the Coastal Plain ...............282 530 Normal distribution p ercentile plot for hafted biface blade length (zone boundary test case) .................................................................................................................................283 531 Normal distribution percentile plot for hafted biface shoulder width (zone boundary test case) ................................................................................................................................ .284 532 Normal distribution p ercentile plot for hafted biface thickness (zone boundary test case) .................................................................................................................................285

PAGE 16

16 533 Normal distribution p ercentile plot for hafted biface shoulder length (zone boundary test case) ...........................................................................................................................286 534 Normal distribution p ercentile plot for hafted biface between notch width (zone boundary test case) ...........................................................................................................287 535 Normal distribution p ercentile plot for hafted biface notch height (zone boundary test case) .................................................................................................................................288 536 Normal distribution p ercentile plot for hafted biface base width (zone boundary test case) .................................................................................................................................289 61 Proposed Coastal Plain macroband territories .................................................................312

PAGE 17

17 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy A COMPLEX WEB OF HISTORY AND ARTIFACT TYPES IN THE EARLY ARCHAIC SOUTHEAST By Kara Bridgman Sweeney August 2013 Chair: Kenneth E. Sassaman Major: Anthropology This research project was designed to answer the question: how are social boundaries, and the processes that shape and transform them, manifested in patterned variation in material culture ? I attempted to answer this question with respect to the Early Side Notched Horizon of the American Southeast, and analyzed numerous ( n=2,020) artifacts from two classes of flaked stone tools: side notched hafted bifa ces and sidenotched unifaces. Artifacts in these tool classes have been given specific type names (including Taylor, Bolen, an d Big Sandy for hafted bifaces, and Edgefield Scraper for unifaces), and have been treated as if they are isomorphic with distinct geographic regions. This research was conducted using artifact collections from throughout seven drainage systems i n the Coastal Plain Southeast. These areas include the following: the SanteeCooper, Savannah Ogeechee, Ocmulgee, Flint, Chattahoochee, AucillaSuwannee, and Tampa Bay drainage systems. I identified evidence for subregional technological traditions within the Early Side Notched Horizon, and suggest that the origin and persistence of these traditions can be traced to similar Paleoindian traditions. Also, I identified specific areas of regional

PAGE 18

18 differentiation, demarcated by a centrally located boundary zone. These areas may correspond to the geographic ranges of distinc t social groups. W ith the presumption of approximate contemporaneity for artifacts from the Early SideNotched Horizon, I identified a high degree of variation for artifacts from the boundary zone, contrasting with the centers of zones where hafted biface types were defined. One key result of this research project has been that basal configuration of side notched artifacts appears to be directly correlated to geography throughout the Coastal Plain. I propose that the Coastal Plain Southeast likely was home to three macroband territories; the Santee Cooper/SavannahOgeechee, the Flint/Chattahoochee, and the Aucilla Suwannee/Tampa Bay. Also, I propose that macrobands were characterized by regular cross drainage mobility, and that macroband group members aggre gated periodically at the Ocmulgee River, for marriage alliances, sharing, and trade. I provide evidence for large scale sharing networks, and suggest that a multiscalar and macroregional approach is essential to studies of social reproduction for the Earl y Archaic Southeast.

PAGE 19

19 CHAPTER 1 INTRODUCTION AND PROBLEM ORIENTATION H ow are social boundaries, and the processes that shape and transform them, manifested in patterned variation in material culture ? Arc haeologists have long sought to make sense of artif act variation using typologies that group similar artifacts together and dissimilar artifacts apart. Too often, archaeologists then equate similar artifacts (i.e., artifact types) with social collectives (e.g., tribes), which are then made distinct from other such groups by disjunctures in the distribution of artifact types through time and space. This was the modus operandi of the culture historical paradigm of Americanist archaeology from about 1910 to 1960 (Lyman et a l. 1997:1) tha t continues to influence the way archaeologists interpret material culture. Culture historical types, as traditionally applied unduly mask variation in social identity and affiliation, resulting in overly simplistic and normative views of ancient cultures More recent research in archaeology ( Barrett 1994; Dobres and Robb 2000; Jones 1997; Sassaman 1998, 2010; Stark 1998) emphasizes the active role of material culture in the assertion and manipulation of identity in the ongoing process of social reproduction. This process operates at multiple scales of time and space, from the interpersonal learning among members of a co resident group to the regional expressions of tradi tion that archaeologists recognize as cultural horizons (after Willey and Phillips 2001 [1958]). How these micro and macro scalar processes relate to one another is an enduring archaeological problem, one that is best served by multiscalar analytical appr oaches to material culture. The Early Side Notched H orizon of the American Southeast exemplifies the complex and multiscalar relationship between artifact variation and cu ltural affiliation among hunter gatherer populations of the Early Holocene. Archaeolo gists have long recognized the widespread

PAGE 20

20 distribution of side notched hafted bifaces across the Southeast, and modern radiometric age estimates show that these forms were made and used for approximately 700 calendar years t hroughout this region, from roug hly 10,300 to 9600 ca l B.P.1 (Driskell 1996:325326; Faught et al. 2003:17; Morse 1996) Side notched hafted bifaces are exemplars of the horizon marker imagined by the architects of North American culture history, when they described the expression of c ultural elements or traits over broad geographical areas where approximate con temporaneity could be presumed. Using horizon markers such as the side notched hafted biface, it was proposed, would allow for the investigation of past c ultural relationships (W illey and Phillips 2001 [1958]:34). Operating at local scales of practice, however, archaeologists have recognized subregional variation in side notched hafted bifaces. Over the years archaeologists have proposed type names to distinguish side notched bifa ces found in South Carolina (Taylor) from those found in Florida (Bolen) and from those found in Alabama (Big Sandy) How these differences in form came into existence and were reproduced over several centuries is not often questioned. As noted above, typologies that recognize diff erences in artifact form implicitly assume differences in cultural identity, as if artifacts are people and clusters of like forms may be equated to the social collectives to which those individuals belonged. It is certainly the case that variation in side notched forms maps onto geography in ways that allow archaeologists to effectively distinguish typical Florida examples from those found in South Carolina for instance I t is also arguably the case that the people who made and used these various forms shared ancestry, and likely interacted on a regular basis despite their geographic separation. Like many hunte r gath erer populations worldwide (Gould 1982; Grinker 1 All radiocarbon estimates are reported in calendar years before present ( or cal B.P. ) following Anderson and Sassaman 2012: 2, 5

PAGE 21

21 1994; Kent 1992, 1993; Wiessner 1977, 1982), those of the Early S ide Notched H orizon in the American Southeast may have moved routinely both within and between habitual use areas. As such, their lives were at once local and regional, resulting in a material record of both cultural affinity at the grand scale and cultural distinction at smaller scales. How patterned variation in material culture results from such multiscalar social and cultural process is the subject of this dissertation. Patterned Variation in the Early Side Notched Horizon My contribution to an archaeol ogical understanding of multiscalar soc ial and cultural process is based on the analysis of large collections of flaked stone artifacts from throughout the lower Southeast. The data considered for this study were derived from two classes of Early Archaic f laked stone artifacts: side notched hafted bifaces known as Taylor, Bolen, and Big Sandy, and side notched hafted uniface s known as Edgefield Scrapers. Figures 1 1 through 15 show examples of typical side notched hafted bifaces. Figure 1 6 shows examples of Edgefield Scrapers. I collected data on artifact s from collections in the vicinity of numerous waterways shown in Figure 17, occupying the area of the lower Southeast that extends across the physiographic divide between Atlantic and Gulf draining rivers in the region, essentially a northwest southeast diagonal line through the state of Georgia Over 2,000 artifac ts were examined in this study to document variation in hafted flaked stone tools of the Early SideNotched H orizon. Hafted flaked stone tools are notably useful for trackin g variation in tool design because they often retain, in archaeological context, the form and technological execution of the toolmakers intended design. In general, flake d stone tools are subject t o change over their uselives as tools become worn or damaged through use and tool

PAGE 22

22 users maintain and rejuvenate tools through further flaking. The blades of knives and the tips of projectiles made from flaked stone are examples of tool elements that are e specially prone to attrition and breakage. The haft elements of these tools, however, are typically bound into a handle (haft) and thus protected from certain forms of damage and inaccessible for modi fication by tool users. Haft elements can also break, and can be quickly repaired after removing them from the haft ( Flenniken and Raymond 1986) Nonetheless the haft elements of side notched hafted bifaces and Edgefield Scrapers offer good opportunity for documenting variations in tool design that were not routinely subject to alteration during tool use life. As such, these haft elements are a reasonably reliable measure of the intended design of at least the proximal (lower) portions of the tools (Austin and Mitchell 2010; Bissett 2003) Data presented in thi s dissertation corroborate to a certain extent the longperceived notion that variation in the haft elements of Early Side Notched Horizon tools maps on to regional geography in a way that allows archaeologists to infer the coexistence of at least three di stinct social groups. In this sense the extant types used to classify side notched hafted bifaces and unifaces appear to be meaningful in more than an analytical sense (i.e., real types, sensu Ford 1952:343). However, the data presented here also reveal a great deal of variation within each of the three subregions, suggesting that personnel and/or ideas were routinely crossing the geographic boundaries of inferred social groups. Additionally some places in the landscape, most notably the physiographic h inge area of the Atlantic Gulf drainage divide, contain assemblages that demonstrate a marked level of formal variation in side notched tools. Several lines of ancillary ev idence would suggest that such zonal interfaces were points of social aggregation for widely distributed groups. At a minimum, two scales of intergroup interactions

PAGE 23

23 can be inferred from these data, revealing at once the complexity of social processes responsible for archaeologically recognized boundaries, and the apparent permeability of such boundaries. Documenting and explaining patterned variation in 10,000 year old material culture is the proximate goal of this dissertation research, while its ultimate goal is to contribute to anthropological understanding of the genesis, reproducti on, and transformation of social boundaries among highly mobile human populations. While people of the Early Side Notched Ho rizon were not the first to settle the American Southeast, they were not far behind the initial settlers of the region. They evident ly traced ancestry to Paleoindian populations who colonized North Americ a before about 14,000 years ago, su ch that the Early Side Notched H orizon can be interpreted as an outgrowth of Paleoindian traditions ( Anderson and Sassaman 2012; Sassaman 2010:22). T he exact timing of the first colonists, their route(s ) into the continent, and their lineages of descent are currently unknown. Theories regarding the peopling of North America are rapidly changing as new discoveries and refined chronologies require archaeologists to rewrite prehistory ( Anderson and Sassaman 2012; Beck and Jones 2012; Fiedel and Morrow 2012) Since the 1930s, people of the Paleoindian tradition known as Clovis were regarded as the initial colonists of North America. Recent reevaluations of radiocarbon dates place the Clovis tradition of fluted point technology at ca. 13,25012,800 cal B.P. (Waters and Stafford 2007:1123) T he Clovis tradition coincides with the last portion of the Pleistocene when numerous species of mega mammals roamed the continent. Clovis points are found across virtually all of North America (Anderson et al. 2010) suggesting that colonization may have been rapid, perhaps concurrent with a wave of migration among many of the mega mammal species known to be tar gets of Clovis weaponry.

PAGE 24

24 With the discovery of sites predating Clovis, archaeologists are close to reaching consensus that Clovis was not the first Paleoamerican cultural tradition. Paleoamerican i s a recent term applied to the first inhabitants of North America, due in part to the uncertain origins of these groups (Anderson and Sassaman 2012:36). As Paleoindians may not have been the initial colonists of the Americas, they cannot be described as the earliest Paleoamericans. Currently, Clovis is recognize d as the earliest and most pervasive Paleoindian tradition. Other fluted or occasionally fluted forms (including Cumberland, Folsom, Quad, Redstone, Simpson, Suwannee, and Waist ed Clovis ) may date to later portions of the Paleoindian sequence (Thulman 2007). Irrespective of the precise nature of the emergence and spread of preClovis cultural traits, this ancestry pre configured the structure and organization of later regional populations in the Southeast. Thus, as Anderson ( 1990, 1996) has argued, certain places in the landscape had already become areas of habitual use, what he has called staging areas for the subsequent expansion of population. Obviously, changes in climate, ground cover, and the distribution of resources factor into long term histories of land use, but if what Anderson (1990, 1996) infers persisted from pre Clovis, to Clovis, and into the Early Side Notched H orizon, patterned variation in side notched tools is likely to be anchored to locations with roots going back thousands of years. In this research project, I remain mindful of potential changes in land use that would have disrupted the historical linkages of sociogeography, while also acknowledging that social memory about places of historical significance does not r equire the sort of cultural continuity or affinity implied by specific artifact types. As discussed in Cha pter 3, the Early Side Notched H orizon is arguably

PAGE 25

25 descended from the Clovis tradition, so we are obliged to look closely at Clovis and its immediate descendants to understand ensuing patterns of cultural affiliation and interaction. Structure of the Dissertation This dissertation is structured as follows. In Chapter 2, I provide an overview of hunter gatherer variability, reviewing evidence for social networking at g reat social and spatial scales. As these studies illustrate the primacy of social interactions to hunter gatherer groups who are connected through time and space, as well as practice, they may be appropriate analogues for the Early Side Notc hed Horizon. Also, I review recent considerations of the linkages among group, material production, and place. These concepts inform discussion of the placebased traditions arguably seen in the Early Side Notched Horizon. Chapter 3 provides an overview of the geography and culture history of the research area This chapter presents discussion of models for the initial settlement of the American Southeast by participants in the Clovis tradition. Increasing evidence indicates these groups were placeoriented, with staging areas serving as the foundation for early regionalization in the Southeast (Anderson 1990, 1996; Anderson and Sassaman 2012:50; Miller 2011; Smallwood 2012; Thulman 2006) The apparent em ergence of subregional traditions during the Paleoindian period has great relevance for contextualizing and modeling the subsequent Early Archaic period. This chapter includes a description of the Younger Dryas, which was the last major cold reversal of th e Pleistocene E poch, affecting later Paleoindian populations (Anderson and Sassaman 2012:3839). Additionally I provide an extended summary of the Early Holocene environmental setting, as this was the setting for the Early Side Notched Horizon. I also provide an introduction to numerous archaeological sites that have informed our knowledge of Early Archaic

PAGE 26

26 chronology, and discuss settlement models that have been proposed for the latter portion of the Early Archaic. Next, I offer a hypothesis for settlement during the earliest portion of the Early Archaic, as no extant mode ls effectively consider this period. Finally, I present a survey of high quality raw material sou rces throughout the region as many of these specific sources are known to have been used t o manufacture sidenotched tools (Austin and Mitchell 1999, 2010; Goad 1979; Goodyear et al. 1983; Goodyear et al. 1985; Upchurch et al. 1982) In Chapter 4, I introduce the specific data sets used in analysis, and describe how I divided these data sets according to closest associated major drainage and watershed to aid in interpretations of patterned variation at a regional scale. Also in this chapter, I prese nt the methods used for this research, as well as my justification for selecting a variety of specific attributes in that analysis. The results of data analyses are provided in Chapter 5. Here, I report on the patterned variation observed for side notched artifacts from throughout the Coastal Plain Southeast First, I describe the study results at the local scale, at the level of individual drainages. Next, I provide a view of patterned variation at the regional scale. The blade portions and haft elements of these tools are distinguished in broad comparisons of these data. One significant finding in this research is that basal configuration of side notched artifacts appears to be highly correlated to geographic region. I argue that t his patterning provides support for the existence of distinct subregional traditions within the Early Side Notched Horizon. Additional evidence for subregional traditions is provided by trends observed within ratioscale data. Also in Chapter 5, I present the results of a test case for certain subsets of data, comparing relative variance for the centers of subregional tradition zones, with the boundary between those zones. Finally, at the

PAGE 27

27 conclusion of Chapter 5, I relate the results of a pilot study for quarry cluster sourcing of hafted bifaces from Florida. This study indicates some degree of mobility between two major river drainages. In Chapter 6, I provide interpretations of the results of this study. I discuss two alternative models for the broad regional tradition that char acterizes the Early SideNotched Horizon. This cultural horizon may reflect the movements of highly mobile groups into new landscapes throughout time. In this scenario, the subregional traditions generally associated with the Taylor, Bolen, and Big Sandy hafted biface types developed as a function of consecutive occupations of distinct physiographic regions. Alternatively, the subregional traditions of the Early Side Notched Horizon may be a product of large scale social interaction. This possibility pres umes that the subregional traditions of the Early Side Notched Horizon are reflective of distinct and contemporary social groups. I argue that t his study of patterned variation lends support to a model of regional interaction for the Early Archaic. I iden tified specific areas of regional differentiation, demarcated by a centrally located boundary zone. These areas may correspond to the geographic ranges of distinct social groups With the presumption of approximate contemporaneity for artifacts from the Ea rly Side Notched Horizon, I identify a h igh degree of variation for artifacts from the boundary zone contrasting with the centers of zones where hafted biface types were defined. Summary In my dissertation research, I have used an explicitly critical perspective with respect to artifact types. I argue that the variation that exists within constructed artifact types must be

PAGE 28

28 explored, and that artifact variation must be appreciated at multip le scales, from local to interre gional In my research, I confront these issues with respect to the Early Side Notched Horizon of the American Southeast where reliance on artifact types remains a significant problem. In this study of patterned variation i n material culture, I attempt to identify the social boundaries and processes that characterized ancient hunter gatherer lifeways throughout the Southeastern Coastal Plain. There must have been mechanisms at work in the maintenance of the larger collective (e.g., periodic aggregation), as well as places (subregionally) that were more habitually used. The advantage of a multiscalar approach to this research question is that it allows for investigation of social interaction and reproduction at both local ( sen su artifact type) and regional levels ( sensu horizon) This study demonstrates that a rchaeological evidence of interaction s at multiple scales does not preclude the emergence and persistence of subregional traditions, and that the genesis of such place bas ed traditions is arguably a product of historical processes. The complex ways in which social boundaries, and the processes that shape and transform them, are manifested in patterned variation in material culture are enduring anthropological issues, for which archaeological data have been under utilized.

PAGE 29

29 Figure 1 1. Side notched hafted bifaces from central South Carolina (Photos courtesy of the author)

PAGE 30

30 Figure 1 2. Side notched hafted bifaces from southcentral Georgia (Photos courtesy of the author)

PAGE 31

31 Figure 1 3. Side notched hafted bifaces from northcentral Florida (Photos courtesy of the author)

PAGE 32

32 Figure 1 4. Side notched hafted bifaces from Tampa Bay, Florida (Photos courtesy of the author)

PAGE 33

33 Figure 1 5. Side notched hafted bifaces from the Florida panhandle (Photos courtesy of the author)

PAGE 34

34 Figure 1 6. A sample of Edgefield Scrapers from southwestern South Carolina (top) and southcentral Georgia (bottom) (Photos courtesy of the author)

PAGE 35

35 Figure 17. Major rivers in the Coastal Plain Southeast

PAGE 36

36 CHAPTER 2 HUNTER GATHERER SOCIAL NETWORKS AND PLACE MAKING It is necessary to situate the Early SideNotched Horizon in an anthropological framework. Here, I provide an overview of hunter gatherer flexibility and variability, foregrounding evidence for large scale social and information sharing networks that may a ssist in modeling Early Archaic social boundaries. Also, I discuss ethnographic examples of place making and social boundary marking, which may be comparable to the Early Side Notched Horizon case study. Hunter Gatherers and Social Networks Identifying the various types of social interactions that existed in hunter gatherer societies has long been of particular interest to anthropologists ( Ames 2004; Lee and DeVore 1968) and archaeologists often look to ethnographic studies of hunter gatherers to assist in mode ling past lifeways. While u se of ethnographic anal ogy is not without problems, it is practically unavoidable for archaeologists (Ames 2004; Kelly 1995; Wobst 1978; Yellen 1977). When relying on ethnographic data, archaeologists would do well to rem ember that while ethnographies provide snapshots of hunter gatherer lifeways, archaeologists must deal with fragments of the entire family album (Jochim 1991:315). Archaeologists typically must work at different spatial and temporal scales than those whi ch are described in ethnographies. Additionally, we must be aware that hunter gatherer variability in the ancient past likely exceeded that observed for extant groups (Yellen 1977). However, the strength of using archaeology in studies of hunter gatherers is that it can highlight the variability of hunter gatherer behavior through time and across space (Jochim 1991; Wobst 1978).

PAGE 37

37 Hunter gatherers are not homogeneous, and their lives and movements may have varied from year to year, and even seasonally. Throu ghout the past century, many typologies of hunter gatherer groups have been made, and archaeologists commonly approach variability within and among hunter gatherer groups with respect to environmental variability, or they explain away incomplete data by ma king reference to a presumed seasonal round ( Barnard 1983; Binford 2001; Fitzhugh et al. 2011; Jochim 1991). While h unter gatherers are often described according to their economy or in terms of environmental constraints, a great deal of ethnographic literature indicates that their social structures are both diverse and flexible. Ethnographic studies of hunter gatherers indicate they often are well served by flexible social groupings, since circumstances often change. Flexible group membership allows people to leave the co resident group if social conflicts arise, or just because they choose to (Kelly 1995; Lee 1992; Marshall 1976; Owen 1965; Radcliffe Brown 1956; Woodburn 1968). A dditionally, a s participants in large scale social networks, hunter gatherers a re known to use mobility for social reasons beyond acquiring their basic subsistence needs (Whallon 2006). Within the context of the Early Side Notched Horizon, these concepts are fundamental in modeling the ways in which social boundaries are manifested i n material culture. Since early twentieth century ethnographic studies of Australian aboriginal groups (Radcliffe Brown 1930, 1931), there was a notion that hunter gatherers lived in certain kinds of patrilineal social bands. These co resident social grou pings were small, and were characterized by egalitarian social relations. B ands were thought to have practiced communal land ownership and patrilocal residence patterns (Steward 1936). This broad characterization of the band has

PAGE 38

38 remained within anthropolog y, and it also has affected the way we imagine ancient lifeways in the Southeast (after Kelly 1995). Griffin (1952) was the first to establish cultural periods/stages for the eastern United States. A seminal book chapter by Griffin (1952) demonstrates the influence of ethnography on archaeological interpretations (Anderson and Sassaman 1996:18) Griffin (1952:354355) suggested that Early Archaic groups most likely were patrilocal and patrilineal, and that cultural exchange of ideas occurred as a result of marital alliances outside local group boundaries. These notions largely have persisted in studies of Southeastern hunter gatherers. Twentieth century hunter gatherers are known to have complex histories and social interactions on great regional scales However, ethnographies of hunter gatherers largely have failed to describe the largescale social and spatial networks of these groups (Wobst 1978) (Grinker 1994 and Wiessner 1982 are notable exception s ). Simply asserting that maximal bands contained 500 people (Kelly 1994) does not explain the need for hunter gatherer aggregation at levels above the minimal band (Kelly 1995:209). Ethnography has demonstrated that hunter gatherers periodically aggregate for social and economic reasons. Aggregations may be seasonal, for the cooperative harvesting of particular resources, or annual or semi annual for ritual or social occasions (Conkey 1980; Damas 1969; 1972; Lee 1979). While the nature of such aggregations in the ancient past remains poorly understood, they have long been postulated as central features of Early Archaic lifeways (cf. Anderson and Hanson 1988; Griffin 1952). Ethnographic models of hunter gatherers provided thus far by anthropologists also cannot effectively account for hunter gatherer group int eractions above the level of the maximal band (as set forth in Steward 1969; Owen 1965; cf. Whallon 2006). In this sense, current

PAGE 39

39 anthropological models about hunter gatherers do not allow for the examination of larger interregional processes, such as thos e I argue we see archaeologically in the Early Archaic Southeast. Much of the extant ethnographic literature provides a worms eye view of reality about hunter gatherers, focusing on the dynamics of small groups in small units of space, who are character ized by little spatiotemporal variability (Wobst 1978:304). Assuming that hunter gatherer band societies are homogeneous in any way fails to take into account the remarkable flexibility of these social groupings. In reality, many of the ethnographically known groups of hunter gatherers actually are multilingual and diverse. This is in large part a result of the exogamy that must be practiced by members of small groups, who like most people throughout the world, recognize the incest taboo (Owen 1965) Accord ing to a great deal of ethnographic research, there is a tendency toward universal kin categorization among hunter gatherer groups, as they apply kinship terms to all people with whom rights and responsibilities are shared. For instance, there may be a division between possible spouses and prohibited spouses, and who can participate in a particular hunting group or trading relationship and who cannot (Barnard 1978, 2011:8081; Kelly 1995:281284). This tells us that evidence for solidarity above the level o f the band or the macroband is a result not of linguistic affiliation, but of a shared sense of ethnic identity, bound together by geography, and often by myth, as well ( after Owen 1965). Sharing and trading doubtless were always integral aspects of hunte r gatherer life, irrespective of spatial or temporal context, since they appear to be fundamental human practices. When asked, many hunter gatherers report that social interaction of all types is essential for a happy life (Kent 1993). However, we cannot s ay that sharing beyond the immediate family is an

PAGE 40

40 essential attribute of hunter gatherer life, since exceptions do exist (after Gould 1982). Among hunter gatherers, the very act of sharing is often valued as much or more than whatever commodity is shared, as sharing helps to maintain the social order, and acts as a form of social storage (Kelly 1995; Kent 1993). In contrast to longheld ideas about the miserable nature of hunter gatherer lifeways, some studies have found that they spend the majority of thei r time thinking about trade and gift giving (after Gould 1982). This demonstrates that there exists a social basis for sharing, and that it is not purely economic in nature (Fitzhugh et al. 2011; Kent 1993; Whallon 2006; Wiessner 1982, 1984). For instance, far from being free of politics San groups in southern Africa have been described as keeping peace where they live by visiting, sharing, and gift giving (Marshall 1976). Sharing food and other resources actually is a wise strategy, since it can provide t he sharer with access to land and other resources outside his or her own immediate landscape or territory. In fact, studies of hunter gatherers who face environmental variability or patchiness indicate that, in general, the more risk people face, the more sharing and social networking you are likely to see (Gould 1982; Whallon 2006). Much of the literature about hunter gatherers in the past decade has focused on the notion of social networking (Apicella et al. 2012; Fitzhugh et al. 2011; Henrich 2012; Lovis et al. 2006; Whallon 2006; Whallon et al. 2011). Taken together, this recent literature is a timely response to the paucity of anthropological models to describe interregional social processes. This research has been based on ethnographic studies (e.g., w ith the Hadza of Tanzania), as well as on archaeological examples (e.g., the Upper Paleolithic and Mesolithic of Germany, and the Kuril Islands in the North Pacific). The thrust of this research has been that large scale cooperation and

PAGE 41

41 social networking a re key features of hunter gatherer life, and that the establishment and maintenance of social relationships must have been a major influence on mobility patterns in the ancient past. The sharing of information and resources even beyond the level of the max imal band is indicated by much of this research. In the ethnographic literature, such large scale social networking has variously been classified as visiting or as aggregation (cf. Kelly 1995; Whallon 2006). The regular maintenance of regional social ti es would have been especially advantageous for groups in uncertain environments, with informationsharing taking place at multiple and overlapping scales of time and place (Fitzhugh et al. 2011; Whallon 2006). The specific strategies employed in informatio nsharing likely varied according to context in the ancient past, much as they are known to vary among modern hunter gatherer groups (Fitzhugh et al. 2011; Whallon 2006; Wiessner 1982). For instance, informal regional networks, characterized by frequent interactions, relatively frequent trade interactions, and walkabouts, would have been well supported by a relatively continuous social landscape with few physical or social barriers to interaction (Fitzhugh et a l. 2011:96). These informal regional networks would allow for the flow of multiple channels of i nformation, from local to macro regional scales. In archaeological cases of high overall environmental predictability and low costs to interaction social networking has been interpreted as being more for sociopolitical purposes than for the exchange of ecological information (Fitzhugh et al. 2011:9697). Long distance social networks are key to hunter gatherer life, since they allow freedom of movement and interaction, and often provide access to resources that otherwise would be unavailable (Gould 1982; Whallon 2006). Perhaps this is why the notion of universal kinship is so central to hunter gatherer life, in

PAGE 42

42 a variety of environments, including in Africa, Australia, Asia, and South America (Barnard 1978, 2011). The implications of this recent research on social networking for patterned variation in the Early Side Notched Horizon are discussed in Chapter 6. For the purposes of this chapter, I next consider the spatial aspect of hunter gatherer social netwo rks, which may best be contextualized in terms of place. Place Making Contemporary approaches to the modeling of social boundaries using patterned variation in material culture allow for the linkage of material culture to place, landscape, and identity. These linkages are key to examining the processes that shaped and transf ormed social boundaries during the Early Archaic period. The concept of place is especially relevant to an understanding of patterned variation in the Early Side Notched Horizon. As I demonstrate throughout Chapters 5 and 6, certain attributes of side notc hed tools are highly correlated to geographic region, providing support for the existence of subregional traditions within the Early Side Notched Horizon. Such place based traditions can arguably be traced to Paleoindian populations, the first placemakers in the American Southeast. The territorial ranges established during the Paleoindian period were oriented along resource rich river drainages, and this trend continued during the Early Holocene (after Anderson 1990, 1996). I would argue that the place ori ented patterning we see for the Early Side Notched Horizon was closely bound to personal and cultural identity (after Tilley 1994:15). Areas of habitual use as well as boundary zones and places of aggregation can thus be seen to reflect historical and mult iscalar social relationships. Following the work of Christopher Tilley (1994), I view landscape as a medium for social activities and identities, rather than a container in which activities and events occurred,

PAGE 43

43 and a surface over which people moved. This view of landscape allows it to be interpreted as a social production that cannot be divorced from human activities and interrelationships. A phenomenology of landscape, as advanced by Tilley (1994), relates to the experiences of people with landscapes. Thes e experiences are deeply connected to temporality, as spaces are always created, reproduced, and transformed in relation to previously constructed spaces provided and established from the past (Tilley 1994:11). In this sense, the re use of places inhabit ed by Paleoindian and more recent ancestors may well have provoked historical memory among Early Archaic period groups ( after Bender 2002). Much of the foundational literature about archaeological landscapes is focused on monuments and other structures, ra ther than archaeological traces such as those examined for this dissertation (Bender 1993; Thomas 1993; Knapp and Ashmore 1999; Tilley 1994). For instance, Julian Thomas (1993:77) has used the notion of the dominant locale to interpret megalithic (Europe an Neolithic) tombs in the landscape as places where people carried out their daily routines and activities. Perhaps this notion can also be applied to places in the Early Archaic landscape, including chert quarries, major river drainages, and interstream divides. While these places were not architectural constructions, they were arguably socially constituted and constitutive places during the Early Archaic that may have served as meaningful locations for human activities and interactions (after Tilley 1994 :17). Recent understandings of unaltered landscapes (Bradley 2000), which are informed by ethnographic research, allow for consideration of natural landscapes in historical context. Ethnographies of certain hunter gatherer societies indicate that the natur al landscape becomes humanized through the use of place names and detailed associations with current use, historic use, or mythical or totemic

PAGE 44

44 activities and events (after Basso 1983; Gould 1969, 1971, 1978; Tilley 1994; Woodburn 1968). In this way, space and time can be seen as bound together by social practices (Tilley 1994). As this dissertation is focused on the multiscalar relationship between artifact variation and cultural affiliation, it is important to note here that places can be experienced and i nterpreted at multiple spatial levels. This concept is closely related to that of social networks, as presented above. Levels of ex perience and interpretation include (but are not limited to) personal space, community space, and regional space, and places at these levels may well be seen to overlap. As people in small scale non Western societies have a sense of being a part of the places where they live out their lives, place is fundamentally connected to the formation of both personal and group identities (Tilley 1994). At the local level, the dwelling perspective can be applied. At a greater regional scale, we can interpret similarities in landscape use as a product of both history and interaction. In this sense, we can read identity as a balance between l ocal and regional scales. In tacking back and forth between the local and the regional perspectives, we must at once consider local places, as well as potential social boundaries in the broader landscape (after Nassaney and Sassaman 1995). Those boundarie s may well represent places of interaction, which may have archaeological expressions ( Barth 1969; Parker 2006; Stark 1998; Wiessner 1982). Additionally, a regional perspective to ancient landscape use may allow for the identification of points of aggregat ion. Below, I review specific ethnographic examples of the linkages between place, landscape, and identity that may be analogous to this study of patterned variation in the material culture of the Early Side Notched Horizon. Places in the landscape may ser ve as metonyms for past events, ancestors, mythical beings, or as reifications of what is done in daily practice (Fullagar and Head 1999; Layton

PAGE 45

45 1999). Certain places also may serve as boundary m arkers. Ethnographically we see that no hunter gatherer group has no boundaries (Kelly 1995:185). Although they are differently maintained and perceived depending on the group in question, boundaries are known to exist within all ethnographically known groups (Kelly 1995:181189). T his knowledge fits well with recen t notions that native people have complex relationships with their landscapes, and that their landscape shapes and is shaped by their social interactions and other activities (Basso 1983; Morphy 1995; Smith 1999). As recent ethnoarchaeological research has demonstrated, hunter gatherer territories are much more than sources for food and firewood; they are the places of ancestors as well as descendants, and they represent and shape social relationships on a daily basis (Hitchcock and Bartram 1998). For example, Australian tribal groups have personal bonds with their home territories, the boundaries of which tend to be marked by natural features that may not be obvious to outsiders. Obvious natural boundaries include rivers, streams, and watersheds, while the less obvious include changes in rock type, slope, or soil type (Williams 1982). In this way, the Aboriginal landscape can be said to be socialized, replete wi th boundaries related to ancestry and mythology (Tilley 1994). Objects of material culture also can be used to mark social boundaries and to share information, essentially referencing other people and places. H xaro exchange of arrows and other objects by San groups in southern Africa is a classic example of this practice (Wiessner 1982). Ethnoarchaeolo gical research (Wiessner 1982, 1984) with the San (or the Kalahari San or !Kung San, as they are often called) is significant for its investigation of issues of artifact variation and social interrelationships in great detail, at large social and spatial s cales. For instance, metal arrows are important in San myth and folklore, and the forms often are

PAGE 46

46 exchanged between people belonging to different bands wit h whom risks are pooled. These people often are recognized as hxaro partners where hxaro is a kind of delayed gift giving of nonconsumable goods. The hxaro system relates individuals with up to forty other people. These people are known as trading partners, and hxaro has created a chain of indebted people that in some cases spans hundreds of miles (Wiessner 1982). Hxaro also is notable as it seems to promote environmental equilibrium, giving hxaro partners the right to procure consumable goods (including water and game) on lands held by the hxaro counterpart (Barnard 2011; Wiessner 1977). According to San informants, hxaro exchanges of arrows allow for greater visibility of weapons which normally are kept (or, hidden) in quivers. While San artisans cannot recognize the arrows made by their own hands, they are able to distinguish between arrows made in their language communities and those made in other communities (Wiessner 1982) This indicates that San arrow types do serve to demarc ate boundaries to some degree. However, it cannot be said that differences in arrows are produced in order to mark soci al boundaries ( David and Kramer 2001; Wiessner 1982). The example of hxaro clearly indicates that hunter gatherers have relationships that are founded on much more than economic need. The arrows, ostrich eggshell beads, and other objects that are traded as part of hxaro are not economic necessities. Rather, they are used to create and maintain social bonds, and this seems to be one of the primary focal points of life in the Kalahari (Wiessner 1977). While this phenomenon should not be surprising to anthropologists, hunter gatherers often are denied any kind of large scale social interaction, and normally are described at the band level of analysis. The example of hxaro is a good reminder to archaeologists, not to deny the social and macroregional aspects of hunter gatherer

PAGE 47

47 lifeways, and the highly materialized nature of hxaro information sharing has inspired archaeologists to develop multiscalar models of hunter gatherer social networks (Fitzhugh et al. 2011:88; Whallon 2006:261; Whallon et al. 2011). Summar y The primary danger in the application of ethnographically derived theories of culture to more ancient contexts is that archaeologists simply may reproduce the form and structure of the perceived realit y of ethnography (Wobst 1978). If current anthropolog ical models are to be taken at face value, archaeologists might think that regional and inter regional processes among hunter gatherers are due to cultural degeneration, and to cont act with capitalist societies. This view suggests that hunter gatherers have no politics of their own, and that they are socially and spatially circumscribe d (Sassaman 1998; Wobst 1978). Alternatively, we can see that multiple social identities exist within and among hun ter gatherer groups, and that an archaeological study of hunt er gatherers stands to make significant contributions to anthropological studies of regional and interregional social interaction and reproduction. In this chapter I provided historical background for hunter gatherer research, relating conventional wisdom about social structure and largescale social and informationsharing networks. Also, I reviewed the linkages of place, landscape, and identity, as these concepts help to historicize the patterned variation evident for the Early Archaic in the lower Southeast (after Sassaman 2010). I suggest that the Early Side Notched Horizon represents a continuation of placebased traditions initiated during Paleoindian times, an argument I return to in Chapter 3 in discussion of culture history for this region. S pecifically, I discussed the concepts of unaltered places and social landscapes, both of which are greatly informed by ethnographic research.

PAGE 48

48 These topics all relate to arguments presented in Chapters 5 and 6 with respect to the Early SideNotched Horizon.

PAGE 49

49 CHAPTER 3 GEOGRAPHY AND CULTURE HISTORY In contextualizing the Early Side Notched Horizon, first it is necessary to review the current state of knowledge about the environmental setting of the Late Pleistocene and Early Holocene epochs. Early Archaic inhabitants of the Southeastern United States did not arrive in the region sui generis Rather, these people arguably were descendants of Paleoindian (and, possibly Pre Clovis) groups who were evidently in place in the Southeast by ca. 13,000 cal B.P. (Waters and Stafford 2007:1122) Immediately prior to the Early Archaic period, t he landscape of the lower Southeast was home to groups who had established areas of habitual use as well as variations on a shared technology by late Paleoindian tim es, during the Terminal Pleistocene. As such I discuss Paleoindian culture history below, with special attention to generalized models of initial colonization and the emergence of subregional cultural traditions. This chapter also provides an overview of the geographical and cultural contexts for the Early Archaic period in the lower Southeast. I suggest that the subregional traditions that apparently initiated during the Paleoindian period continued throughout the earliest portion of the Early Archaic. Th erefore, comp arisons between the two culture historical periods are appropriate in modeling the social boundaries manifested in the Early Side Notched Horizon. I relate how the Taylor, Bolen, Big Sandy, and Edgefield Scraper flaked stone artifact types wer e defined and provide a broad overview of several Early Archaic period sites that have aided in the building of chronology for the Early Side Notched Horizon. Also, I review specific settlement models that have been applied to the latter portion of the Early Archaic, when sidenotching ceased as a cultural tradition in the lower Southeast. In concluding this discussion, I propose a

PAGE 50

50 tentative model for Early Archaic social boundaries by projecting the staging area hypothesis (Anderson 1990, 1996) from Pal eoindian times into the Early Holocene. A generalized survey of highquality Coastal Plain chert raw material sources throughout the study region also is presented here. Many of these sources were demonstrably used for the manufacture of side notched tools (Austin and Mitchell 1999, 2010; Goad 1979; Goodyear et al. 1983; Goodyear et al. 1985; Upchurch et al. 1982). The PleistoceneHolocene Transition The PleistoceneHolocene transition was a time of dramatic environmental changes, which are recognized from recent paleoenvironmental data throughout the lower Southeast. Detailed interdisciplinary studies of the paleoenvironment in this region have been undertaken in central and northern South Carolina and in Florida. These zones comprise the northern and sout hern boundaries of my study area, and current data suggest that these regions differed significantly with respect to climate and available resources during the onset of the Holocene. The implications of this regional climate variability on cultural express ions are discussed below. The Late Pleistocene Environmental Setting and Culture History The terminal Pleistocene was remarkably cool and dry, and while most water was trapped in ice across much of the continent, the project study area itself was not glaci ated. As a result of glaciation elsewhere in North America, the water table was much lower than it is today. Most early Southeastern sites are located near permanent sources of fresh water, which often was perched on the limestones of the Floridan Aquifer, and which was available in lakes, prairies, or springs (Dunbar 2002; Dunbar and Waller 1983).

PAGE 51

51 The first large scale influx of inhabitants to North America occurred during t he Late Pleistocene, and hunter gatherers quickly settled into the Southeast during that time. Southeastern populations at the end of the Paleoindian period witnessed the Younger Drya s which was the final major c old advance of the Pleistocene Ep och, occurring ca. 12,85011,650 cal B.P. (Anderson and Sassaman 2012:38). This sudden cool period, followed by rapid warming, doubtless affected later Paleoindian populations (Morse et al. 1996). For example mobility patterns and seasonal movements may well have been impacted by these extreme env ironmental shifts. Apparently, L ate Paleoindian groups (e.g., those engaged in the Dalton cultural tradition, discussed below) were forced to alter their subsistence strategies as numerous genera of megafauna became extinct, and hardwoods replaced boreal forests (Delcourt and Delcourt 1985; Delcourt and Delcourt 1987; Martin 1984). A recent hypothesis (Firestone et al. 2007) holds that a comet may have caused the rapid (perhaps, within a few decades) cooling of the Younger Dryas; this remains a controversial hypothesis (Anderson and Sassaman 2012:58). Pr ior to discussing Early Archaic culturehistory, some statements must be made regarding the antecedents of side notched tool forms in the Southeast, namely, Clovis, the primary type artifact of the Paleoindian period. Clovis refers to a type of fluted lanc eolate biface that was used in the hunting of megafauna (including, but not limited to, mastodon) (cf. Clausen et al. 1979; Webb et al. 1984); these tools often were manufactured from exotic and highquality raw materials. Clovis has also come to refer to the earliest Paleoamerican populations, who are presumed to have been ancestral to participants in the Early SideNotched Horizon. Traditionally, Clovis has been classified as an adaptation for the hunting of large mammals (Dunbar and Webb 1996; Hemmings 2004; Kelly and Todd 1988). Refined chronologies place

PAGE 52

52 t he Clovis tradition ca. 13,25012,800 cal B.P. (Waters and Stafford 2007:1123). As such, the Clovis tradition (as defined by Waters and Stafford 2007) is considered Early Paleoindian, and is closely as sociated with the final period of existence f or numerous species of megamammals. Establishing when the first migrations to North America occurred is outside the scope of this dissertation, and consensus is lacking about the origins and chronology of the fi rst Americans (Anderson and Sassaman 2012; Beck and Jones 2012; Fiedel and Morrow 2012). For instance, while some archaeologists continue to assert that an ice free corridor (Beringia) was the m ost likely route, bringing east A sian peoples in from the nort hwest portion of North America, others suggest that groups may have arrived by watercr aft from South America or even W estern Europe (Fiedel 1999; Stanford and Bradley 2002). Regardless of the exact timing and origin of the initial peopling of North America, the settlement models discussed below have great relevance for the current study. One model for Paleoindian coloniza tion of North America (generally known as the technology oriented model of Clovis radiation through the continent) suggests the process was rapid, and that mobility was tied to the absence of extant populations with local knowledge of resources, as well as a tendency to change territories as local resources became scarce (Kelly and Todd 1988). This model (as it was offered prior to mounting evidence at a number of sites in North and South America) presumes that no Pre Clovis occupation of the continent ha d occurred, and suggests that Paleoindians made frequent territorial moves in pursuit of terrestrial game, many species of which were rapidly becoming extinct due to environmental stresses of their own. The technology oriented model (Kelly and Todd 1988) predicts short term Paleoindian sites across North America, apparently with no immediately successive occupations.

PAGE 53

53 This model suggests that Paleoindians were technology oriented, rather than placeoriented. According to this model, Paleoindians used portabl e technology (bifaces) made of highquality raw materials, since in unknown terrains, they primarily were interested in the exploitation of fauna. It is important to note that the technology oriented model was offered as an alternative to the overkill model ( Martin 1973, 1984), which suggested that Paleoindians essentially caused megamammal extinction by over hunting. Another model regarding the peopling of North America the staging area hypothesis, suggests that Clovis settlement was gradual, and that g roups slowed their initial migration to establish territorial ranges based along resourcerich river drainages. According to this model, Clovis populations were place oriented rather than technology oriented, and these staging areas were areas of habit ual use that formed the foundations for early regionalization in the Southeast (Anderson 1990, 1996; Anderson and Sassaman 2012: 50). According to the staging area hypothesis, interaction networks enabled Paleoindians to colonize North America despite low p opulation density, while maintaining reproductive viability. These social networks may have been maintained by way of loosely scheduled meetings (Anderson and Gillam 2001:530) or occasional instances of social aggregation. Such networks arguably also all owed for the development of subregional traditions ( Anderson 1995; Anderson and Gillam 2001), such as those evidenced for the subsequent Early Side Notched Horizon. Patterns of regionalization predicted by the staging area hypothesis have been supported by recent studies of fluted points (that apparently post date Clovis forms) from throughout the Southeast (Smallwood 2012; Thulman 2006). One study of bifacial forms f rom sites in Tennessee, South Carolina, and Virginia lends support to the specific aspects of timing and

PAGE 54

54 directionality predicted by the staging area model (Smallwood 2012). Another study considered Paleoindian bifaces from numerous Florida river drainages Based on the degree of similarity between these regions, a distancedecay relationship was considered, such that the collective styles of closely related groups would be expected to be subtle and to vary in few aspects. This study identified the earlie st evidence for regionalization in Florida points at the end of the Early Paleoindian period. This regional variation appears to have increased over time, as Florida Paleoindians apparently intensified their use of specific river drainage locations (Thulma n 2006). A dditional recent research has demonstrated that possible Paleoindian aggregation locations tend to be located at the intersections of major rivers, at raw material sources, and other physiographic boundaries (Miller 2011). Such placebased tradit ions arguably laid the foundations for early regionalization in the Southeast. This GIS based research also appears to support the staging area hypothesis (Anderson and Sassaman 2012:50 51). In Chapter 6, I evaluate the strength of the staging area model w ith respect to the origin and persistence of subregional cultural traditions in the Early Side Notched Horizon. Researchers have been unable to explain the genesis of apparent subregional variation in later Paleoindian bifacial forms (e.g., Cumberland, Fol som, Quad, Redstone, Simpson, Suwannee, and Waisted Clovis ) (cf. Smallwood 2012; Thulman 2006, 2007). Cultural transmission theory (as advanced in Bettinger and Eeerkens 1999) may be an avenue for special consideration (after Smallwood 2012). According to that hypothesis, in some cases, point manufacturers may have copied the designs made by the groups most successful hunter, such that all designs may be seen to converge on that single design. In other cases, point manufacturers may have acquired their ini tial design from another individual, and then modified

PAGE 55

55 different attributes on a trialand error basis, resulting in the degeneration of correlating attributes across the region. This use of cultural transmission theory was applied to explain regional diff erences in the transition from use of the atl atl to the bow and arrow in the Great Basin (Bettinger and Eerkens 1999). Most scholars agree that Clovis groups were ancestral to later groups, who continued to fill in the landscape during the Late Paleoindi an and Early Archaic periods (Anderson and Sassaman 2012; Anderson et al. 1996; Bradley 1997; Sassaman 2010). In the Southeast, these later groups often manufactured hafted bifaces known as Dalton, followed by the side notched tools which are the focus of this study. The Dalton tradition dates to the end of the Pleistocene, during the Late Paleoindian period (Anderson et al. 1996; Anderson and Sassaman 2012:5). The tradition is named for the distinctive hafted biface (the Dalton point) made during that time. Dalton points, in turn, were named after an avocational archaeologist who recovered numerous examples in the first part of the twentieth century (Chapman 1948). Daltons are basally thinned, lanceolate, and often are fluted. There is every indication to vi ew Daltons as part of a descendent technological tradition that began with Clovis (Bradley 1997; Goodyear 1982). The earliest known cemeteries were established i n North America as part of the Dalton cultural tradition, and at least one of these (the Sloan site in Arkansas) was associated with numerous examples of Dalton hypertrophy (e.g., some of the hafted bifaces were as long as 18 cm) ( Morse 1997 a :17). Microwea r analyses indicate those large tools never had been used, or else had been resharpened prior to their apparent ceremonial placement with human interments (Yerkes and Gaertner 1997). This sort of hypertrophy has not been documented for other tool types unt il at least the Middle Archaic (Anderson and Sassaman 2012:61)

PAGE 56

56 The dating of many Paleoindian artifact forms is difficult, and there is no exception with the Dalton form. While some have been able to obtain absolute dates for Dalton components (e .g., Good year 1982), others recognize that relating the tool type to its stratigraphic position, as well as to its position relative to other tool types in the vicinity, is essential (Sherwood et al. 2004). Important Dalton sites in the Southeast include the Stanfi eld Worley site in Alabama (DeJarnette et al. 1962) and the Brand and Sloan sites in Arkansas (Goodyear 1982, 1995 [1974] ; Morse 1997b). Data from many sites throughout the Southeast indicate that Dalton points were not coeval with Early Archaic sidenotc hed forms, and that the transition to side notching was relatively rapid, occurrin g between 10,500 and 10,200 years ago. Some Daltons appear to be almost side notched, which suggests they may have been a transitional form between Paleoindian fluted lanceol ates and Early Archaic sidenotched tools (Morse 1994). However, most Dalton points are readily distinguished from later side notched tools. In contrast to earlier fluted points, Dalton points generally have a shorter haft element, which would have provide d a greater cutting edge. This edge would have been ideal for butchering deer, which is the proposed function of Daltons (Goodyear 1995 [1974]). Later side notched points have even smaller haft areas, which some suggest reflects dramatic changes in the toolkit, including perhaps the advent or improvement of the atl atl (Morse 1994). The Early Holocene Environmental Setting and Culture History The posit ion of the Early Archaic at the PleistoceneHolocene transition may help to account for the numerous shifts in subsistence, settlement, and technology observable in the archaeological record throughout North America. As such, it is necessary to review the evidence

PAGE 57

57 for these environmental conditions, which provided a backdrop for a diversity of lifeways throughout the lower Southeast. This evidence comes from a few excavated sites, as well as from pollen diagrams and geomorphological studies. Early Holocene environmental setting The Early Holocene bega n after the Younger Dryas which as discussed above, marks the end of the Late Pleistocene. Seasonal variation was greater than it is today, with warmer summer temperatures and colder winter temperatures. At t he end of the Pleistocene, temperatures increased, and sea levels rose as a result of glacial retreat. Modern forms of fauna replaced megafauna, after dozens of the large mammal species went extinct. These extinctions are thought to have been completed by about 10,000 years ago (Anderson et al 1996:3; Grayson 1987; Meltzer and Mead 1983; Morse et al. 1996). At the onset of the Holocene, shorelines were approximately 60 meters below modern sea levels, and the Florida peninsula was approximately double its c urrent size (Faught 1996). Terrestrial and underwater archaeologists are only beginning to document the distribution of Early Archaic (and more ancient) artifacts along the continental shelf, south and west of what is the current Gulf Coast of Florida. For instance, investigations of a submerged channel of the Aucilla River (Faught 2004a, 2004b), and along the continental shelf off the west coast of Florida (Adovasio and Hemmings 2011) demonstrate that sealevel rise attendant with postglacial warming inund ated large sections of the lower Southeast. These events no doubt influenced settlement during the Pleistocene Holocene transition, and settlement and mobility models that fail to consider these environmental constraints remain incomplete (Anderson and Sas saman 2012:42).

PAGE 58

58 There is no evidence for the domestication of any plants or animals during the Early Holocene. One exception is the dog, which apparently was introduced to eastern North America by Paleoindian groups. These first dogs in North America are t hought to have come from Northeast Asia, and presumably arrived to the continent with the initial colonists (Swartz 1997). Hardwoods (including Quercus and Carya ) covered most of the lower Southeast during the Younger Dryas (Morse et al. 1996). North of 33 latitude oak, hickory, elm, and birch replaced the pine boreal forests by ca. 9,000 years ago, and south of that point (with the exception of peninsular Florida), stable oak hickory forests already were established (Anderson et al. 1996; Delcourt a nd De lcourt 1987; Watts 1980 a, b). North of the Savannah River, mixed hardwoods likely dominated the landscape, while at the Savannah River and to the south, forests likely were a combination of mixed hardwoods and oak and hickory trees. Environmental variatio n was greatest within the Savannah River drainage, as well as to the south of it (Daniel 1998; Steponaitis 1986; Ward 1983). The first portion of the period saw an increase in rainfall, and even some flooding in parts of the study area (Faught and Carter 1998). Pollen diagrams from northern and central South Carolina indicate that Pinus and Quercus dominated the landscape after ca. 9,500 years ago, and this pattern continued throughout the Holocene ( Hussey 1993; Watts 1980a ). Therefore, cool mesic hardwood forests appear to have been the norm for the Early Archaic, especially to the south of the 37th parallel and, in many ways, the environment was much like it is today (Delcourt and Delcourt 1985; Hussey 1993; Morse et al. 1996; Watts 1980a ). However, the description of forests as mesic should perhaps best be used with caution, as more palynological data from the region still are needed (Watts 1980 a, b).

PAGE 59

59 Efforts to model subsistence patterns for the Early Archaic have been challenging, and are based on relatively few directed studies. The lack of organic preservation that characterizes the region has been a particular problem for researchers; this ideally will change with the advent and use of refined techniques for data col lection. Although the Windover archaeological site in Brevard County, Florida apparently dates to the latter portion of the Early Archaic, materials recovered there were remarkably well preserved, and presumed evidence of subsistence in particular has been extrapolated for the entire Early Archaic period (Austin and Mitchell 1999). The Windover materials were identified in part using stable isotope analyses, and a wide variety of flora and fauna exploitation is indicated by these data (Doran and Dickel 1988; Turross et al. 1994). Plan ts used by Early Archaic hunter gatherers are presumed to have been similar to those used by their Paleoindian ancestors. Hickory nuts in particular seem to have figured prominently into the Early Archaic diet (Detwiler 2000; Driskell 1996). Edible flora indicated by Windover data include acorn, hickory, persimmon, prickly pear, maypop, wild plum, and wild grape. Fauna indicated by these data include deer, opossum, rabbit, raccoon, and squirrel (Doran and Dickel 1988). White ta iled deer in particular are thought to have comprised a large part of the Early Archaic diet, and were found in much of the region during the Early Holocene (Anderson et al. 1996; Morse et al. 1996). Lithic blood residue analyses of samples from one Florida site indicate that some side notched tools were used to kill or to process rabbit and bear (Hornum et al. 1996). Evidence of bird species including duck and heron also was recovered in deposits at Windover (Doran and Dickel 1988; Turross et al. 1994), and bird remains also have been identified for the PleistoceneHolocene transition at Dust Cave (Walker 2007).

PAGE 60

60 Fish also apparently formed an important part of the Early Holocene diet; apparent evidence for the exploitation of numerous fish species (includin g large mouth bass and catfish) was recovered at the Windover site (Doran and Dickel 1988). Additionally fish hooks have been recovered at the Dust Cave site in northern Alabama (Driskell 1996). Pieces of turtle shell also were recovered at Dust Cave, as well as at the Windover site, suggesting that turtles comprised part of the Early Archaic diet (after Doran and Dickel 1988; Driskell 1996). Possible evidence for the exploitation of alligators, certain frogs, and snakes, also was recovered at the Windover site (Doran and Dickel 1988). Isotopic evidence from burials at the Windover site indicates the presence of aquatic resources (both freshwater and riverine) in the Early Archaic diet (Tuross et al. 1994). Taken together, these data indicate that a wide ra nge of fauna and flora was exploited by Early Holocene groups throughout the Southeast, and a greater range of diet may have been related to the recent extinction of megamammals. This general overview of the breadth of subsistence patterns demonstrates tha t while some degree of environmental uncertainty may have existed for participants in the Early Side Notched Horizon, their responses were highly diverse (after Anderson and Sassaman 2012:3839). This apparent flexibility with respect to diet may also be expected with regard to social groupings, as anticipated in Chapter 2 and as further explored in Chapters 5 and 6. Periods of rapid karst drainage at the onset of the Holocene are indicated in the hydrology of the Page Ladson site, which is located in a sinkhole of the Aucilla River, and these periods seem to be the result of La Nina drought conditions. The water table dropped at around the same time at numerous other Florida sites, including Little Salt Springs in southwest Florida, resulting in a water tab le some seven meters below its previous stand during the late Pleistocene. These

PAGE 61

61 trends suggest a dramatic hydrologic response to more arid conditions throughout Florida by ca. 11,500 cal B.P. One researcher has dubbed these collective events part of the Bolen Drought (Dunbar 2002:143). The duration of the drought has been date d to three centuries, based on radiometric dates obtained from Unit 6L at the Page Ladson site (Dunbar 2002:148). Some archaeologists have suggested that Floridas water table did n ot approach modern levels until approximately 8,500 years ago (Dunbar 1991; 2002; Gunn 1996). A centuries long drought would result in desert like conditions, as well as the likely extinction of numerous floral and faunal communities. It also would result in a settlement shift for human groups, who likely would need to make strategic settlement decisions, or even significantly alter their settlement strategies. Prior to the stabilization of the water table, it would have been necessary for native Floridians to get water from aquifer fed springs, or from lakes perched on limestone. As many researchers have noted, the distribution of Early Archaic sites in Florida demonstrates that native groups elected to live close to dependable sources of fresh water. Elsew here in the Southeast, groups made more use of other environments, including riverine and coastal settings (Austin and Mitchell 1999; Goodyear et al. 1983). More data from submerged contexts (e.g., along Floridas Gulf Coast) obviously are needed to determ ine if estuarine resources were as valued in Florida as they were elsewhere in the Coastal Plain (cf. Anderson and Sassaman 2012:4243). The large groundwater systems held in Floridas Tertiary karst regions, couple d with the availability of high quality chert raw materials in those same regions, also seem to have influenced Paleoindian settlement, as Paleoindian sites tend to be based in drought tolerant Tertiary karst regions (Dunbar 1991). In my own research, I also noted that large assemblages of

PAGE 62

62 side notched tools in the state were collected from these Tertiary karst regions. Certainly side notched tools also have been recovered in other portions of Florida, but not in such high frequencies (cf. Sassaman 2003:90) This suggests that Early Archaic group s chose to settle in the same environments as did their ancestors, and quite likely for the s ame reasons; desires for a high quality chert, and a consistent source of water (after Goodyear et al. 1983). Currently, it is not known if other portions of the S outheast experienced such a dramatic trend toward aridity at the onset of the Holocene. Until these data are forthcoming, it seems that settlement patterns at Floridas Early Holocene sites may be markedly different from sites in other portions of the lowe r Southeast, where higher elevations were present. Also, water sources may have been more predictable in regions outside of Florida. For instance, features known as Carolina Bays may have been excellent sources of water, and some sites have produced evidence of Early Archaic occupation at these locations ( Brooks et al. 2010; Sassaman 1996). Much more data are needed from throughout the Southeast, to add to existing models and interpretations of ancient landscapes. Furthermore, some criticisms have been made about an over reliance on highly general environmental models developed in the 1980s ( Delcourt and Delcourt 1985; Delcourt and Delcourt 1987). The thrust of such critiques is that more attention should be paid to environments people occupied at localized, rather than regional, scales. Further, it has been stated, archaeologists must be wary of modeling ancient human behaviors, based solely on the attributes of their environments (Dincauze 1996; Sassaman and Holly 2011). Early Side Notched Horizon artifact types The Early SideNotched Horizon of the American Southeast is broadly used by archaeologists to reference the widespread distribution of side notched hafted bifaces across the

PAGE 63

63 region from approximately 10,300 to 9600 cal B.P. (Driskell 1996:325326; Faught et al. 2003:17; Morse 1996). The side notched hafted biface is the generalized timemarker for the Early Side Notched Horizon. Its chronological significance at the beginning of the Early Holocene has been well establis hed through stratigraphically controlled excavations throughout the Southeast, and particularly in South Carolina, Florida, and Alabama (Austin and Mitchell 1999; Carter 2003; Carter and Dunbar 2006; Driskell 1996; Goodyear et al. 2007; Sherwood et al. 2004). This tool class is presumed to have performed a number of functions, including use as a projectile, and in a variety of cutting tasks (after Austin and Mitchell 2010). My use of the term hafted biface is to avoid implied reference to specific shape o r function, as terms such as projectile point, projectile point/knife, and even point tend to do (after Thulman 2006). Hafts for side notched bifaces may have been made of antler, bone, wood, or a material such as rivercane (Scott Jones and Kenn Stef fy, personal communication 20032005). Direct evidence of Early Archaic hafts is not available for the Coastal Plain Southeast. Notching would have enhanced the binding of bifaces into hafts. When a biface is notched such that the width between the notches equals the width of the shaft, a binding such as sinew can be tightly secured to the foreshaft or shaft using a natural adhesive (such as pine or spruce tree sap), allowing for proper alignment of the biface and shaft (Scott Jones, personal communication 2003; Van Buren 1974). Side notching likely facilitated the hafting of thicker biface bases, and would have reduced the risk of tool breakage as a result of impact fractures. The innovation of side notching in the Early Archaic Southeast arguably accompani ed a radical shift in hunting techniques (Anderson and Sassaman 2012:72; cf. Van Buren 1974).

PAGE 64

64 Side notching of hafted bifaces is thought to have begun at the end of the Paleoindian period, with the introduction of the atl atl, and a shift in hunting adapta tions to smaller game (Anderson 1990; Gardner 1974). Precisely dating the advent of side notching in the Southeast is quite difficult, as archaeologists are dependent on varying contexts and stratigraphic associations, as well as a paucity of radiocarbon d ates at many ancient sites. Additionally, they face challenges related to site location. For example, dating submerged or deeply stratified deposits can yield mixed results. There is some disagreement over what constitutes true side notching. One reason for the lack of consensus in whether a tool is side notched is the somewhat subjective nature of such nominal categories. By definition, side notching is literally when notches are removed from the lateral edges of a biface, near its base, through the use of pressure flaking. Use of this definition requires many tool types to be considered side notched, which often are seen more as transitional forms, and which likely date to the Late Paleoindian period. These transitional forms include: Greenbriar, Osceola G reenbriar, Union, and Hardaway (Bullen 1975; Coe 1964; Daniel 1998; DeJarnette et al. 1962; Lewis and Kneberg 1960; Powell 1990; Schroder 2002). Artifact type names have been proposed (Bullen 1958, 1975; Kneberg 1956; Michie 1966) to denote observed subre gional variations in tool forms within the Early Side Notched Horizon. While the use of distinct type names (including Taylor, Bolen, and Big Sandy) for similar side notched artifacts has been meaningful at localized archaeological scales of practice, the differences and relationships among these types has been less clear. How did these apparent subregional differences within the Early Side Notched Horizon come to exist, and how and why were such like forms reproduced over time and space?

PAGE 65

65 The variation with in and between side notched hafted biface types has yet to be demonstrated; this is the proximate goal of this dissertation. The geographic zones for these hafted biface types also have not been effectively demonstrated to date, although a review of the ex tant literature would suggest that Taylor, Bolen, and Big Sandy side notched forms all have relatively well mapped and understood geographical zones (cf. Bissett 2003; Bullen 1975; Cambron and Hulse 1983; Michie 1996; Milanich 1994). As a result, the broader regional landscape of the Early SideNotched Horizon remains poorly understood. Largely state oriented research strategies have led to distinct names and descriptions (e.g., Taylor, Bolen, and Big Sandy) for very similar tool forms. For example, if a si de notched hafted biface is recovered in northcentral Florida it is called Bolen, and if it is recovered in southwestern South Carolina it is called Taylor. If it is recovered in southwestern Georgia or Alabama, it commonly is called Big Sandy. While stat e boundaries are not ideal units for considering the expansive landscapes that arguably comprised the world of Earl y Archaic hunter gatherers (after Wright 1996), it is true that variation in these tools is such that typical South Carolina examples can be distinguished from those found in Florida. What is much less clear, and less frequently explored, is why certain similarities and differences in side notched hafted bifaces existed and persisted in the first place. Descriptions of the Taylor, Bolen, and Big Sandy artifact types are provided below. Taylor. Taylor Side Notched hafted bif aces were defined in 1966 by James Michie of the Archaeological Society of South Carolina (and, later, of the South Carolina Institute of Archaeology and Anthropology). Tayl ors were named after the Taylor site (38LX1) in central South Carolina, which Michie (1966, 1996) excavated. According to Michie (1966), the basal edges of Taylors are always square, and the bases have a slight concavity. The blade area is

PAGE 66

66 alternately beveled and serrated, and the lateral edg es, basal concavities, and sidenotching areas are ground smooth. According to Michie (1966), chert and orthoquartzite were the most commonly used raw materials for Taylor Side Notched manufacture. Michie (1966) described the distribution of Taylor Side Notched as rather limited, from the fall line south of Columbia, South Carolina and in the Santee River drainage. He suggested that the highest frequencies of Taylor Side Notched could be found in the Santee River area, a nd that the points fell somewhere between the Dalton and Palmer Corner N otched horizons. Taylor P oints never have been well dated radiometrically, but historically have been compa red to Bolen Points in Florida (Michie 1996; after Dunbar et al. 1989). This would place the artifact type in the earliest portion of the Early Archaic. In the last four decades, much more has been learned about Taylor Side Notched distribution. Now we know that the form occurs state wide in South Carolina, as well as throughout South Georgia and into northern Florida, and that it is coeval with Bolen Side Notched hafted bifaces (Austin and Mitchell 1999; Michie 1996). This makes it difficult to determine when it is appropriate to call a side notched tool Taylor, and when to call it Bolen. We also know now that a variety of raw materials were used to manufacture Taylor Side Notched hafted bifaces, although it is true that Coastal Plain chert from sources near the lower Savannah River was used most often. Unfortunately, the variation that exists within the Taylor Side Notched type remains poorly understood. Bolen. Bolen Side Notched hafted bifaces were defined by Ripley Bullen (1958, 1975) of the Florida Park Service (which at the time operated out of the Florida State Museum, at the U niversity of Florida), following his excavations at the Bolen Bluff site (8AL439) in Alachua

PAGE 67

67 County, Florida. Bullen (1975) described two types of Bolens, Bolen Plain and Bolen Beveled, each of which he divided into five subtypes. Bolen Plain is basically the same as Bolen Beveled, according to this distinction, except that the latter is beveled. Bullens (1975) subtypes include the following: concave based, side notched, high notched, corner notched, and expandednotched. This description states that the w idth of the base is the same as the width of the blade in the first three subtypes. Meanwhile, the width of the base is narrower than that of the blade in corner notched examples, and in expandednotched examples it may be the same or narrower. It has sinc e been suggested (Sc hroder 2002) that no unbeveled Bolens actually exist, and that even the examples Bullen (1975) called Bolen Plain are beveled, albeit lightly. However, this apparent desire for parsimon y ends abruptly when it is suggested that Bolens st ill can be divided into nine subtypes (Schroder 2002). Subtyping Bolens remains a common pursuit, and has been based on ear type, blade shape, and notch type (Bissett 2003). One recent suggestion is that the Bolen type name be used to refer only to side notched examples, since corner notched forms may in fact be Kirk corner notched hafted bifaces (Farr 2006). This suggestion is difficult to apply to assemblages with apparently coeval side and corner notched hafted bifaces (Austin and Mitchell 2010). In general, Bolens have beveling on alternate blade margins. The base usually is straight or excurvate, and frequently is ground. Recent excavations in Florida suggest an Early Archaic date for Bolens (Austin and Mitchell 1999). The distribution of Bolen Side No tched hafted bifaces extends from the Tampa Bay region of Florida, north to the southeastern portion of Georgia (Austin and Mitchell 1999; Dunbar and Webb 1996).

PAGE 68

68 Recently, some archaeologists have suggested that we use the Bolen type name to refer to all s ide notched hafted bifaces throughout the Southeastern Early Archaic (Carter and Dunbar 2006). Specifically, they suggest the Bolen type should include the Big Sandy side notched tool type, which has been used throughout Alabama and Georgia, and the Taylor type, used primarily in South Carolina. This suggestion is based on the fact that the Big Sandy type name originally was used to describe similar looking Middle Archaic tools, and as a result, use of the Big Sandy name to describe earlier sidenotched tools is misleading and confusing. Also, they argue that use of the Taylor type name in South Carolina post dates use of the Bolen name. Unfortunately, while these suggestions fulfill a widespread desire for parsimony in Southeastern archaeological discourse, conflating three widely used types into one has not yet been justified by regional analysis of the variations within and among these types. In fact, this is one of the aims of my own research. Big Sandy. Big Sandy Side Notched hafted bifaces were defined by Madeline Kneberg (1956), who named the form after examples she excavated at the Big Sandy site in Henry County, Tennessee. Big Sandys typically have incurvate bases and elongated blades. Excavations five decades ago at the Eva Site in Tennessee suggested that the Big Sandy type dated to the earliest portion of the Middle Archaic (Lewis and Kneberg Lewis 1961). Plates of Big Sandys in the Eva site report look like Early Archaic sidenotched forms, so their stratigraphic position at the site is not clear. Recent excavations and radiocarbon dating of Big Sandy deposits at the Dust Cave site in northwestern Alabama place the tool form in the earliest portion of the Early Archaic, but use of the Big Sandy type name continues to be problematic (Driskell 1996; R andall 2002).

PAGE 69

69 Since many regional variants have been dubbed Big Sandy, it recently has been suggested that the name Big Sandy be rejected, in favor of Early SideNotched (Driskell 1996; Morse 1994; Randall 2002). The distribution of Big Sandy Side Notche d hafted bifaces extends from the panhandle of Florida, north to the western portion of Georgia. However, the highest frequencies of what many people call Big Sandys are found in the central and northwestern portions of Alabama (after Randall 2002). Other diagnostic Early Archaic artifact t ypes While side notched hafted bifaces are the diagnostic artifacts most commonly found at the earliest Early Archaic sites, other diagnostic flaked stone tools also are known. These include Edgefield Scrapers (Michie 1968) Aucilla Adzes (Gerrell et al. 1991) and Waller Knives (McGahey 1996; Waller 1971) I analyzed numerous Edgefield Scrapers for this research, but found that fewer collectors are aware of Aucilla Adzes and Waller Knives. Hence, when I broadened my rese arch scope from side notched hafted bifaces, to include contemporaneous tools, I limited analyses to Edgefield Scrapers. Edgefi eld Scrapers were defined by James Michie (1968), based on examples he saw from Edgefield County, in the southwestern portion of South Carolina. The tools typically are side notched and bifacially flaked, with steep edge angles suggestive o f use as scrapers (Michie 1968). Edgefield Scraper s may only have been seasonally used, for working bone and wood implements. The tools commonly are found in association with both Taylor and Bolen Side Notched hafted bifaces (Goodyear et al. 1980). Interestingly, while Edgefield Scrapers typically are con sidered to date to the earliest portion of the Early Archaic, they also have been found in association with later corner notched artifacts (Sassaman et al. 2002).

PAGE 70

70 Goodyear et al. (1980) noted that the Edgefield Scraper is remarkable, both in form and geogr aphic distribution, as there appears to be a geographical continuity in the scraper form from South Carolina to Florida. The haft elements of Edgefield Scrapers have been described as similar to Big Sandy, Taylor, and Bolen Side Notched hafted bifaces (Goo dyear et al. 1980). The majority of the Edgefield Scrapers has been recovered from the Coastal Plain regions of South Carolina, Georgia, and Florida, and most of the specimens are made from Coastal Plain marine cherts (Sassaman 1996). Rarely, Edgefield Scrapers made of quartz and metavolcanic rock have been recovered by avocational archaeologists. Some Edgefield Scrapers also reportedly have been recovered in southern Mississippi (McGahey 1996). However, for the purposes of my research, I focused only on those Edgefield Scrapers which occur in association with Taylor, Bolen, and Big Sandy Side Notched hafted bifaces. The effective northern boundary for Edgefield Scrapers appears to be the northernmost extent of the Coastal Plain, encompassing the Savannah Ogeechee and the SanteeCooper River drainages. The association of Edgefield Scrapers with the southern coastal plain of South Carolina b etween the Sa vannah and Santee r ivers is reinforced by the occasional use of raw materials (including orthoquartzite and Black Mingo chert) local to this region (Bridgman Sweeney et al. 2008). Summary of Early Side Notched artifact t ypes There clearly is blurring betw een most of the geographic boundaries of Early Side Notched Horizon flaked stone tool types. However, the general locations of these boundaries (especially those of side notched hafted bifaces) largely are taken for granted in Southeastern archaeology, due to interest in employing common regional typologies. Some critiques of the typological approach are based on the very nature of hafted

PAGE 71

71 bifaces, which may not be stable as artifact types, as a function of their manufacture, maintenance, and use (after Flen niken and Wilke 1989). The assumption being made here is that the artifact types constructed by archaeologists have inherent meaning (cf. Ford 1952; Spaulding 1953a, 1953b, 1954a, 1954b). For instance, in some cases, it has b een suggested that variants of side notched hafted bifaces may reflect a long te rm northward movement by hunter gatherer groups throughout the Early Holocene (Gunn and Rovner 2003). I provide discussion of this possible source of p atterned variation in the Early Side Notched Horizon in Chapter 6. Preliminary consideration of morphological variation in many of these Early Archaic artifact types suggests that it is more productive to unpack the established types, and to compare specific attributes of side notched hafted bifaces and other diagnostic flaked stone tools across space. As previously stated, these types encompass a wide range of variation, and it seems appropriate to consider the technology of the Early Side Notched Horizon in a less restrictive, and more regional, manner than ha s been attempted to date. Early Archaic s ites and settlement models Clearly, the PleistoceneHolocene transition in the American Southeast was a time of great environmental and cultural change, yet extant models of settlement during this period largely are synchronic. Perhaps, variation in settlement throughout the Early Archaic may best be seen as reflective of time and ethnicity, rather than function (after Sassaman et al. 2002). Below, I review evidence from a number of sites that assist in refining the chronology of the Early Side Notched Horizon. I also relate why a reconstruction of social boundaries in the Early Archaic is dependent on received wisdom about Paleoindian patterns of place making throughout the Coastal Plain. How can we link variation in the Early Side Notched Horizon to processes that

PAGE 72

72 distributed Early Archaic groups differentially across the landscape? I suggest that historical and multiscalar processes were at work that ultimately resulted in variations on a shared technology. As previ ously stated, the subregional technological traditions evident in the Early Side Notched Horizon arguably have culture historical roots in post Clovis Paleoindian cultural traditions. In recent years, numerous Early Holocene sites have been excavated throu ghout the Southeast, and many of the Early Archaic artifact bearing strata of those sites have been radiometrically dated. These data rarely have been compiled to assemble an effective chronology across the lower Southeast (after Anderson and Sassaman 1996, 2012). As a result, comparisons of side notched artifact assemblages largely have not been made to date. One unfortunate effect of this trend has been the perpetuation of certain regional assemblages as being especially distinctive, or as reflecting the true chronology of specific artifact types. Early Archaic s ites Sites central to this discussion include the Dust Cave (1LU496) and Stanfield Worley (1CT125) sites in Alabama; the Page Ladson (8JE591) and 8LE2105 sites in Florida; the Big Pine Tree and Topper sites in South Carolina (38AL143 and 38AL23, respectively); and the Hardaway site (31 ST 4) in North Carolina. While sites in Alabama and Florida have produced the most radiometric dates, and the most complete information about site stratigraphy and chronology, sites in other locales have helped archaeologists to refine their understandings of Early Archaic tool typology. Alabama. Dust Cave (1LU496) is recognized as one of the best stratified late PleistoceneEarly Holocene sites in the Southeast. Exc avations at the site have produced evidence of discrete components from the Late Pleistocene through the Middle Holocene, and

PAGE 73

73 the Early Archaic period occupation of the site appears to have been particularly dense (Driskell 1996; cf. Randall 2001, 2002). A t Dust Cave, the Early Archaic component is marked by high frequencies of side notched artifacts. Generally, these artifacts are described not by type names, but simply as Early Side Notched. This decision appears to have been made for the sake of parsim ony. More specifically, however, these artifacts are described as generally similar to those described elsewhere as Big Sandy I or Big Sandy and Bolen Plain (Driskell 1996:323). S amples taken within the Early Side Notc hed component at Dust Cave have been radiometrically dated to 11,20010,500 cal B.P. (Driskell 1996:325326; Sherwood et al. 2004:546). While these dates were recently calibrated (using Stuiver et al. 1998), resulting in refined dates for the Early SideNotched component the chronology of t he Early SideNotched Horizon in the Midsouth is not well understood (Sherwood et al. 2004). The Stanfield Worley bluff shelter (1CT125) is another important Alabama site. One sealed deposit at the site contained both Dalton and Big Sandy type points. This component was radiometrically dated to between 10,150 and 11,200 cal B.P. (DeJarnette et al. 1962; Josselyn 1964). According to some these dates may not reliably reflect the association of components, and Dalton should not be considered contemporary with side notched forms (Goodyear 1982) Dalton and Big Sandy toolkits recovered from the Stanfield Worley site look identical except for hafted biface morphology (Futato 1996; cf. Randall and Detwiler 2002). Florida. Two sites in Florida recently have been described as producing some of the earliest Holocene evidence for the Southeast. These sites include the Page Ladson site (8JE591)

PAGE 74

74 and 8LE2105. Both sites have produced notched points known to Florida archaeologists as Bolen, and both have been radiometrical ly dated (Faught et al. 2003). Page Ladson (8JE591) is a submerged site in the Aucilla River. Both late Pleistocene and Early Holocene sediments are found at the site, and numerous radiometric samples have been taken in recent years. Strata containing Bolen side and corner notched points produced five radiometric dates, ranging from 11,50011,250 cal B.P. (Faught and Waggoner 2012:154; Faught et al. 2003:17). Site 8LE2105 produced five side notched, and seven corner notched projectile points identified as Bolen. Three radiocarbon dates were taken at 8LE 2105; these ranged from 11,36011,200 cal B.P. ( Faught and Waggoner 2012:154; Faught et al. 2003:17). This evidence from Florida suggests that side and corner notched points are contemporary, a notion that does not fit with evidence from other excavated and radiometrically dated sites in the lower Southeast. It is possible that Florida is the only portion of the region where this pattern holds true. Alternatively, it is possible that problems exist with the samples taken to date, or that mixed contexts are reflected in the radiometric evidence. North Carolina. Initial excavations at the Hardaway site (31 ST 4) in North Carolina were part of a larger effort by the Research Laboratories of Archaeology (RLA) at UNC Chapel Hill to identify stratified sites in the Carolina Piedmont. One primary goal of this work was to identify Archaic complexes, since at the time, the Archaic sequence in the area was not well known. Archaeologists at Hardaway looked for isolated stratigraphic deposits, but no well developed deposits were found there. Additionally, no deposits at the site could be effectively radiometrically dated. Nonetheless along with Lowders Ferry and Doershuck, the site became

PAGE 75

75 critical for the establishment of an Early Archaic projectile point chronology (Anderson and Sassaman 1996), and Coes (1964) sequence of cultural phases in the Southeast was based larg ely on findings at the Hardaway site. This sequence relied heavily on tool typology as a proxy for actual chronology, and it still is employed today throughout much of the region (Daniel 1998). South Carolina. The Big Pine Tree site (38AL143) in South Carolina also is important to an understanding of Early Archaic projectile point chronology. At this site, Taylors are stratified with or above Dalton, and at the neighboring Topper site (38AL23), they are recovered directly above Clovis deposits, recently radiometrically dated to 12,841 +/ 62 cal B.P. (Goodyear 2013: 11). An approximate 2,000 year hiatus in site occupation may be indicated by soils observed between Clovis and Taylor bearing layers at the Topper site; the length of this hiatus is based on conte mporary sites that have been radiometrically dated (Goodyear et al. 2007; Albert Goodyear, personal communication 2012). While no radiocarbon dates have been obtained from Early Archaic bearing strata at these South Carolina sites, they have helped Southea stern archaeologists to better understand the relative chronology of the region (Goodyear 2001, 2003; Gunn and Rovner 2003). Early Archaic settlement m odels Any number of processes may have resulted in the differential distribution of groups across the Southeast in the earliest portion of the Early Archaic period. Variation in material culture for the Early Side Notched Horizon may reflect patterns of any or all of the following: successive migration episodes (cf. Gunn and Rovner 2003), territorial circu mscription (e.g., due to environmental constraints or social avoidance mechanisms), group fissioning, or large scale social alliances and networks. A review of ethnographic literature

PAGE 76

76 would suggest that Early Archaic period groups may well have been engage d in all of these processes, as well as others, and that these may not be archaeologically distinguishable or visible (cf. Yellen 1977). Furthermore, as reviewed in Chapter 2, we might expect Early Archaic social group membership patterns to have been rela tively fluid, with flexible social group structure and relatively permeable social boundaries (Kelly 1995; Woodburn 1968; Yellen and Harpending 1972). Below, I review models that have been proposed for Paleoindian and (later) Early Archaic settlement in th e lower Southeast, prior to presenting a hypothesis for settlement for the Early Side Notched Horizon. Some recent attempts have been made to model Early American social organization. One such effort focused on modeling social organization for the Early and Middle Paleoindian periods of Florida That work was based on a high frequency ( n =980) of hafted bifaces, the base forms of 504 of which were compared using various statistical analyses (Thulman 2006:103). Certain results of that research effort seem to suggest that the beginnings of regionalization may be sourced to the end of the Early Paleoindian period, and that during the Middle Paleoindian period, groups used collective styles in biface forms to signal group identity and membership. Certain territor ies may have been used exclusively by specific Middle Paleoindian social groups, whose material culture can now be studied to identify edge areas that overlapped with the territories of neighboring groups (Thulman 2006). While that research was heavily inf luenced by cultural evolutionary theory, and focused on Paleoindian social group differentiation and the identification of specific geographical (and by extension, social) boundaries, it has relevance to this work in that it is a recent attempt to identify some of the earliest territorial group units in the Coastal Plain Southeast. Simply put, this dissertation is an attempt to ask similar questions using

PAGE 77

77 slightly later cultural materials, at a larger scale than what was accomplished with Floridas Paleoind ian data. Since the late 1980s, two major models (Anderson and Hanson 1988; Daniel 1996, 1998, 2001) have been proposed for relating material culture patterning to social organization in the latter portion of the Early Archaic when corner notching suppla nted side notching as a hafting technique. These models include the bandmacroband, and the Uwharrie Allendale models. The bandmacroband model (Anderson and Hanson 1988) holds that bands largely restricted their mobility to individual river drainages; eig ht bands were thought to have occupied the South Atlantic Slope region ( Figure 3 1). Where distinct physiographic features (such as mountain ranges, and shifts in drainage flow) occur, according to this model, there would have been a potential for social differentiation between groups. Band mobility was marked by seasonal shifts involving movement between the Piedmont and Coastal Plain physiographic regions, and periodic aggregations are thought to have occurred among numerous bands, at the greater regiona l level of the macroband. Environmental constraints in the form of food resources are proposed to have been the primary factor for mobility, with information sharing at the level of the macroband an important benefit of seasonal macroband aggregations. Arc hitects of the bandmacroband model (Anderson and Hanson 1988) posit that aggregation loci were likely located at Fall Line river terraces. Examples of such aggregations are well known ethnographically, and have long been postulated as central features of Early Archaic lifeways (after Griffin 1952; Kelly 1995). Aggregation has been described as essential to the maintenance to Early Archaic group ties, since low density settlement is postulated for the period (Anderson 1996).

PAGE 78

78 At least four macrobands have be en proposed by Anderson and Hanson (1988) for the lower Southeast during the later portion of the Early Archaic. From north to west, and then to the south, these are the Middle Atlantic Macroband (encompassing regions north of the Neuse River drainage), the Tennessee River Cumberland Plateau Macroband (encompassing regions west of the Savannah and Pee Dee drainages), the South Atlantic Macroband (including environs from the Ocmulgee to the Neuse drainages, as described above), and the Eastern Gulf Coast Florida Macroband (which encompasses areas south and southwest of the Ocmulgee drainage, and which also abuts the southern portion of the Tennessee River Cumberland Plateau Macroband). These macrobands are thought to have contained an ywhere from 500 to 1,500 people. The Uwharrie Allendale model (Daniel 1998) was offered as an alternative to the bandmacroband model of (later) Early Archaic settlement. This model postulates that bands regularly c rossed drainages to obtain highquality raw material. Daniel (1998 ) evaluated the band macroband model (Anderson and Hanson 1988) with respect to the YadkinPee Dee River region, examining collections of (later) Early Archaic hafted bifaces in terms of two transects. The first transect cross cut several drainages, and the second paralleled the Yadkin Pee Dee drainage. The resulting raw material distributions were determined to reflect band ranges (i.e., perhaps one group preferred Allendale chert, and the other preferred Uwharrie rhyolite). The Uwharrie Allendale model (D aniel 1998) epitomizes the technological organization approach, as it suggests that groups were tethered to their raw material source of choice. According to this model, stone was a critical resource to Early Archaic groups, and sources of knappable stone (rather than watersheds) formed the focus of (later) Early Archaic settlement. Furthermore, according to this model, two distinct settlement ranges are represented by the

PAGE 79

79 Uwharrie and Allendale regions, and their respective group members preferred to utili ze different sources of high quality raw material (Uwharrie rhyolite versus Allendale chert). Models for settlement have not been developed for the earliest portion of the Early Archaic. In attempting to model social boundaries for the Early Side Notched H orizon, I must critically examine both the Anderson and Hanson (1988) and Daniel (1998 2001) models for (later) Early Archaic settlement. In general, I would argue that we must move beyond conceptualizing ancient landscapes in purely functional and physio graphic terms. For example, it is unlikely that Early Archaic groups were tethered to specific drainages or raw material sources, and knappable stone cannot have been the only determinant of social organization. Given the depth of time in consideration, and the presumably flui d and flexible nature of hunter gatherer social groupings, it also is not appropriate to extrapolate either the bandmacroband model or the Uwharrie Allendale model back in time for application to the Early Side Notched Horizon. With r espect to the earliest portion of the Early Archaic, I suggest it is most appropriate to look to models of (antecedent) Paleoindian regionalization (Anderson 1990, 1996). As previously noted, a growing body of research into (post Clovis) Paleoindian subreg ional traditions lends support to the staging area hypothesis of colonization and settlement (cf. Smallwood 2012; Thulman 2006). Therefore, I would argue that this model can be effectively projected forward in time to assist in modeling social boundaries a s reflected in the material culture of the subsequent Early Side Notched Horizon. Specifically, relative to the staging area model, I would expect to find evidence of Early Archaic settlement focused along large river systems, especially in locations wher e food and raw material resources were plentiful. As noted above, these areas had been established as places to

PAGE 80

80 dwell by ancestral populations during the Paleoindian period (after Randall 2010). I would expect to see a shift in geographical (and by extensi on, social) boundaries at physiographic shifts or zonal interfaces whether it be the Piedmont Coastal Plain boundary (Fall Line), or the interface of the Atlantic and Gulf watersheds (after Anderson and Hanson 1988; Blanton and Snow 1989; Daniel 1998). Additionally, I would expect to find evidence of periodic aggregation among regional bands at these locations (Anderson 1996; cf. Anderson and Hanson 1988; Miller 2011). Also, if side notched artifact type definitions are correct, I can presume the existence of a Taylor Zone, a Bolen Zone, and a Big Sandy Zone. These zones might be taken to reflect macroband territories, such that I might expect evidence of core areas within inferred macroband territories. These would be areas of high artifact density, where artifacts show less variation as compared to inferred boundary areas where zones intersect The groups who made and used various side notched tool forms (defined as Taylor, Bolen, an d Big Sandy) arguably shared ancestry, and may have had some level of regular interaction despite appreciable geographic separation. Participants in the Early Side Notched Horizon may have moved routinely both within and between habitual use areas. There i s support for this settlement patterning in the ethnographic literature (Gould 1982; Grinker 1994; Kent 1992, 1993; Wiessner 1977, 1982) as well as in analogous Paleoindian contexts throughout the lower Southeast. Raw Material Sources and Selection Coastal Plain chert raw material sources were used for the manufacture of sidenotched flaked stone tools throughout the lower Southeast (Austin and Mitchell 1999, 2010; Goad 1979;

PAGE 81

81 Goodyear et al. 1983; Goodyear et al. 1985; Upchurch et al. 1982). Below, I prese nt a generalized survey of highquality chert raw material sources throughout the study region. Tertiary age cherts provided the bulk of the cryptocrystalline raw materials for Paleoindian and Early Archaic peoples in the Southeastern Coastal Plain (after Goodyear et al. 1985). While the exact sources of these raw materials have not been systematically mapped throughout this region, a number of archaeological studies have been done toward that end. Additionally, lessons from sources such as the United State s Geological Survey are extremely helpful when considering why ancient hunter gatherers collected raw materials, and may even have chosen to settle, in particular places. It is relevant here to consider the geological surveys that have been done, on a stat e by state basis. Here, raw material studies in South Carolina, Georgia, Florida, and Alabama, are considered, with particular attention to how they have informed our knowledge of Coastal Plain chert acquisition and availability. Suggestions about how thes e materials may have been procured also are provided. Chert For this study, I restricted my analyses to tools made of chert (commonly known as Coastal Plain chert in my study area). While it is difficult to describe chert as a single raw material type, s ince it can be found in various geological formations, and often can be sourced to distinct qua rry clusters, as in Florida ( Upchurch et al. 1982), for the purposes of this research I justified use of the term Coastal Plain chert, as it is markedly (and v isibly) petrologically distinct from quartz, metavolcanics, rhyolites, silicified corals, quartzites, and orthoquartzites. It often is possible, again, in the cases of Florida and South Carolina, to identify distinct fossil inclusions which can help to det ermine the geologic formations where the material was formed

PAGE 82

82 (Austin 1997; Goodyear et al. 1985; Upchurch et al. 1982). Such studies can only stand to increase our knowledge of past lifeways, and perhaps also can inform discussions of exchange, trade, and movement of individuals or groups. Defining chert and its distinction from flint can be a challenge. Both are found in chalk and limestone deposits; however, cherts are normally associated with limestones. Millions of years ago, ancient seas inundated the Coastal Plain, creating limestone f ormations throughout the region. Surficially exposed cherts subsequently were found in these limestones, such that chert can be described as silicareplaced limestone (Austin and Mitchell 1999:5). In general, cherts can be described as cryptocrystalline and silica based, and chert fractures conchoidally. As such, it is an ideal material for the predictable manufacture of tools (Goad 1979). Recent geological analyses of the Coastal Plain province involved mapping the Floridan aquifer system, which is a sequence of permeable carbonate rock of Tertiary age. These rocks all are hydraulically connected, and they range in age from Jurassic to Holocene (Bush and Johnston 1988). Figure 3 2 shows this dissertation project research area, superimposed over a map of the Floridan aquifer produced by the United States Geological Service ( modified from Bush and Johnston 1988 :C4). Where Coastal Plain sediments consist of carbonate material, the rocks are somewhat dissolved by percolating water, which results in karst topography where those rocks are near the ground surface. These sediments first were laid down on a surface developed on igneous and other rocks, and today they are found to thicken toward the coast from a featheredge where they outcrop against metamorphic and igneous rocks of the Piedm ont and Appalachian provinces. Then these Coastal Plain sediments dip toward the Atlantic Ocean or the Gulf of Mexico, except

PAGE 83

83 where they are faulted or warped (Bush and Johnston 1988). Native Americans recognized these areas as prime sources of raw material for making stone tools; archaeologists call this material chert, and think of it as a sort of imperfect form of flint (Dunbar 1991, 2002; Goodyear 1979; Purdy 1986). More correctly, chert may be defined as an insoluble sili ca found in limestone deposits. Variations among chert types found throughout the Coastal Plain include color, porosity, fossiliferous inclusions, fracture properties, and texture. Differences in original package size also may exist, but not enough regiona l data exist to answer this question. Chert has been found to occur in tabular form throughout much of the Coastal Plain. As others have effectively described, high quality Coastal Plain cherts are found from Allendale County, South Carolina, west into southeastern Georgia and southeastern Alabama, and south into the regions of Tampa Bay (after Goodyear et al. 1985). South Carolina The density of Paleoindian hafted bifaces in South Carolina long has been linked to the availability of high quality cherts. T hese Tertiary Age cherts are of marine origin, and are known to occur in the Flint River Formation (Upchurch et al. 1984). In South Carolina, these cherts primarily are located in Allendale County, in the southwestern portion of the state (Daniel 1998; Goo dyear 2010; Goodyear et al. 1985). Allendale chert nodules up to 500 mm in diameter have been described (Smallwood 2010, 2012:693), although the consistency of the quality of such large nodules is unknown. A survey of Oligocene Flint River Formation chert sources in South Carolina was conducted in the early 1980s, and petrologic analyses were conducted at eleven identified quarry sites in the Coastal Plain (Goodyear et al. 1985:47). All of these quarries were

PAGE 84

84 located in Allendale County, where Paleoindian s ites already had been recorded. The quarry material was described as silicified grainstone; what long has been known to archaeologists and local collectors as Allendale chert. This material was recognized as being native to the Flint River formation, and stratigraphic testing at two sites (38AL23 and 38AL135) revealed deep (at 150 cm below the surface) Paleoindian and Archaic deposits containing artifacts of this material (Goodyear et al. 1985:47). These findings echo those made by James Michie for the Ea rly Archaic. Michie (1966, 1968) suggested that the earliest of Early Archaic artifacts (including what he defined as Edgefield Scrapers and Taylor Points) were most commonly made with Coastal Plain cherts. What is interesting to note, is that many of the artifacts Michie discussed were recovered in central South Carolina, along the Fall Line, in an area devoid of Allendale cherts. Perhaps the procurement of Allendale cherts was performed as part of a seasonal round, in that groups came to Allendale Cou nty in large part to obtain this highquality raw material for assembling their toolkits (Anderson 1996; Anderson and Hanson 1988; Daniel 1998). Georgia Figure 3 3 shows a map of Tertiary chert bearing formations of the Georgia Coastal Plai n. Both Oligocene and Eocene epoch formations are shown here. This map is modified from a study produced for the University of Georgia, as a contribution to the Wallace Reservoir project in the 1970s (Goad 1979:6, 9). Paleocene and Miocene formations are not shown on this map. As shown in this figure, cherts occur inside the Coastal Plain, south of the Fall Line. This is typical throughout the region; hence, the term Coastal Plain cherts.

PAGE 85

85 Cherts can be found throughout the east central and southwest portions of Georgia, as displayed in Figure 33 (after Goad 1979). These raw materials may vary according to color, texture, type of inclusions, and other variables, but generally all are known as cherts. The Coastal Plain cherts that extend throughout central Georgia date to the Oligocene and Eocene epochs, but the majority of deposits in the region date to the Eocene. It is important to note that most ancient lithic artifacts found in Georgia were manufactured from Coastal Plain cherts dating to the Oligocene and Eocene epoch s which are Tertiary Period formations (Goad 1979). Georgias Coastal Plain geologic formations dating to the Oligocene include those of the Appalachicola Group, and include the Alum Bluff and Chattahoochee Formations. The Vicksburg Formation also dates t o the Oligocene. Geologic formations dating to the Eocene include those of the Claiborne Group, including the Barnwell sand and McBean Formations. The Jackson, Twiggs clay, Ocala lime, Wilcox, and Midway Formations also date to the Eocene (Goad 1979). Olig ocene Epoch Formation cherts are found in a diagonal band that extends from northcentral Georgia to the southern border of the state. These cherts also can be found by the Fall Line and along the Savannah River. Outcrops along the Savannah River have been described as having different characteristics than those found elsewhere. In southwestern Georgia, between Albany and Bainbridge, cherts occur in large boulders, along the Flint River. Meanwhile, large blocks of chert can be found outcropping along the Sa vannah River, and often are found along ridgetops (Goad 1979) [refer to the above discussion of South Carolina cherts regarding the high number of chert outcrops found just across the state border, also along the Savannah River].

PAGE 86

86 Eocene Epoch cherts are fo und along the Fall Line and extend along the northern edge of the Coastal Plain. These cherts often occur in large nodules and blocks, throughout the western, central, and eastern portions of the state. Jackson Formation cherts are found along the crests of hills in Burke County, where high frequencies of artifacts were examined for my project (after Goad 1979). Eocene Epoch cherts are considered high quality raw material. High quality cherts are absent from other portions of Georgias Coastal Plain. Florid a Various Coastal Plain chert quarry clusters have been defined by researchers in Florida. Features (such as distinctive fossil inclusions and porosity) of these clusters have been identified by geologists and archaeologists. These studies have gone a long way toward assisting archaeologists with questions about ancient raw material procurement (Austin 1997; Austin and Mitchell 1999; Goodyear et al. 1983; Upchurch et al. 1982). The findings of these research efforts are briefly described here. A total of ni neteen chert quarry clusters have been identified in Florida. These clusters were defined based on particular features within the raw material that could be discerned microscopically. Figure 3 4 presents the locations of Florida chert quarry clusters (adap ted from Upchurch et al. 1982:14). Several of these quarry clusters are in the vicinity of the collected areas I examined for this study (refer to Chapter 4 for data sets used). For example, the most northwestern sites I studied (in Taylor, Jefferson, Jackson, and Lafayette counties) are in quarry cluster areas known as 1, 2, and 3 (after Upchurch et al. 1982). Respectively, these correspond to the Wrights Creek, Marianna (both just south of the Alabama border), and Wacissa quarry clusters (at the southern portion of the Florida panhandle, along the Gulf Coast). The collection I

PAGE 87

87 examined in northcentral Florida, along the Santa Fe and Suwannee Rivers, corresponds to quarry cluster number 8, which is the Lower Suwannee. For the more southern collections I st udied, in the Tampa Bay environs, quarry clusters 17 and 18 (the Hillsborough River and Turtlecrawl Point) have been defined for the region (a fter Upchurch et al. 1982) ( Figure 3 4). While the locations of these quarry clusters have been clearly defined, i t is beyond the scope of this research to determine whether the local quarry clusters were actually used for local side notched tool manufacture and/or deposition, at all localities analyzed for this project. However, some examples of such research are pro vided in Chapter 5, with respect to a pilot study performed as part of this research effort. Alabama While geological maps exist for Alabama, the extent of quartzes, orthoquartzites, and (most relevant to my studies) cherts has not been clearly demonstrated (after Copeland 1968). For example, various types of cherts have been identified by both collectors and archaeologists. Blue Fort Payne cherts are found in northern Alabama, for instance, while what others would call Coastal Plain cherts can be found out cropping in at least two counties in the southeastern portion of the state. A recent geoarchaeological study by the University of South Alabama identified several geologic outcrops of knappable stone. In southeastern Alabama, a material identified as Ocala Chert, found throughout Henry and Houston counties, fits the general description for Coastal Plain chert. While this study has not been formally published, its results are available online (Philip Carr, personal communication 2012; University of South Ala bama 2004) Through discussions with both Eugene Futato of the University of Alabama Museums Office of Archaeological Services, and an anonymous private collector, I have learned a great

PAGE 88

88 deal about the raw material distribution throughout the eastern and southern portions of Alabama. In the northeastern portion of Alabama, encompassing Jackson, DeKalb and Cherokee counties, some Coastal Plain chert has been seen, but most of the material observed can be colloquially called flint. South of that portion of t he state, in Cleburne, Randolphe, Chambers, Lee, and Russell counties, raw material seems to be approximately 90% quartz, with what is known as a band of quartz beginning near the central portion of Russell County and extending to the north. In Barbour, He nry, Houston, and the eastern portion of Dale County to the west of Henry County, Coastal Plain cherts begin to be evident. Finally, from the western portion of Dale County, to the south in Geneva County, and then to the entire western portion of the state extending to the north through Wilcox and Lowndes counties, what is known as Tallahatta quartzite dominates local assemblages and outcrops. The recent geologic study by the University of South Alabama sugge sts that Ocala (Coastal Plain) c herts are limite d to Henry and Houston counties in southeastern Alabama ( University of South Alabama 2004). It seems that the Barbour County artifacts made of Coastal Plain cherts referenced by the anonymous collector in his description of local chert availability, may actually be an example of distance from raw material source. A Regional View of Coastal Pl ain Chert Distribution As previously stated, despite the lack of formal chert sourcing studies in all portions of the Coastal Plain, it is possible to at least broadly describe the locations of highquality chert raw material in the Southeast. Figure 3 5 presents the distribution of high quality Coastal Plain chert throughout the study region. This map is an effort to portray information gleaned from a variety of sources designed to provide state level coverage of geological as well as archaeological

PAGE 89

89 information for chert distribution (Goad 1979; Goodyear et al. 1985; University of South Alabama 2004; Upchurch et al. 1982). Summary This chapter provided a review of numerous lines of archaeological and environmental evidence in a reconstruction of the geological and cultural history of the Pleistocene Holocene transition. First, I presented background for Paleoindian colonization and settlement, in the context of dramatic environmental changes throughout the lower Southeast. I suggested that variation manifested in the flake d stone tools of the Early Side Notched Horizon arguably can be traced to subregional cultural tradit ions first evident in the Middle Paleoindian period. I reviewed the type definitions of side notched tool forms (including Taylor, Bolen, Big Sandy, and Edgefield Scraper) examined for this dissertation, and critically examined certain extant models for (l ater) Early Archaic settlement. Finally, I proposed a working hypothesis for settlement in the earliest portion of the Early Archaic period, suggesting that the staging area hypothesis (Anderson 1990, 1996) can effectively be projected into the Early Holoc ene toward modeling social boundaries evident in the Early Side Notched Horizon. This chapter concluded with a regional review of highquality Coastal Plain chert sources throughout the lower Southeast, as raw material is necessarily an important component in modeling Early Archaic settlement.

PAGE 90

90 Figure 3 1. Band ranges for the South Atlantic Macroband in the Early Archaic ( after Anderson, David G., and Glen T. Hanson. 1988. Early Archaic Settlement in the Southeastern United States: A Case from the Savannah River. American Antiquity 53(2):262286. [Page 269, Figure 3] )

PAGE 91

91 Figure 3 2. Project research area and the Floridan aquifer (modified from Bush, Pe ter W., and Richard H. Johnston. 1988. GroundWater Hydraulics, Regional Flow and GroundWater De velopment of the Floridan Aquifer System in Florida and in Parts of Georgia, South Carolina, and Alabama. U.S. Geological Survey Professional Paper 1403C. United States Government Printing Office, Washington, D.C. [Page C4, Figure 1] )

PAGE 92

92 Figure 3 3. Coastal Plain chert distributions throughout Georgia ( adapted from Goad, Sharon I. 1979. Chert Resources in Georgia: Archaeological and Geological Perspectives. Wallace Reservoir Project Contribution 3. University of Georgia Laboratory of Archaeology Ser ies Report Number 21, Athens. [Page 6, Figure 1 and Page 9, Figure 3])

PAGE 93

93 Figure 3 4. The locations of Florida chert quarry clusters ( adapted from Upchurch, Sam B., Richard N. Strohm, and Mark G. Nuckels 1982. Methods of Provenance Determination of Flo rida Cherts. Report prepared for the Florida Division of Archives, History, and Records Management by the Department of Geology, University of South Florida, Tampa. [Page 14] )

PAGE 94

94 Figure 3 5. The distribution of high quality Coastal Plain chert throughout t he study region

PAGE 95

95 CHAPTER 4 RESEARCH METHODS AND DATA SETS In this dissertation, my research question relates to the use of patterned variation in material culture to infer social boundaries for the Early Archaic Southeast. I began this research by selecting a set of artifact attributes for analysis in order to measure the variation I observed in two classes of Early SideNotched material culture; side notched hafted bifaces and unifaces. This chapter begins with discussion of those attributes and the statistical techniques used to clarify observed patterning in collected data. In selecting particular attributes, I remained mindful that artifact types (e.g., Taylor, Bolen, and Big Sandy) are unlikely to be delimited by clear geographic boundaries Als o in this chapter, I describe the research area, and introduce all data sets used in the context of state as well as more localized provenance. Next I explain how the data were assigned to specific drainage and watershed systems for comparative regional analyses of patterned variation Finally, I describe how I analyzed variation in these assemblages relating my efforts to compare the range of patterned variation at the core areas of Taylor, Bolen, and Big Sandy artifact type distribution, to that of infe rred boundary areas within the Early Side Notched Horizon. Attributes for Analysis I performed analyses at two levels: at the gross level of the individual objects (e.g., the individual hafted bifaces and unifaces), and at the finer level of individual art ifact attributes. These units i nclude both nominal and ratioscale variables, and were plotted through space, as well as in association to one another. Numerous attributes have been employed to distinguish one artifact type from another. Many of these were chosen to be diagnostic of certain cultural, spatial, or temporal boundaries.

PAGE 96

96 Some examples of these include overall tool shape, tool size, base shape, presence or absence of grinding or resharpening, presence of haft element, presence and position of not ches, and presence or absence of lateral thinning or fluting. When a minimum number of such attributes can be seen to consistently co occur, an artifact type may be constructed by archaeologists, who then use the types to make inferences about past lifeway s (Adams and Adams 1991; Andrefsky 1998; Krieger 1944). Individual Objects All individual artifacts in my analyses were coded according to the county and state in which they were collected, or the specific location or state numbered archaeological site wh ere they were recovered. Each of these objects also was coded according to a variety of specific attributes. These attributes, both nominal and ratioscale, are described below. Distinct identification numbers then were assigned to each artifact in the analysis. These numbers correspond to all data entries made for each individual artifact. All metric attributes were taken with digital calipers. The same set of calipers was used throughout the project to maintain consistency, and all measurements were taken by the author for the same reason. Most artifacts in this study had been broken at some point, along at least one edge. I attempted to determine whether these breaks had occurred during the use life of the tool, or whether they were more likely to be products of post depositional events or processes. Approximately half of all experimentally produced tools have been found to break during first use (Cheshier and Kelly 2006). Fracture types recorded include: hinge fracture, step fracture, perverse fracture, l ateral snap, crenated, potlid, impact fracture, half snap, miscellaneous

PAGE 97

97 shoulder damage, base break, and radial break. These artifact breakage patterns were based on a variety of published reports (Cheshier and Kelly 2006; Purdy 1975). Very little ethnogr aphic data are available about the effects of resharpening and rejuvenation on stone tools (Binford 1986; Towner and Warburton 1990). Few data also are available regarding the incidence of fractures that occur during the manufacturing process. However, man y inferences may be made using the archaeological record, and these have been bolstered by a great deal of experimental archaeology (Johnson 1979; Keeley 1982; Purdy 1975; Titmus 1985; Towner and Warburton 1990). Metric data were tabulated as to specific collection or site, as well as to region. Subsequent analyses of metric attributes took only those measurements of complete attributes into account for the purpose of comparison. For example, if a hafted biface had one snapped shoulder, its remaining shoulder width was not used in this analysis. Similarly, if an Edgefield Scraper was missing one ear, its base width was not used in this analysis. I photographed all artifacts using a Sony CD MavicaTM camera and metric scale. All images were recorded onto compact discs for future reference. While a representative sample of artifact photographs is provided in this document, specific images of interest to archaeologists or other members of the public are available upon request. Individual Artifact Attributes Attributes collected from artifacts varied depending on the artifact type. For instance, certain attributes can be recorded from hafted unifaces, which cannot be recorded from hafted bifaces. All metric attributes (with the exception of maximum thickness) for b oth types of artifacts are displayed in Figures 4 1 and 42. These attributes are described below.

PAGE 98

98 Hafted bifaces Attributes recorded for each individual side notched hafted biface i nclude both nominal and ratio scale attributes ( Figure 4 1). Nominal attri butes were selected since it is reasonable to assume that native toolmakers also were conscious of nominal variation (after Spaulding 1982). This is not meant to imply the existence of a mental template, but rather to suggest native groups observed variati on in thei r tools as we do today. Ratioscale attributes are commonly taken during artifact analysis, and are seen to complement knowledge of nominal attribution variation (Cowgill 1982). Attributes for analysis of the hafted bifaces were selected with the goal of identifying technofunctional attributes as well as possible stylistic markers. Attributes I recorded for all side notched hafted bifaces in this study include many of those used by previous researchers with respect to the Early Archaic (Ahler 1971; Bissett 2003; Randall 2002). These include the following: Condition which portion of the tool remains? Proximal, distal, medial, lateral? Manufacturing condition is the tool intact? Broken/aborted? Is the break use related (e.g., impact fracture)? Fracture type is there shoulder damage? Step fracturing? An impact fracture? A lateral snap? Basal grinding rated 0 4, with 4 as the heaviest amount of grinding, and 0 having no grinding Cross section is it biconvex? Rhomboidal? Plano convex? Later al condition is the tool complete? Or, is it missing one or both shoulders? Blade length (mm) taken using digital calipers, measuring from the shoulder edge to the distal tip of the blade Shoulder width (mm) taken using digital calipers, measuring fro m the very tips of the shoulders

PAGE 99

99 Between notch width (mm) taken using digital calipers, measuring from the deepest part of the notches Maximum thickness (mm) taken using digital calipers at the thickest portion of the artifact, which normally was found to be near the base of the blade, above where it meets the haft element Beveling is there evidence of blade maintenance or rejuvenation, indicated by the removal of small flakes from one or more edges of the blade to create a resharpened surface? None? Right? Left? Alternate? Not ch type one of four basic side notching types Basal configuration [incurvate, straight, and excurvate], based on conformity or variance from a straight line that parallels the bottom of the proximal tool edge Basal concavity width (mm) taken for incurvate artifacts, using digital calipers Base width (mm) taken using digital calipers Base depth (mm) taken for incurvate artifacts, using digital calipers Auricle height (mm) taken using digital calipers Notch height (mm) taken using digital calipers Notch width (mm) taken using digital calipers Shoulder length (mm) the length of the haft element, taken using digital calipers Basal configuration is presumed to remain relatively unchanged during the use life of a tool. This is because, as an attribute of the haft element, basal configuration is typically not affected by resharpening. Descriptions of basal configuration have comm only been used to define different tool types identified in the archaeological record. As noted above, the three classes of basal configuration I observed in side notched tools throughout the study region include incurvate, straight, and excurvate. These c lasses correspond to Bullens (1975) use of the

PAGE 100

100 categories of concave, straight, and convex; I chose different verbiage due to the ease of selecting a single abbreviating descriptive letter (e.g., I, S, or E) during data analyses. I am assuming that incurv ate, straight, and excurvate base shape variations were coeval At one Early Archaic site (the Jeanies Better Back site) where hafted biface basal configuration was recorded during previous research, archaeologists observed no apparent distinctions in ver tical distribution for variations in basal configuration, suggesting that these variations were contemporary (Austin and Mitchell 2010). As other researchers have done, I considered each sidenotched hafted biface to have two gross parts; the blade (distal portion) and the haft element (proximal portion). The haft element is that portion of the artifact below the shoulders, and the length of the haft element in my study is called the shoulder length. This distinction has been of use to other researchers, in understanding the differential patterns of use wear and tool maintenance known to occur during the Early Holocene (Ahler 1971; Austin and Mitchell 2010; Bissett 2003; Goodyear 1995 [1974]). Most typological emphasis has been placed on the haft element, be cause bifaces can break during use, and most diagnostic attributes tend to be found there (Flenniken and Wilke 1989). Unifaces The only unifaces examined for this dissertation are Edgefield Scrapers (cf. Michie 1966). As described in Chapter 3, these artif acts are similar in form to side notched hafted bifaces, with the exception that the latter are worked bifacially. Therefore, most of the metric attributes recorded for Edgefield Scrapers largely are similar to attributes I would record for any side notche d flaked stone tool ( Figure 4 2). Attributes specific to the Edgefield Scraper include the length and angle of the primary edge, which in almost all cases, meant the left lateral edge.

PAGE 101

101 The presence or absence of any concavity along the primary edge was not ed. Handedness (right versus left) also was recorded for each uniface. Since Edgefield Scrapers typically are asymmetrical, I recorded both sides of each example separately. The attributes I considered in my study of Edgefield Scrapers are similar to those considered by other researchers in the past (Goodyear et al. 1980). As with respect to side notched hafted bifaces, I considered each side notched hafted uniface to have two gross parts; the blade (distal portion) and the haft element (proximal portion). Statistical Techniques Descriptive statistics used in the analysis of data for this project include the following: coefficient of variation, standard deviation, standard error, skewness, and kurtosis. I also performed Levenes test for all ratio scale data; this is an inferential statistic used to assess the equality of population variances (Gastwirth et al. 2009; Levene 1960). Additionally, I conducted chi square analyses as tests of association between specific variables. These are typical statistical tools used in anthropology and archaeology (Bernard 2002; Drennan 1996; Kvamme et al. 1996). Here, I provide brief descriptions of these techniques, and note the situations where they were applied. Descriptive Statistics Coefficient of variation (c.v.) is calculated as a ratio of the standard deviation to the mean of a sample. This is a measure of dispersion of a probability distribution, and is a way to standardize comparisons of standard deviation when means are not equal (Drennan 1996) A c.v. typically is interpreted relative to other coefficients of variation in a sample. Measurements of coefficient of variation can be compared across data sets to determine relative variation for those

PAGE 102

102 samples For instance, a higher coefficient of variation for a speci fic measured attribute of hafted bifaces from one site assemblage, compared to many others in the region, would indicate a wider range of variation for that measured attribute at that particular site (cf. Austin and Mitchell 2010) I calculated coefficient of variation for all individual and aggregated data sets. The results of numerous measurements of coefficient of variation conducted for this research appear throughout Chapter 5. Measurements of standard deviation (s.d.) are based on deviations from the mean. Standard deviation is the typical distance from any point in the data set to the mean (after Bernard 2002; Drennan 1996). As previously noted, standard deviations are needed for calculations of coefficients of variation. The results of numerous meas urements of standard deviation appear in descriptive statistics tables throughout Chapter 5. Standard error (s.e.) is the standard deviation of a sampling distribution of means. This measurement indicates how much the mean of a sample varies from the mean of a given population (Bernard 2002; Drennan 1996). Standard error is calculated by dividing the standard deviation of a sample by the square root of the population size. I calculated standard error for all aggregated (drainage wide) data sets. Skewness m easures the asymmetry of the probability distribution of a random variable around its mean. Normal distributions are symmetric because they have a skewness value of zero. A negative skew indicates that the tail on the left side of the probability density f unction is longer (toward more negative values) than the right side, and that the bulk of the data lie to the right of the mean. A positive skew indicates that the tail on the right side is longer than the left side and the bulk of the values lie to the le ft of the mean (Bernard 2002). For instance, a high

PAGE 103

103 skewness value might indicate that most of the artifacts in the sample have measurements near the maximum value. Not all symmetric distributions are normal. Kurtosis measures the peakedness or flatness of a probability distribution of a random variable. Positive kurtosis indicates a relatively peaked distribution. For example, a comparatively high kurtosis value would indicate that that metric variable is the least varied among considered attributes Nega tive kurtosis indicates a relatively flat distribution. Like skewness, kurtosis can be used to describe the shape of a probability distribution The kurtosis for a standard normal distribution is 3. In a leptokurtic (which means thin bulge in Greek) distribution, the kurtosis is greater than 3, and in a platykurtic (which means flat bulge), distribution the kurtosis is less than 3 (Bernard 2002). A l eptokurtic distribution might indicate that the data in consideration are highly concentrated around the mean due to low variations for the attributes in a sample. A platykurtic distribution might point to a high degree of variation for the attributes in a sample. I calculated skewness and kurtosis for all aggregated data sets. The results of numerous measurements of skewness and kurtosis conducted for this research appear throughout Chapter 5. For all ratio scale attributes considered in this disse rtation, I provide individual descriptive tables including the following: n, mean, min, max, s.d., and c.v. These tables are included in Chapter 5. Inferential Statistics The Levenes test of equality of variance (Gastwirth et al. 2009; Levene 1960) allows for the simultaneous testing of multiple samples, testing the null hypothesis that variances of populations from which samples are drawn are equal (known as homogeneity of variance, or

PAGE 104

104 homoscedasticity). If the resulting pvalue of Levene's test is less t han some critical value (typically 0.05, which is based on a 95% confidence interval), the obtained differences in sample variances are unlikely to have occurred based on random sampling from a population with equal variances. A rejected null hypothesis fr om this test suggests that there is a statistically significant difference between the variances in the population. The Levenes test of equality of varianc e was conducted for all ratio scale data; resulting tables are presented in Chapter 5. Tests of Asso ciation Chi square is a test of the association between variables of two populations. This test allows for the comparison of samples by estimating population proportions, and resulting in a single probability value. The chi square test is accomplished by c omparing a table of expected values with observed values. Resulting deviations are observed deviations from these expected values (Drennan 1996). I used chi square tests to make comparisons of types of basal configuration for all artifact samples in this s tudy. The results of these chi square tests appear in Chapter 5. Sample Selection and the Use of Collected Material This examination of patterned variation in the Early Side Notched Horizon holds raw material constant. I analyzed sidenotched artifacts that were made from only Coastal Plain chert, the sources for which were discussed in Chapter 3. This restriction on ra w material allowed control over at least one variable in the study of Early Archaic hafted bifaces and unifaces. Oftentimes, patterned variation among suites of lithic tools is explained by pointing to the diversity of raw material types within artifact as semblages. Alternative explanations for observed variation must also be considered. Considering only those tools manufactured from Coastal Plain

PAGE 105

105 chert should result in observed variation being a function of something other than raw material choice. For ins tance, variation in the use of Coastal Plain cherts throughout the lower Southeast may well be a function of distance from source; this idea is explored in Chapter 5 with respect to specific research results. As part of my research for this dissertation p roject, I conducted analyses of a few excavated collections housed in curation facilities and at universities. I also conducted analyses of numerous private collections, many of which currently are housed at the homes of individual artifact collectors, or in banks or local museums. These collections are from various locales in my study area, which includes South Carolina, Georgia, Florida, and Alabama. In this chapter, I provide details about the various data sets I studied in each state. One artifact colle ctor requested I keep his identity private, and so I identify him here as X. The use of artifacts collected by amateur archaeologists is not without controversy. Many archaeologists think that working with collectors serves as a kind of tacit approval of t heir actions, and that artifact collecting is destroying the nonrenewable archaeological record. Currently, there are no established professional standards for the use of private artifact collections in professional research. This is likely because there i s no consensus about the ethical issues involved (after Thulman 2006). My own stance is that we would do well to learn as much as possible from extant collections, and that interactions between archaeologists and collectors can be mutually beneficial (afte r Luchterhand 1970; Sassaman 2003). If it is not necessary to excavate sites, and if cultural resources are not being actively threatened, I would argue that archaeologists should consider other means to learn about the ancient past (cf. Anderson 2003).

PAGE 106

106 Ar tifacts collected by avocationalists form the bulk of this study ( n=1,962). As the majority of assemblages considered for these investigations were not professionally collected, certain collector biases were widely apparent. For example, some of the collectors I interviewed collect all obvious tools that they encounter. Other collectors admit to being more interested in collecting hafted bifaces than other tools (such as unifaces) they see. Certain collectors collect only complete or almost complete tools, and others collect whatever tool fragments they find. Some professionally excavated Early SideNotched Horizon artifacts also were recorded for this research ( n =58); the variation in these assemblages in turn reflects professional biases toward complete do cumentation of excavated materials. All of these variables, and others, together have created the artifact assemblage used for my analyses of 2,020 sidenotched artifacts from across the South Atlantic Slope and Eastern Gulf Coast. Research Area and Data Sets Used The research area for this dissertation project is the Coastal Plain physiographic province, which begins in central South Carolina, and continues west throughout Georgia and southeastern Alabama, as well as south into northcentral Florida, unti l jus t south of the Tampa Bay region. This area encompasses some 240,283 km. The topography of much of the research area is characterized by extensive and dissected plains, as well as low hills. The basis for recognizing the Coastal Plain as a physiograph ic province distinct from the Piedmont, which exists to the north of what geologists call the Fall Line (the boundary between the Coastal Plain and the Piedmont physiographic provinces), is rooted in geological history, and specifically, the distinctive or der and character of sedimentary rocks in these regions.

PAGE 107

107 Since American cultural resources often are managed at the level of the state, I first consider the side notched artifacts with respect to their state of origin. Within each state, I then consider these artifacts with respect to the specific locations where they were found. Next, I explain how I aggregated these data, in order to develop a regional perspective for the side notched artifacts in this portion of the Southeast. Figure 4 3 presents the res earch area, in which examined materials were ascribed to one of seven analytical units, as discussed i n detail later in this chapter. Table 4 1 presents the data sets I examined according to state, as well as the individuals who assisted me in gaining acce ss to artifact collections. Where state site numbers have been assigned to these locales, I provide that information. Other collections were made from entire counties, and where that is the case I simply provide the county name. County level data are commo nly and successfully used by Southeastern archaeologists, especially those who work with regional modeling (e.g., Anderson 1996; Charles 1986; Sassaman et al. 1988). Table 42 displays the frequencies of hafted bifaces and unifaces examined from assemblages throughout the Coastal Plain. South Carolina In South Carolina, I examined artifacts from two different regions. I analyzed material that had been collected, in the northeasternmost portion of the Coastal Plain (in Berkeley, Calhoun, and Clarendon count ies), in the Lakes Marion and Moultrie region. I also collected data from professionally excavated as well as collected locales in Allendale County, South Carolina. Figure 4 4 shows the locations of these sites and collection areas.

PAGE 108

108 Table 4 1. Data sets used in analysis State Region Site Collection Contact SC Lakes Marion and Moultrie Elloree Museum Elloree Museum Elloree Bank Elloree Bank Skinner Jimmy Skinner Hebert Ted and Darby Hebert Allendale County Big Pine Tree SCIAA Charles SCIAA Topper SCIAA Strong Larry Strong GA Burke County Owl and Cotton Creek Danny Greenway, Danny Ely, and Seaborn Roberts Burke County Greenway, Ely, and Roberts Coffee County Feronia Site Complex Frankie Snow Coffee County Snow Dooly County Dooly County John Arena and John Whatley Worth County Worth County Tommy Bailey Early County Early County Marvin Singletary Miller County Miller County Bud Carter AL Barbour County Barbour County X Henry County Henry County Don Marley Houston County Houston County X, Marley Geneva County Geneva County Marley FL Jackson County Jackson County Marley Jefferson County Applewhite Billy Dean Jefferson County Dean Taylor County Lake Bird Brian Evensen Lafayette County Jeanie's Better Back SEARCH, Inc. Santa Fe and Suwannee Rivers Hendrix Alvin Hendrix Hillsborough County Harney Flats Jerry Spencer Harris Grove Evensen and Spencer

PAGE 109

109 Table 4 1. Continued State Region Site Collection Contact Pinellas County Kepler Jim Drawdy Pinellas County Evensen and Spencer St. Petersburg Area Evensen Hernando County Spring Hill Spencer Pasco County Pasco County Spencer Table 4 2. Frequencies of artifacts from assemblages throughout the Coastal Plain Assemblage Hafted Bifaces (n) Unifaces (n) Elloree Museum 6 Elloree Bank 22 Skinner 33 4 Hebert 5 Big Pine Tree 15 4 Charles 3 Topper 15 Strong 411 149 Owl and Cotton Creek 134 35 Burke County 25 Feronia Site Complex 37 9 Coffee County 37 Dooly County 45 Worth County 52 Early County 26 Miller County 97 Barbour County 43 1 Henry County 23 Houston County 48 Geneva County 66 Jackson County 81 Applewhite 87 Jefferson County 14 Lake Bird 74 Jeanie's Better Back 21 Hendrix 268 39 Harney Flats 5

PAGE 110

110 Table 4 2. Continued Assemblage Hafted Bifaces (n) Unifaces (n) Harris Grove 44 7 Kepler 2 1 Pinellas County 14 1 St. Petersburg Area 3 3 Spring Hill 7 1 Pasco County 2 1 Total 1,765 255 Lakes Marion and Moultrie region artifacts Several investigations into Early Archaic sites have been conducted in the central portion of the state, below the Fall Line. Examples include the Mattassee Lakes project, along the lower Santee River, as well as several surveys along the Congaree River dr ainage. These projects, both along and south of the Fall Line, indicate a great deal of raw material diversity, a pattern that may suggest aggregation by groups from different portions of the state (Anderson 1996; Anderson and Hanson 1988; Anderson et al. 1982; Goodyear 1975; Michie 1969). Early Archaic sites in the middle and lower Coastal Plain of the state have been investigated using different coverage than that used in the Savannah River valley (discussed below), and they have produced very different kinds of evidence. In particular, it is not surprising that these sites have produced a greater diversity of lithic raw materials (Sassaman 1996). Central South Car olina is located away from highquality cherts, and many artifacts in this region are made of locally available quartzes and orthoquartzites. A s the nearest locations of high quality Tertiary cherts are to the south and west, chert artifacts found in this region must reflect tool replacement at some travel cost to the toolmakers (Figure 4 5). I co nducted analyses of artifact collections from central Sou th Carolina ( Figures 4 4 and 45). Most of these collections were from the Lakes Marion and Moultrie region, which is a

PAGE 111

111 drowned riverbed for the Santee. These collections now are housed in a variety of places, including one bank (Elloree Bank and Trust), one museum (Elloree Museum), and in several private residences, including those of Dotty Overson, Jimmy Skinner, Ronnie Caison, and Gary LeCroy of Pinopolis, Ted Hebert of Santee, and Darby Hebert of Summerville. Artifacts examined from Caison, LeCroy, Overson, and Skinner are collectively referred to as the Skinner Collection since these were all examined at Skinners place of residence in Pinopolis. Ted and Darby Herberts artifacts are collectively referred to as the Hebert Collection. A total of 66 hafted bifaces was recorded from the Lakes Marion and Moultrie region; half of these were from the Skinner Collection ( n=33) ( Table 4 2). The Skinner Collection is also the only data set among the Lakes Marion and Moultrie group that had unifaces ( n =4) analyzed for this investigation. Allendale County artifacts Allendale County lies adjacent to the Savannah River, in the southwestern portion of South Carolina ( Figure 4 4). The county is predominantly rural, and is the location of high quality Coastal P lain cherts ( Figure 4 5). The location of Allendale C ounty near fresh water and high quality raw materials would have made it an id eal location for ancient hunter gatherers (Goodyear and Charles 1984). Re cent attempts to document Paleoindian settlement in the region have been undertaken at several sites along the river. These excavations have been conducted by staff and volunteers in coordination with the SCIAA and USC (Goodyear 1999, 2013; Goodyear and St effy 2003). I examined material from three of these professionally excavated sites, as well as from one large personal collection (belonging to Larry Strong), in Allendale County, South Carolina.

PAGE 112

112 I analyzed material recovered from controlled excavations at three multicomponent archaeological sites in Allendale County, South Carolina. These sites include the Big Pine Tree site (38AL143), the Charles site (38AL135), and the Topper site (38AL23). All three sites were investigated as part of the Southeastern P aleoamerican Survey, headed by Albert Goodyear of the South Carolina Institute of Archaeology and Anthropology ( SCIAA ) at USC (Goodyear 2010). The Topper site is of special note, as it appears to contain the purest Taylor side notched component in the regi on (after Sassaman et al. 2002; Albert Goodyear, personal communication 2012). Larry Strong of Allendale has collected flaked stone artifacts throughout Allendale County for over twenty years, and has maintained a high degree of communication with archaeol ogists at SCIAA especially with Albert Goodyear and Tommy Charles. Strong recently donated a large part of his personal artifact collection to SCIAA, and he continues to collect to the present time. A total of 444 hafted bifaces was recorded from Allendal e County sites/collection locales; the majority of these were from the Strong C ollection (approximately 93% ) (Table 4 2) A total of 153 unifaces was examined from Allendale County sites/collection locales; the Big Pine Tree site ( n =4) and the Strong Colle ction ( n=149). Georgia Cultural resource management archaeologists have excavated several Early Archaic sites in the Georgia Coastal Plain, and many side notched artifacts have been recovered from stratified contexts. These efforts, most of which were focused in the eastern port ion of the state, have been described in cultural management reports (Elliott 2006; Elliott and Doyon 1981). One site, called

PAGE 113

113 the Theriault site (9BK2), was excavated in the 1960s and 1970s, but data collected from that location have not been fully examine d by archaeologists, and no complete report of findings there has yet been made (cf. Brockington 1971; Elliot t 2006; Midgette 2005). Additionally, an informal and ongoing survey of a 20 county area in southcentral Georgia has identified numerous sites wit h Early Archaic components; the cultural materials from these sites await further study (Elliott 2006; Snow 1977a, b). Some extensive Early Archaic sites from the Piedmont (specifically, the Oconee River drainage) have been examined by archaeologists (OSt een 1983, 1996), and many of those collected materials were reportedly manufactured from Coastal Plain chert. In discussions with one of the principal investigators, this raw material identification may have been incorrect (Lisa OSteen, personal communica tion 2004). Therefore, these materials were not examined as part of this research. Data I collected from Georgia do not include any professionally excavated artifacts. Rather, since I was interested in examining large frequencies of artifacts, I was depend ent on collected materials throughout the state. Georgia collection locales examined for my research include the specific locations of the Owl and Cotton Creek sites (collectively recorded as 9BK113) in Burke County, as well as the Feronia Site Complex in Coffee County. Collections made from entire Georgia counties include those from Burke, Coffee, Dooly, Worth, Early, and Miller counties. As discussed in Chapter 3, high quality cherts are found throughout a large portion of central Georgia. However, they are not found in the southeas tern portion of Georgia ( Figure 3 5).

PAGE 114

114 Burke County artifacts I was able to examine large personal collections made throughout Burke County in southeastern Georgia, as well as at two archa eological sites in Burke County: the Owl and Cotton Creek sites (collectively recorded as 9BK113). The Owl and Cotton Creek sites are located along a s wamp of the Ogeechee River ( Figure 4 4). The Owl Creek site occupies approximately 2.4 hectares of a small hill, which is at only slightl y higher elevation than the surrounding terrain. The Cotton Creek site is located approximately 0.8 kilometers (km) upriver from the Owl Creek site. Personal collections made at the Owl and Cotton Creek sites are housed in several homes in Swainsboro, Geor gia. The Owl and Cotton Creek sites have been collected since 1984, when the area was logged and artifacts were exposed. I observed artifacts from the site which ranged in date from the Paleoindian through the Mississippian periods; occupation of the site seems to have been most frequent during the Archaic culture periods. I had the opportunity to interview Danny Greenway and three of his friends and fellow collectors (namely, Danny Ely, Emery Fennell, and Seaborn Roberts), who also keep collections of Owl and Cotton Creek site material at their homes. For the most part, these individuals keep only those materials which they personally have collected at their homes. However, some trade of individual artifacts has taken place among the friends, and they all r ecall which artifacts they found, and which they were traded. All four men are deeply interested in keeping the site collection in Georgia for posterity. Unfortunately, the current relationship between archaeologists and artifact collectors in this state h as been weakened by recent legislation which

PAGE 115

115 prohibits people from collecting artifacts on private lands without written permission of the landowner. A total of 159 hafted bifaces was recorded from Burke County sites/collection locales; the majority of the se were from the Owl and Cotton Creek sites (approximately 84% ) (Table 4 2) A total of 35 unifaces was examined from Burke County sites/collection locales, all of which were from the Owl and Cotton Creek sites. Coffee County artifacts One collection from Georgia had especially good integrity, as it was made from a complex of numerous sites along the Ocmulgee River in the central part of the state. This area is known as the Feronia locality, and it contains many archaeological sites which have state site numbers. The Feronia locality is located in northern Coffee County, in southcentral Georgia (Figure 4 6). Frankie Snow of South Georgia College collected the Feronia locality beginning in 1978, when the sites were exposed by land clearing in advance of construction. Snow recorded sixteen sites at the Feronia locality (Blanton and Snow 1986) and is to be commended for his careful collecting and recording activities, which he performed alone, and without funding. Frankie Snow is not a formally trained archae ologist, and only performed surface collections at the sites. As no excavations have been performed at the locality, no subsurface artifacts or features have been recorded. Fortunately, a great deal of information about the cultural periods and the natures of activities performed at the Feronia locality still can be gained from the extant collection material.

PAGE 116

116 Twelve of the sites in the Feronia locality are located on a ridge above the Ocmulgee River floodplain, and the remaining four sites are located withi n that floodplain. The site complex comprises approximately 4 km. While cultural materials from later historical periods (up to the early twentieth century) were found to be present at many of the sites, the Feronia locality primarily is known as an inten sive concentration of Paleoindian and Early Archaic artifacts. The Feronia locality is notable for the presence of four pitted egg stones, made of ferruginous sandstone. These objects, the function of which is unknown, were produced by pecking and grinding and a dimple or facet was observed on one end of each stone (Blanton and Snow 1986, 1989). The pitted egg stones recovered at the Feronia sites were reportedly similar to those reported from early sites in Florida where they were described as bola stone s (Blanton and Snow 1986; Nei ll 1971; Purdy 1986:23, 30). The Feronia locality is situated far from any known source of lithic raw material; this is atypical of Early Archaic site placement (cf. Daniel 2001). Figure 4 7 displays the location of the Feronia sites in the context of Coastal Plain cherts (after Goad 1979). Specifically, while roughly 99% of the lithic artifacts recovered at the Feronia sites are manufactured from Coastal Plain chert, the closest source of that material is located ap proximately 83 km to the northwe st (Blanton and Snow 1986, 1989) Nearer cherts were available, but as discussed in Chapter 3, Eocene age cherts are considered high quality raw material, and these are located at a considerable distance from the Feronia sites. Such a s ite location is unusual given most models of early land use in the Southeast (cf. Daniel 1996; Goodyear 1979). The locale makes more sense when one recognizes its position at the boundary of the Gulf and Atlantic watersheds; boundaries proposed over two de cades ago to explain Early Archaic settlement patterns in this

PAGE 117

117 region (Anderson and Hanson 1988; Blanton and Snow 1986, 1989). Additionally, the position of the locality near two springs, as well as near the Ocmulgee River, would have attracted Native Amer icans to the region. Previous researchers have suggested that the Feronia sites may have served as a place of aggregation (Blanton and Snow 1986, 1989). Three of the sites in the Feronia Site Complex were helpful for my study (9CF22, 9CF124, and 9CF132). Site 9CF22 is situated along the Ocmulgee River terrace. Site 9CF124 is located along the floodplain of the Ocmulgee River. Site 9CF132 is situated along the side of the sand ridge located to the south of the Ocmulgee River floodplain. Since these sites we re recorded as surface artifact scatters during salvage archaeological investigations, the artifacts analyzed from these three sites are collectively referred to as the Feronia locality with regards to my investigation. Additional artifacts from various lo cations within Coffee County that are not a part of the Feronia locality were also examined for my study. These artifacts are currently housed at Snows office at South Georgia College in Douglas. Collectively, these other artifacts from Snows collection that are not a part of the Feronia locality are referred to as the Coffee County Collection. A total of 74 hafted bifaces was recorded from Coffee County sites/collection locales, half of which are from the Feronia locality and half of which are from the Coffee County Collection ( Table 4 2). Nine unifaces were examined from Coffee County sites/collection locales, all of which were from the Feronia locality. Dooly County artifacts I analyzed artifacts that were recovered from Dooly County, Georgia that had been collected by John Arena of Evans, and John Whatley of Thomson. Whatley, a professional

PAGE 118

118 mineralogist, has used his personal collection of artifacts as a partial basis for his 2002 publication on the typology of Georgia stone tools (Whatley 2002). Artif acts from both Arena and Whatley are unprovenienced with regards to a formally recorded site. I combined the artifacts analyzed from both collections into the Dooly County Collection ( Figure 4 6). A total of 45 side notched hafted bifaces was recorded from Dooly County ( Table 4 2). Unifaces are absent among the Dooly County sample of Early Archaic artifacts. Worth County artifacts Ty Ty, Georgia resident Tommy Bailey had a collection of Early Archaic artifacts from near his home in Worth County. These artif acts were from various locations, and do not correspond to a state accessioned archaeological site. Collectively, all the material from Baileys collection is referred to as the Worth County Collection ( Figure 4 6). I recorded a total of 52 hafted bifaces from the Worth County Collection, and no unifaces were represented among this artifact assemblage ( Table 4 2). Early County artifacts Artifacts examined from Early County were collected by Marvin Singletary on his private lands in Blakely. Singletary estimated that he had recovered several thousand artifacts from his land, some of which are manufactured using chert from local sinkholes in the nearby area. In the past, he has shared his collection data with other professional archaeologists, including D aniel Elliott, Michael Faught, and James Waggoner. No formal state accessioned site numbers correspond to the locations which Singletary collected these artifacts. The artifacts I analyzed from Singletarys collection are collectively referred to as the Early County Collection (Figure 4-

PAGE 119

119 8). A total of 26 side notched hafted bifaces was recorded from t he Early County Collection ( Table 4 2). No unifaces were present among the Early County Collection assemblage. M iller County artifacts Artifacts from Miller Co unty, Georgia, were collected by Bud Carter, a resident of Colquitt, Georgia. Carter collected artifacts from multiple private lands for over 40 years. Since the artifacts in Carters collection cannot be correlated to specific archaeological sites, they are collectively referred to as th e Miller County Collection ( Figure 4 8). A total of 497 side notched hafted bifaces was recorded from Miller County; no unifaces were identified among this sample of Early Archaic artifacts ( Table 4 2). Alabama As previousl y discussed in Chapter 3, Alabamas Coastal Plain chert sources are restricted to the southe astern corner of the state ( Figure 3 5). As with Georgia, data I collected from Alabama are not from professionally excavated artifact assemblages; collected materials provided all samples for the investigation. Three individuals provided the collections that were examined for this study, each of whom was able to assign their material to a specific county. Don Marley and Steve Lamb of Geneva, Alabama, shared their ar tifact collections from Geneva, Henry, and Houston counties. Both collections were made over the past 30 years from various areas along the surface of creek edges or in recently plowed privately owned lands. The anonymous collector referred to as X provide d artifacts that were recovered from Barbour and Houston counties. None of the artifacts from these three collectors was recovered from any specific archaeological site. Artifacts from X comprised the Barbour County Collection. Artifacts aggregated from the collections of Marley

PAGE 120

120 and Lamb were aggregated to form the Henry County Collection and the Geneva County Collection. The Houston County Collection is comprised of artifacts recovered by Marley, Lamb, and X. Refer to Figure 4 8 for the locations of these collection areas. A total of 180 side notched hafted bifaces was recorded from Alabama collection locales; the Geneva County Collection ( n=66, approximately 37% ) has the highest samp le size of these artifacts ( Table 4 2). Only one uniface, which was among the Barbour County Collection, was examined from the Alabama collection locales. Florida The relationship between collectors and professional archaeologists in Florida may be described as strained. This was not always the case, but the dissolution of the Isolated Finds policy in June 2005 has created a great deal of resentment among avocationalists, many of whom say they were complying with the states requests for information, and now feel archaeologists are no longer interested in their contributions. Fortunately, I was able to work with a number of individuals who own large collections. Jackson County artifacts Jackson County is located in the Panhandle of Florida, and is immediately south of Houston County, Alabama. Don Marley and Steve Lamb, who had a lso shared some of the artifacts examined in Alabama, provided artifacts they collected from various locations in Jackson County, Florida. None of these artifacts were recovered from any specific archaeological site within the county, and they are collecti vely referred to as the Jackson County Collection ( Figure 4 8). A total of 81 side notched hafted bifaces was recorded from the Jackson County Collection locales ( Table 4 2). No unifaces were identified among this artifact assemblage.

PAGE 121

121 Jefferson County arti facts Billy Dean allowed me to examine many artifacts at his home near Gainesville, Florida. Numerous Paleoindian and Early Archaic lithic tools are among the artifacts he has collected. His collections typically come from recently plowed fields on privat e lands. Most of the artifacts that Dean shared with me were from one location in Jefferson County, along the Aucilla River. This location is referred to as the Applewhite site. No formal state site form or accession number exists for this site. Dean also has collected numerous artifacts from throughout the county, many of which I examined. These other artifacts, which are not from the Applewhite site, are collectively referred to as the Jefferson County Collection. Figure 4 9 presents the locations of Jeff erson County and the Applewhite site. A total of 101 side notched hafted bifaces was recorded from Jefferson County sites/collection locales; most of these ar e from the Applewhite site ( Table 4 2). No unifaces were examined from Jefferson County sites/coll ection locales. T aylor County artifacts Taylor County is represented by one archaeological site, the Lake Bird site (8TA143). This site is located in the northern portion of Tay lor County, approximately 4.8 km south of the Madison County boundary, near sev eral ponds ( Figure 4 9). The Lake Bird site is well known to archaeologists and avocationalists alike. The site is also colloquially known as Bird Lake, and as the Ross Bay site, after Dennis Ross of Perry, Florida. Lake Bird has been collected since the 1980s, when development made the site accessible. Then in 1987, 1988, and 1990, a group of amateur archaeologists from New York and Florida conducted limited excavations at the site. The group recovered artifacts dating to the Early and Middle Archaic periods, including eight

PAGE 122

122 Bolens and one Edgefield Scraper. No caches or other cultural features were identified at that time (Gramly 1994). Recently, a graduate student at Florida State University analyzed 31 Bolen hafted bifaces from Lake Bird, which were found by a private collector in Tallahassee (Bissett 2003). Additionally, the Florida Bureau of Archaeological Research has several ( n =23) Bolens from the Lake Bird site in their research collection (James Dunbar, personal communication 2005). Brian Evensen of St. Petersburg, Florida, had a personal collection of recovered materials from the Lake Bird site that were available for analysis. A total of 74 sidenotched hafted bifaces was recorded from this site in Taylor County ( Table 4 2). No unifaces were examin ed from this artifact assemblage. A total of 37 of the hafted bifaces from Lake Bird was part of a pilot stu dy for quarry cluster sourcing. This study was conducted as part of this research with the generous assistance of Robert Austin of Southeastern Archaeological Research, Inc. in Riverview, Florida and the study results are presented in Chapter 5. Lafayette County artifacts The Jeanies Better Back site (8LF54) is a multicomponent Archaic site located in the northern portion of Laf ayette County near the confluence of Mill Creek and Bethel Creek, which drai ns into the Suwannee River ( Figure 4 9). In 1999, Southeastern Archaeological Research, Inc. of Gainesville conducted testing excavations within a portion of the site for a Florida Department of Transportation project (Austin and Mitchell 1999). The Early Archaic component of the examined portion of the site was interpreted by Austin and Mitchell (1999) as a base camp and habitation area. More than 4,000 artifacts recovered from thi s site were attributed to the

PAGE 123

123 Early Archaic occupation. The majority of these artifacts were debitage from tool manufacture and maintenance. Analysis of lithic biface technology by Austin and Mitchell (1999:104) identified minimal evidence for early stage biface reduction, and a high frequency of finished bifaces. This indicated that prepared preforms and finished tools likely were brought to the site from elsewhere. It may be that primary biface production occurred at procurement sites (Austin and Mitchell 2010). Austin and Mitchells (1999) analysis also provided evidence that basal notching of Early Archaic bifaces occurred prior to final thinning. This interpretation of biface production technique differs markedly from interpretations of side notched tools at Page Ladson, a submerged site in the Aucilla River. However, evidence for the bifacial production sequence at Jeanies Better Back suggests it is broadly similar to the Late Paleoindian biface manufacturing process (Austin and Mitchell 2010). Numerous hafted bifaces recovered at Jeanies Better Back were typed as Bolen; these included side notched and corner notched examples. Archaeologists observed no apparent distinctions in vertical distribution for variations in notch type Examination of beveling and edge angles for many bifaces recovered at Jeanies Better Back indicate that they were much more frequently used for cutting, sawing, scraping, and shaving tasks, than they were used as projectiles (Austin and Mitchell 2010). I had the opportunity to examine the Early Archaic tools from Southeastern Archaeological Research, Inc.s excavation at the site. For the purposes of this research, I examined only side notched examples. Twenty one hafted bifaces were recorded from the Jeanies Better Back site in Lafayette County; no unifaces were identified among the examined sample ( Table 4 2).

PAGE 124

124 Austin and Mitchell (1999) typed numerous Early Archaic biface preforms and fragments recovered at the Jeanies Better Back site as to geological formation and quarry cl uster as part of their contract with the Florida Department of Transportation. I analyzed a total of 21 of the se hafted bifaces Table 43 presents data collected by Austin and Mitchell (1999:43, 46) for quarry cluster sourcing of these artifacts I estimated data regarding distance from site using ArcGis 10TM. A majority of side notched hafted bifaces recovered at the Jeanies Better Back site that were analyzed for this study could be sourced to quarry cluster. These quarry clusters all are part of the S uwannee geological formation. As Table 4 3 shows, m ost of the identified cherts were sourced to quarry clusters located locally or relatively near to the archaeological site (Figure 4 10) This seems to suggest that groups using this site had a relatively restricted range (Austin and Mitchell 2010). However, one of these side notched hafted bifaces is manufactured from chert with inclusions typical of the Brooksville quarry c luster (Austin and Mitchell 1999). This quarry Table 4 3. Quarry cluster sourcing for side notched hafted bifa ces from 8LF54 Quarry Cluster Total artifacts Distance from site (km) # of days to source (@ 15 km/day Brooksville? 1 150 10 Marianna? 1 173 11.53 Probably Upper Suwannee 1 2 0.13 Unidentified 5 n/a n/a Upper Suwannee 8 2 0.13 Upper Suwannee/Santa Fe 1 2 or 47 0.13 or 3.13 White Springs 2 46 3.07 White Springs or Wacissa 1 46 or 38 3.07 or 2.53 White Springs? 1 46 3.07 Total 21

PAGE 125

125 cluster is located approximately 150 km south of 8LF54, and it is considered an exotic raw material in comparison to the other identified quarry clusters. The maximum distance traveled per day by hunter gatherers for resources in a variety of envir onments is estimated to be 15 km (Kelly 1995:133). Use of this formula, which does not take the possibility of watercraft into account, indicates a distance of ten days (one way) from the Jeanies Better Back site to the Brooksville quarry cluster. It is more likely that Early Archaic groups traveled b y canoe when possible, so these travel time estimates may not be particularly meaningful in this research context. While direct evidence for watercraft is lacking for the Early Archaic, adzes have been recovered in association with Late Paleoindian and Ear ly Archaic period artifacts throughout the region (Jodry 2005; Morse and Goodyear 1973). One biface recovered at the Jeanies Better Back site may be made of chert from the Marianna Quarry Cluster, l ocated some 173 km west of 8LF54 (a distance of 11.53 day s, without the use of watercraft) (after Kelly 1995:133). Interactions with groups to the west, east and south may be indicated by these data. It is also possible that these bifaces are evidence of long distance gift exchange (Austin and Mitchell 2010). Long distance exchange, and not direct procurement, may be the most likely scenario, given the great distances involved in these two examples. Santa Fe and Suwannee Rivers region artifacts I also analyzed numerous flaked stone artifacts in Alvin Hendrixs collection, which is housed at his home in McIntosh, Florida. Mr. Hendrixs artifacts were collected throughout the 1970s, from underwater contexts throughout northcentral Florida. The collection is especi ally representative of the Santa Fe and Suwannee Rivers, since many of the artifacts were collected from those contexts.

PAGE 126

126 Many of the artifacts in Alvin Hendrixs collection have been featured in publications, including Ripley Bullens (1975) guide to the interpretation of Florida flaked stone tools. Throug hout recent decades, Hendrix has shared information with numerous Southeastern archaeologists, including James Dunbar, Albert Goodyear, Andrew Hemmings, and Barbara Purdy. He also maintains a close relationship with the Silver River Museum of Ocala, to whi ch he has donated many artifacts. Hendrixs future plans are to donate most (if not all) of his collection to that museum, as he is deeply interested in keeping Floridas cultural patrimony in Florida. Alvin Hendrixs artifact collection is somewhat unusua l in my experience, in that he has kept detailed records of where each individual artifact was recovered. For instance, he has created site names such as Gator Hole, or Santa Fe Site #1. Each artifact recovered from those locations has been labeled wit h an individual number, which can be used to cross reference its site location in a single, handwritten ledger. All of the artifacts that I examined were aggregated into a sample that is referred to as the Hendrix Collection ( Figure 4 9). I recorded a tota l of 268 hafted bifaces from the Santa Fe and Suwannee River area ( Table 4 2). This collection also provided 39 unifaces that were examined for my study. Tampa Bay region artifacts I also examined Early Archaic artifacts that were recovered from four counties in the central peninsular Gulf Coast (Hernando, Hillsborough, Pasco, and Pinellas). Tampa Bay is the drowned source area of the Early Holocene Hillsborough River (after Goodyear et al. 1983). As discussed in Chapter 3, the Florida peninsula was approximately twice its current size at the onset of the Holocene (Faught 1996). Diagnostic Paleoindian bifaces have been recovered from

PAGE 127

127 dredged portions of Tampa Bay, and submerged Early Archaic sites also likely exist in the region (after Goodyear et al. 1983). As previously mentioned in Chapter 3, Upchurch et al. (1982) identified several chert quarries in the region, one of which (the Hillsborough River Chert Quarry Cluster) is loca ted in Hillsborough County ( Figures 3 3 and 34). Hernando County is the northernmost of the four counties I examined in this region. The Spring Hill site is in southern Hernando County, just north of the Pasco County boundary, approximately one mile east of Weeki Wachee Springs State Park (Figure 4 11). During the construction of several ponds for a neighborhood in Spring Hill, Jerry Spencer of St. Petersburg collected artifacts from the exposed surface. As no formal Florida Site Form has been completed for t his site, no state accession site number has been assigned. Artifacts from Spencers collection serve as the data set for this site. Pasco County is located just south of Hernando County in the central peninsular Gulf Coast. The artifacts that I examined f rom this county were also from Spencers collection. None of the artifacts within this county could be assigned to a specific archaeological site; hence they are all collectively referred to as t he Pasco County Collection ( Figure 4 11). Within Hillsborough County, I examined artifact s recovered from one archaeological site (the Harney Flats site, 8JE507) and one collection locality ( the Harris Grove Collection) The Harris Grove collection locality is located both inside and adjacent to the southeastern boundary of the Hillsborough River State Park, in northern Hillsborough County ( Figure 4 11). The area, comprising a three mile radius, is situated along a ridge that overlooks Flint Creek. Artifacts exam ined for my study from the Harris Grove Collection were collected by both Brian Evensen and Spencer. The Harney Flats s ite (8JE507) is located approximately 8.0 km so uthwest of the

PAGE 128

128 Harris Grove collection locality in northern Hillsborough County. Exc avations at the site recovered side notched bifaces in association with Paleoindian artifacts The Harney Flats si te is thought to reflect long term and diverse exploitation of riverine and marsh resources (Daniel and Wisenbaker 1987). Spencer also provide d artifacts analyzed for this investigation that he collected from the Harney Flats site. Pinellas County is located west of Hillsborough County in the central Gulf Coast area of the Florida peninsula. The Kepler site (8PI862) is the only formal archaeolog ical site with Early Archaic artifacts that I examined in this county ( Figure 4 11). Jim Drawdy collected artifacts from this site from a field adjacent to the Intracoastal Waterway at Indian Rocks. Two collection locales that were examined for this resear ch were also from Pinellas County. Evensen collected the artifacts that were examined from the St. Petersburg Area Collection. Both Evensen and Spencer provided artifacts that belong to the Pinellas County Collection. None of the artifacts from the St. Pet ersburg Area Collection or the Pinellas County Collection could be assigned to a specific archaeological site. Seventy seven side notched hafted bifaces were recorded from the Tampa Bay region, the majority of which (approximately 57% ) were from the Harri s Grove site ( Table 4 2). Fourteen unifaces, half of which are from the Harris Grove site, were examined from the Tampa Bay region. Numerous hafted bifaces from the Tampa Bay region were part of a pilot study for quarry cluster sourcing. This study was conducted as part of this research with the generous assistance of Robert Austin of Southeastern Archaeological Research, Inc. in Riverview, Florida and the study results are presented in Chapter 5.

PAGE 129

129 Mapping Artifacts to Drainage and Watershed I aggregated al l data collected from South Carolina, Georgia, Florida, and Alabama according to watershed and closest associated major drainage system. Figure 412 displays the watersheds and Coastal Plain drainages examined for this dissertation. As discussed in Chapter s 2 and 3, river drainages and their corresponding watersheds would have been obvious natural and physiographic boundaries. Numerous archaeological and ethnographic examples suggest these features also may have demarcated social group boundaries during the Early Archaic period (after Anderson and Hanson 1988; Blanton and Snow 1989; Fullagar and Head 1999; Miller 2011; Thulman 2006; Williams 1982). As such, the river drainage and the watershed are meaningful analytical units with a history of application to the PleistoceneHolocene transition. A total of seven drainage systems is represented by the data I collected. Three of these are part of the Atlantic watershed; the remaining four are part of the Gulf watershed. Drainage Systems of the Atlantic Watershed Three drainage systems of the Atlantic watershed are represented here; the SanteeCooper, the Savannah Ogeechee, and the Ocmulgee ( Figure 4 12). Brief descriptions of these drainage systems are provided below. The easternmost drainage system represented is the SanteeCooper. This drainage system begins at the confluence of the Congaree and Wateree Rivers, which drain into the drowned Santee River (today, the Santee Lakes, Moultrie and Marion). Artifact collections I examined from the SanteeCooper drainage system include: the Elloree Museum Collection, the Elloree Bank Collection, the Skinner Collection, and the Hebert Collection. A total of 66 hafted bifaces

PAGE 130

130 and four unifaces was examined from the Santee Cooper drainage system. The Santee Cooper drainage sy stem is comprised o f approximately 13,741 km ( Figure 4 3). The second drainage system examined the Savannah Ogeechee, is comprised of two river drainages. Justification for considering these drainages together as a single analytical unit is based on the location of Burke County, Georgia within both the Savannah and Ogeechee River basins (SavannahUpper Ogeechee Water Planning Council 2011:22). As previously discussed, I examined 194 side notched artifacts from Burke County. The SavannahOgeechee drainage system begins in the Coastal Plain of southwestern South Carolina and extends west of Burke County, Georgia, encompassing approximatel y 18,227 km ( Figure 4 3). This drainage system contains a number of recorded Early Holocene sites, and has been the focu s of a great deal of professional interest in ancient lifeways (as in Anderson and Hanson 1988). Collections examined from this drainage system include those from the Big Pine Tree site (38AL143), the Charles site (38AL135), and the Topper site (38AL23) in Allendale County, South Carolina, and the Owl and Cotton Creek sites in Burke County, Georgia, as well as more generalized collections made from throughout Allendale (Strong Collection) and Burke counties (Burke County Collection). I examined 603 hafted bifaces and 188 unifaces that were recovered from the Savannah Ogeechee drainage system. The third drainage system, and the final one in the Atlantic watershed, is the Ocmulgee. This system begins in north central Georgia, and flows southeast to its confluence with the Oconee River, which ultimately drains into the Altamaha River. The Ocmulgee River drainage system is comprised of approximately 10,799 km ( Figure 4 3). Artifacts I examined from the Ocmulgee drainage system include collections from the whole of Coffee County, Georgia, as

PAGE 131

131 well as material from the Feronia Site Complex in that same county. The Feronia Site Complex lies at the interface of the Gulf and Atlantic watersheds, which may have made the site attractive to Native Americans in search of a place for social aggregation (Blanton and Snow 1986, 1989). I analyzed a total of 74 hafted bifaces and nine unifaces from the Ocmulgee drainage system. Drainage Systems of the Gulf Watershed Four drainage systems of the Gulf watershed are represented in this research; the Flint, the Chattahoochee, the Aucil la Suwannee, and Tampa Bay ( Figure 4 12). I provide brief descriptions of these drainage systems below. The Flint River begins in western Georgia, and continues to the southwest until its confluence wi th the Chattahoochee in Lake Seminole, just north of the Florida panhandle. This drainage system also is the location of a recent statefunded effort to inventory Paleoindian and Early Archaic artifacts (David Crass, personal communication 2005). The Flint drainage system encompasses approximately 11,499 km ( Figure 4 3). Artifacts I examined from the Flint drainage system include the Dooly County Collection and the Worth County Collection. Ninety seven hafted bifaces from the Flint drainage system were analyzed for this investigation; no unifaces were identified among the examined assemblages. The Chattahoochee drainage system begins in northwestern Georgia, where it flows south along the states border with Alabama. The Chattahoochee drainage continues int o north Florida, until its confluence with the Flint River in Lake Seminole. The Chattahoochee drainage system is comprised of approximately 32,020 km ( Figure 4 3). I examined numerous personal collections from the Chattahoochee drainage, including those made in Barbour, Geneva, Henry, and Houston counties in Alabama, Early and Miller counties in Georgia, and Jackson County,

PAGE 132

132 Florida. A total of 384 hafted bifaces and one uniface was recorded from the Chattahoochee drainage system. The Aucilla Suwannee drai nage system begins in northwestern Florida, at the Aucilla River, and continues to the southeast to include much of the Santa Fe and Suwannee Rivers. The Aucilla Suwannee drainage system encompasses approximately 29,978 km ( Figure 4 3). These areas have b een of great interest to a number of archaeologists, as well as collectors, in recent years (Austin and Mitchell 1999; Carter 2003; Carter and Dunbar 2006; Faught 2004b; Upchurch et al.1982). Artifacts I examined from the Aucilla Suwannee drainage include those from the Applewhite site and the Jefferson County Collection in Jefferson County, the Lake Bird site (8TA143) in Taylor County, the Jeanies Better Back site (8LF54) in Lafayette County, and the Hendrix Collection from the Santa Fe and Suwannee River s. I examined 464 hafted bifaces and 39 unifaces from the Aucilla Suwannee drainage system. The final and southernmost portion of my research area is Tampa Bay, which cannot properly be described as a drainage. The geological history of this region lends s upport to the notion of Tampa Bay as a circular basin during the Early Holocene. As such, it is an analytical unit in the same way that river drainages are for other portions of this study. The Hillsborough River is the largest freshwater source for Tampa Bay. Tampa Bay is comprised of approximately 11,583 km ( Figure 4 3). Artifacts I examined from Tampa Bay include those collected from the Harney Flats site and the Harris Grove Collection in Hillsborough County, the Kepler site (8PI862) in Pinellas County and the Spring Hill site in Hernando County, as well as from general county wide collections made in both Pasco and Pinellas counties. I analyzed a total of 77 hafted bifaces and 14 unifaces recovered from the Tampa Bay area.

PAGE 133

133 Analyses of Patterned Variat ion Having introduced the samples of Early Side Notched Horizon artifacts examined throughout the lower Southeast, as well as the attributes selected for analysis, I am obliged to describe how I analyzed variation in these assemblages and how patterns were sought As discussed in Chapter 3, no settlement model has been developed for the earliest portion of the Early Archaic period. While there have been some recent attempts to model antecedent Paleoindian regionalization and social org anization (Anderson 1990, 1996; Miller 2011; Smallwood 2012; Thulman 2006), the social boundaries potentially reflected by subregional traditions within the Early Side Notched Horizon remain underexplored. As described above, I first considered patterned v ariation for hafted biface s, plotting nominal and ratioscale attributes for all examined artifacts at the level of assemblage, and then aggregating these data to drainage and watershed for comparisons across space. Subsequently, I performed the same level of analysis for Edgefield Scrapers, although only five drainage systems produced those tool forms, while hafted bifaces from seven drainages were examined. Although I am interested in exploring social boundaries and boundary processes in the Early Archaic, I would like to explicitly avoid the classification of these groups as bounded entities (after Jones 1997). Nonetheless, as addressed in Chapter 3 with respect to Early Side Notched Horizon artifact types these investigations necessarily begin with the presumption of a Taylor Zone (after Michie 1966), a Bolen Zone ( after Bullen 1958, 1975), and a Big Sandy Zone ( after Kneberg 1956) In order to test the hypothesis that these zones are geographically distinguishable ( relative to measurable shifts in the patterned variation of associated side notched

PAGE 134

134 tool forms ) I designated approximate central points ( sensu core areas) within zones, as well as boundary areas at zonal intersections The approximate center of the Taylor Zone is Allendale County; the Strong Collection ( n= 411 hafted bifaces) assemblage is from the core of the Taylor Zone. The approximate center of the Bolen Zone is the Aucilla Suwannee Ri ver drainage; the Hendrix Collection ( n=268 hafted bifaces) assembla ge is from the core of the Bolen Zone. Unfortunately, large frequencies of hafted bifaces were not available from the center of the Big Sandy Zone. The Feronia Site Complex (in Coffee County, Georgia) in the Ocmulgee River drainage is situated at the inter section of the Taylor, Bolen, and Big Sandy Zones Therefore, I assigned assemblages from Feronia ( n=37 hafted bifaces) and Coffee County ( n=37) spec ial status due to their provenance at a boundary area. Next, I applied various statistical tests (most effi cacious of which was coefficient of variation) to compare the range of patterned variation in core areas (i.e., in the Strong and Hendrix Collections), to that of the boundary area. I expected that the widest range of va riation for nominal and ratio scale artifact attributes would be evidenced at the zonal boundary of the Ocmulgee River, given the likelihood that while social boundaries surely existed during the Early Archaic, those boundaries were permeable and group interactions were common. Summary This chapter presented the research methods and data sets used throughout this study. First, I related the ways in which I recorded numerous artifact attributes, and I defined the attributes selected for the analyses of side notched hafted bifaces and unif aces throughout the Coastal Plain. These data were introduced in the context of state as well as more specific provenance. Next, I explained how I assigned these data to specific drainage systems within the

PAGE 135

135 Atlantic and Gulf watersheds in preparation for r egional analyses. In all, some 2,020 side notched hafted bifaces and unifaces from a total of seven Coastal Plain drainages were analyzed for this research, the results of which are presented in Chapter 5. This chapter concluded with discussion of the methods applied in pattern recognition within the Early Side Notched Horizon, toward the identification of social boundaries as manifested in apparent subregional technologic al traditions.

PAGE 136

136 Figure 4 1. Measurements taken for side notched hafted bifaces

PAGE 137

137 F igure 4 2. Measurements taken for side notched hafted unifaces

PAGE 138

138 Figure 4 3. Research area in the Coastal Plain Southeast

PAGE 139

139 Figure 4 4. Locations of Lakes Marion and Moultrie, and Allendale County, South Carolina; and Burke County, Georgia sites and collection areas

PAGE 140

140 Figure 4 5. Locations of Lakes Marion and Moultrie, and Allendale County, South Carolina; and Burke County, Georgia sites and collection areas, in the context of Coastal Plain cherts

PAGE 141

141 Figure 4 6. Locations of Coffee, Dooly, and Worth counties, Georgia sites and collection areas

PAGE 142

142 Figure 4 7. Location of Feronia sites in the context of Coastal Plain chert distribution ( adapted from Goad, Sharon I. 1979. Chert Resources in Georgia: Archaeological and Geological Perspectives. Wa llace Reservoir Project Contribution 3. University of Georgia Laboratory of Archaeology Series Report Number 21, Athens. [Page 6, Figure 1 and Page 9, Figure 3] )

PAGE 143

143 Figure 4 8. Locations of Early and Miller counties, Georgia; Barbour, Geneva, Henry, and Houston counties, Alabama; and Jackson County, Florida, sites and collection areas

PAGE 144

144 Figure 4 9. Locations of Jefferson, Lafayette, and Taylor counties, Florida sites and collection areas

PAGE 145

145 Figure 4 10. Locations of Florida archaeological sites in the context of chert quarry clusters ( adapted from Upchurch, Sam B., Richard N. Strohm, and Mark G. Nuckels 1982. Methods of Provenance Determination of Florida Cherts. Report prepared for the Florida Division of Archives, History, and Records Management by the Department of Geology, University of South Florida, Tampa. [Page 14] )

PAGE 146

146 Figure 4 11. Locations of Hernando, Hillsborough, Pasco, and Pinellas counties, Florida sites and collection areas

PAGE 147

147 Figure 4 12. The Atlantic/Gulf watershed boundary in the context of Coastal Plain drainages

PAGE 148

148 CHAPTER 5 RESULTS The proximate goal of this research is to document and explain patterned variation in Early Side Notched Horizon material culture. My ultimate goal is to understand the genesis and reproduction of social boundaries throughout the lower Southeast during the earliest portion of the Early Archaic period. I considered Early Archaic hafted bifaces and unifaces that were collected from seven different river dr ainages in the South Atlantic Slope and Eastern Gulf Coast. Methods for these analyses were discussed in Chapter 4. In this chapter, I relate the results of detailed analyses of some 2,020 side notched artifacts, discussing side notched hafted bifaces and unifaces across drainages, and then comparing those artifact classes throughout the Coastal Plain. Subsequently, I discuss zones of patterned variation across drainages, and compare the centers of the Taylor and Bolen Zones with the inferred boundary area where those zones intersect. Finally, I present the results of a pilot study for raw material sourcing of Florida side notched hafted bifaces. Side N otched Hafted Bifaces across Drainages The hafted biface data set consists of 1,765 Early Archaic specimen s. The results of analy ses of both nominal and ratioscale attributes are presented below, with respect to their associated drainages. Then, comparisons of patterned variation among the data collected from all river drainages are made. Statistical analyses of select attributes that highlight broad similarities and differences among these data sets are also presented. Mapping Nominal Data I documented ten different nominal attributes among the hafted bifaces. As discussed in Chapter 4, I restricted my research to Coastal Plain chert artifacts to allow for some level of

PAGE 149

149 control of the sample set. Therefore, raw material selection is consiste nt throughout the data and is not considered a variable. Among the ten nominal attributes, eight (condition, manufacturing condition, fracture type, basal grinding, cross section, beveling, lateral condition, and notch type) are artifacts of attrition, use and design. With regard to notch type, only side notched hafted bifaces were examined; however, a notch type described as sloping sidenotched may be seen in assemblages where reworking of the haft element changed the shape and angle of the notch area. T hese trends are often reflected in the auricle height, which was seen to decrease in value as a result of changes towards a more sloping notch type throughout the life of the tool. While all technological choices are situated in social contexts (after Lemo nnier 1992), multiple and overlapping decisions that occur while a tool is manufactured, used, lost, and discarded are often not knowable due to the reductive nature of lithic technology. Trends observed in basal configuration are discussed below, with res pect to each of the seven drainages in the study area. Santee Cooper hafted bifaces Sixty four of the 66 hafted bifaces examined from the Santee Cooper River drainage have a recognizable basal configuration (Appendix A). The majority of these hafted biface s ( n=46, approximately 72% ) have an incurvate base. Straight bases were the next most frequent ( n=17), consisting of approximately 27 % of the sample. Only one excurvate base (approximately 1% ), was observed. Savannah Ogeechee hafted bifaces I recorded 593 hafted bifaces that have a recognizable basal configuration from the Savannah Ogeechee River drainage (Appendix A). Ten hafted bifaces from the river drainage did not have a recognizable base shape. Incurvate bases were the most frequent among the

PAGE 150

150 sample ( n=393, approximately 66% ). The second most frequent base shape was straight ( n =149, approximately 25% ). Excurvate bases were minimally represented ( n=51, approximately 9% ) among the Savannah Ogeechee hafted bifaces. Ocmulgee hafted bifaces Within the Ocmulgee River drainage, 66 of the 74 hafted bifaces have an identifiable base shape (Appendix A). Straight is the most frequent basal configuration ( n=31, approximately 47% ) among this sample. The second most frequent basal configuration is incurvate ( n=29, ap proximately 44% ). Excurvate bases were the least frequent base shape among the Ocmulgee sample; only six (approximately 9% ) were observed. Flint hafted bifaces A total of 94 out of 97 hafted bifaces recovered from the Flint River drainage had observable ba se shapes (Appendix A). The most common base shape was straight, comprising 65% ( n=61) of the sample. Incurvate bases were the second most frequent type observed in this drainage ( n=31, approximately 33% ). Only 2% of the sample from the Flint River drainag e ( n=2) had excurvate bases. Chattahoochee hafted bifaces I recorded 375 hafted bifaces that have a recognizable basal configuration from the Chattahoochee River drainage (Appendix A). Nine hafted bifaces from the river drainage did not have a recognizable base shape. Straight bases were the most frequent among the sample ( n=206, approximately 55% ). The second most frequent base shape was incurvate ( n=155, approximately 41% ) Excurvate bases were minimally represented ( n=14, approximately 4% ) among the Chattahoochee hafted bifaces.

PAGE 151

151 Aucilla Suwannee hafted bifaces A total of 454 of the 464 hafted bifaces I analyzed from the Aucilla Suwannee River drainage have a recognizable basal configuration (Appendix A). The majority of these hafted bifaces ( n =217, approxim ately 48% ) have a straight base. Excurvate bases were the second most frequent ( n =154), consisting of approximately 34% of the sample. Incurvate bases were the least represented among the AucillaSuwannee sample; 83 bases (approximately 18 % ) were inc urvate. Tampa Bay hafted bifaces Within the Tampa Bay area, all 77 of the hafted bifaces from the region have an identifiable base shape (Appendix A). Straight bases ( n=36, approximately 47% ) are found most frequently in this sample. The second most freque nt basal configuration is excurvate ( n=27, approximately 35% ). Incurvate bases were the least frequent base shape among the Tampa Bay area sample, w ith 14 (approximately 18% ) such examples observed. Comparisons across Drainages A comparison of the data col lected for basal configuration may inform inquiries into Early Archaic social boundaries. Here, I compare these attributes between the seven river drainages in an attempt to determine if any trends can be identified spatially. In an effort to map all patterning in basal configuration for side notched hafted bifaces across the study area, I used data which had been aggregated to Southeastern drainage (as detailed above ), and recorded the apparent trends in each area. Table 5 1 sh ows the patterns that emerged.

PAGE 152

152 Table 5 1. Absolute and relative frequencies of h a fted bifaces by drainage and basal configuration Drainage Incurvate Straight Excurvate Total n % n % n % Santee Cooper 46 71.88 17 26.56 1 1 .56 64 Savannah Ogeechee 393 66 .27 149 25 .13 51 8.6 593 Ocmulgee 29 43.94 31 46.97 6 9 .09 66 Flint 31 32.98 61 64.89 2 2 .13 94 Chattahoochee 155 41 .33 206 54.93 14 3.73 375 Aucilla Suwannee 83 18 .28 217 47.8 154 33.92 454 Tampa Bay 14 18 .18 36 46.75 27 35 .06 77 Total 751 717 255 1,723 Figure 5 1 displays the information from Table 5 1 in terms of percentage. Figure 52 presents the same information, and also allows for the graphic representation of differentiation in basal configuration for hafted bifaces throughout the Coastal Plain. I n the SanteeCooper and Savannah Ogeechee Ri ver drainages, there is a strong tendency for sidenotched hafted bifaces to be made with incurvate bases. Specifically, approximately 72 % (46 out of 64) and 66% (393 out of 593) of those samples have incurvate b ases, respectively. This is in accordance with Michies (1966) type description of Taylors, which generally are thought to be distributed throughout southwest ern South Carolina and eastern Georgia (after Sassaman 1996). This pattern changes in the Ocmulgee River drainage, where I observed a shift in basal configuration. Early Archaic visitors to Coffee County, Georgia, which includes the Feronia sites, may have had no strong preference for one kind of basal configuration; s traight ( approximately 47% ) and in curvate ( approximately 44% ) bases occur almost equally in these samples. Alternatively, variation in basal configuration in the Ocmulgee samples may be interpreted as an amalgam of multiple base shape traditions from throughout the region. The Feronia site s are located at the intersection of the Taylor, Bolen, and Big Sandy Zones, and these

PAGE 153

153 sites arguably reflect social group aggregation. This possibility is discussed at length later in this chapter, as well as in Chapter 6. This apparent geographic shift f rom the production of mostly incurvate based tools can be seen to continue to the south and west of the Ocmulgee River drainage ( Figure 5 2). This adds some credence to the suggestion that the Feronia sites may have ser ved as a place of aggregation. This i dea first was offered to explain the location of these sites, which are located some 83 km away from sources of highquality raw material. Additionally, these sites are located at the interface of the Gulf and Atlantic watersheds, boundaries proposed to explain Early Archaic settlement patterns in this region (Anderson and Hans on 1988; Blanton and Snow 1986, 1989) ( Figure 4 12) To the west in the Flint River drainage, I examine d samples from Dooly and Worth c ounties, Georgia ( n=94). At the Flint River, I observed a s hift toward increased frequencies of straight bases. Straight bases comprise approximately 65% of this sample, while incurvate examples account for only about 33 % of the sample. This trend continues in the Chattahoochee River drainage and sout h into Florida ( Figure 5 2). In the Aucilla Suwannee and Tampa Bay regions, I observed another shift in basal conf iguration. The trend for straight bases continues here, and now excurvate bases occur with the second highest frequencies ( Figure 5 2). Straig ht bases in the Aucilla Suwannee River drainage comprise some 48% of the sample, and approximately 34 % of the side notched tools in t his area have excurvate bases. In Tampa Bay, approximately 47 % of the side notched hafted bifaces have straight bases, and some 35% have excurvate bases. Clearly, incurvate bases are significantly less common in Florida than they are elsewhere, especially when compared to the

PAGE 154

154 SanteeCooper and SavannahOgeechee River drainages at the northern extent of the study area where th ey were predominant. Notably, Bullens (1975) type descriptions for Bolens (defined as distinctive Florida projectile points) include a subtype known as Subtype 1/Concave based. This subtype corresponds to incurvate basal configuration, as applied in this study. Bullen (1975) included this subtype for two distinct Bolen types, Bolen Plain and Bolen Beveled. All other subtypes of Bolens defined by Bullen (1975) are described as having straight or convex bases. Convex bases correspond to excurvate bases, as a pplied in this research. Out of a total of ten subtypes defined for Bolen Plain and Bolen Beveled, only two were defined as Concave based (Bullen 1975). It is interesting to note that trends for basal configuration of a majority of Florida provenienced haf ted bifaces examined for this analysis, correspond well to the broad definitions of the Bolen artifact type provided over three decades ago (after Bullen 1975). Another recent analysis of 200 side notched hafted bifaces from the Aucilla Suwannee River drai nage found that approximately 77% of all examined samples had either straight or excurvate bases (Bissett 2003:64). Interestingly, in that study, the majority of bifaces ( n=83, or approximately 42% ) had excurvate bases; the next highest proportion of bifac e bases was straight ( n =69, or approximately 35% ). In this study of 454 hafted bifaces from the Aucilla Suwannee River drainage, some 82% of the examined tools were either straight or excurvate. However, as discussed above, the higher proportion of bifaces ( approximately 48% ) had straight bases. Future studies of side notched hafted bifaces throughout the region should continue to consider basal configuration as an important area of analysis, since geographic patterning of this attribute has long been poorly understood.

PAGE 155

155 I performed a chi square test in order to test the significance of these apparent trends in basal configuration throughout the Coastal Plain, creating contingency tables that compared the basal configuration of hafted bifaces to specific river drainage location. Table 52 shows the observed number of hafted bifaces of different basal configuration from the seven examined river drainages; Table 5 3 shows the expected number of hafted bifaces of different basal configuration from these regi ons. Table 5 2. Hafted biface basal configuration (Observed) Drainage Incurvate Straight Excurvate Total Santee Cooper 46 17 1 64 Savannah Ogeechee 393 149 51 593 Ocmulgee 29 31 6 66 Flint 31 61 2 94 Chattahoochee 155 206 14 375 Aucilla Suwannee 83 217 154 454 Tampa Bay 14 36 27 77 Total 751 717 255 1,723 Table 5 3. Hafted biface basal configuration (Expected) Drainage Incurvate Straight Excurvate Total Santee Cooper 27.904 26.624 9.472 64 Savannah Ogeechee 258.548 246.688 87.764 593 Ocmulgee 28.776 27.456 9.768 66 Flint 40.984 39.104 13.912 94 Chattahoochee 163.5 156 55.5 375 Aucilla Suwannee 197.944 188.864 67.192 454 Tampa Bay 33.572 32.032 11.396 77 Total 751.228 716.768 255.004 1,723 Th e chi square value that resulted is 348.5599, with 12 degrees of freedom and a probability of <.0001. This probability indicates that type of basal configuration and drainage location are not independent variables. In other words, the base shape of hafted bifaces appears

PAGE 156

156 to be directly correlated to geography within the South Atlantic Slope and Eastern Gulf Coast during the Early Ar chaic. Mapping RatioS cale Data As discussed in Chapter 4, numerous ratio scale attributes were collected among the Early Archaic hafted bifaces used in this research. It is vital to distinguish between the blade portions of these tools, versus their haft elements. Blade length and shoulder width are attributes known to change during the life of a tool, as it is used and resharpened. Maximum thic kness is another attribute of the blade portion of a biface. This attribute is less likely to change during tool use life, as further biface thinning is unlikely to occur beyond initial manufacture. Based on the results of previous research on variation in Early Holocene bifaces, I expected that the aspect of side notched hafted bifaces least subject to use alteration, and most affected by initial design, would be the haft element ( Austin and Mitchell 2010; Bissett 2003; Goodyear 1995 [1974] ; Keeley 1982). The metric variables I studied relating to haft element in clude: shoulder length, between notch width, notch height, and base width. These measurements were taken for all bifaces and are presented in Appendix B My interpretations of data from the seven Coastal Plain drainages are highly influenced by knowledge of local raw material availability (cf. Bamforth 1986, Goodyear 1979; Kelly 1988). Five of the seven drainages are located in close proximity to Coastal Plain cherts (Figure 53). The Santee Cooper and Ocmulgee R iver drainages are located away fr om Coastal Plain chert sources. As such, I anticipated that side notched hafted bifaces from these chert poor drainages would be characterized by smaller blades as well as haft elements and expected to see general evidence of attrition as a result of distance from raw material source. While the

PAGE 157

157 expectation of smaller blades as a function of distance from source is somewhat intuitive, haft elements generally were secured in hafts, and their sizes typically ar e not thought to have changed much during tool use life. Therefore, the expectation of smaller haft elements for assemblages from chert poor locales must be justified. Toolmakers likely anticipated raw material shortages whe n traveling to areas where highquality cherts were lacking. As such, they may have used bifaces almost to the point of exhaustion, resharpening and maintaining tools at both the distal and proximal ends. For instance, a biface broken by haft snap may have been refreshed by discarding th e initial haft element, and reworking the blade portion into a completely new (and significantly smaller) biface. Obviously, this scenario is speculative, but this type of ad hoc strategy ma y have been commonplace if highquality cherts were locally absent In general, I expected that side notched hafted bifaces from the SanteeCooper and Ocmulgee River drainages would show some indication of curation or raw material maintenance, prim arily due to distance from high quality chert sources. Meanwhile, I predicted that most metric attributes of side notched artifacts from the remaining five drainages, including the SavannahOgeechee, Flint, Chattahoochee, AucillaSuwannee, and Tampa Bay, would exhibit similar patterned metric variation with respect to one anothe r. Since raw material was readily available in those five drainages, I expected that there would be a similarly high degree of variation in metric attributes for those suites of hafted bifaces. As I assumed that fresh raw material could easily be obtained in these drainages, such that new hafted bifaces c ould be manufactured, I did not expect to see evidence for intensive curation of side notched bifaces from these drainages. Of course, depositional patterns only reflect where these tools ended up and were collected; they may have been transported to and

PAGE 158

158 from areas far from chert sources, and some level of curation may have been common practice for all of these highly mobile groups. For instance, groups may have traveled to the Fall Line in South Carolina or Georgia with a limited Coastal Plain chert toolkit, conserving and maintaining tools as they traveled and then discarding them back in the Coastal Plain. One final expectation of the data was that local proximity to high quality raw material, such as we s ee for the Savannah Ogeechee R iver drainage, may have allowed for the emergence of local cultural traditions with respect to hafted bifaces, especially relating to the adoption of some standardization of the haft element. Data rel ating to the mapping of ratio scale data for the seven drainages where sidenotched hafted bifaces were studied are presented below. Descriptive statistics (including n, mean, min, max, s.d., and c.v.) and inferential statistics (including Levenes test for equalit y of variances) for all ratio scale data also are provided below, with respect to specific discussions of the blade portions and haft elements of hafted bifaces. Appendix C presents detailed data regarding the range of variation observed for all metric attributes at the individual river drainage level of analysis. Santee Cooper hafted bifaces The greatest values for skewness in the Santee Cooper Ri ver drainage are for blade length (1.25) and shoulder length (.94). Blade length and notch height have the highest kurtosis va lues (1.68 and 1.3, respectively). The kurtosis value for shoulder length is 1.25. These data suggest that a greater percentage of hafted bifaces in the San teeCooper Ri ver drainage have blade and haft leng ths nearer the minimum values. Although blade leng th, shoulder length, and notch height are also the least varied among the metric attributes for this sample, they still have a

PAGE 159

159 platykurtic distribution, along with all of the attr ibutes examined in the sample. The distribution of the data for all of the ob served at tributes for the Santee Cooper R iver drainage have flattened peaks that result from the data being less concentrated around the mean due to large variati on within observed attributes. These data may indicate a t rend toward tool conservation. This would be expected, since this locale is not located near Coastal Plain chert raw material; the nearest sources are located approximately 97 km to the west near the Savannah River ( Figure 5 3). Savannah Ogeechee hafted bifaces The highest skewness v alues in the Savannah Ogeechee R iver drainage were found for shoulder length (1.48) and notch height (1.34). A greater percentage of hafted bifaces in this drainage have haft element attributes nearer the minimum values for those two variables. Kurtosis values (9.25 and 4.14, respectively) for these two attributes are also the highest amon g the Savannah Ogeechee sample. These kurtosis values show a leptokurtic distribution, indicating that the data are highly concentrated around the mean due to low variations for t hese attributes in this sample. These attributes are the least varied in distribution among metric attributes, perhaps suggesting a standardized haft length and notch height for tool makers in that area. Ocmulgee hafted bifaces Measured sh oulder width (0.83) and between notch width (0.84) have the highest skewness among the hafted bifaces examined from the Ocmulgee River drainage. The highest kurtosis values were found for maximum thickness (2.0) and shoulder width (0.43). The most negative kurtosis val ues were observed for blade length and base width ( 0.63 and 0.56). This suggests a greater percentage of hafted bifaces from the Ocmulgee River drainage h ave shoulder widths and between notch widths near the minimum values. Also, maximum thickness and

PAGE 160

160 s houlder width are the least varied in distribution among the ratio scale attributes, while blade length and base width are the most varied in distribution. The higher skewness values for shoulder width and betweennotch width may suggest a trend toward tool conservation. The negative kurtosis value for blade length also likely reflects tool conservation since the sample appears to contain many different hafted bifaces that were in various stages of resharpening and reuse, accounting for the flatter distribu tion of blade lengths. This would be expected given that the nearest source of high quality Coastal Plain chert is located some 83 k m to the northwe st (after Blanton and Snow 1986). Flint hafted bifaces Among the sample of hafted bifaces examined from the Flint River drainage, shoulder width (0.97) and notch height (0.92) have the highest skewness. This suggests that a greater percentage of hafted bifaces here have shoulder widths and notch heights nearer the minimum values. Shoulder length (1.37) and notch height (1.23) have the greatest kurtosis values. Shoulder length and notch height are the least varied in distri bution among metric attributes. Taken together, these data suggest a somewhat standardized haft element compared to the blade element for bifaces in this river drainage. Chattahoochee hafted bifaces The highest skewness values for the analyzed sample from the Chattahoochee River drainage are for between notch width (1.32) and blade length (0.87). The highest kurtosis values are for between notch width (2.61) and shoulder width (1.4). It would appear that a greater percentage of h afted bifaces here have between notch widths and blade lengths nearer the m inimum values, and that betweennotch width and shoulder width are the least varied in

PAGE 161

161 distribution among ratio scale attributes. This may reflect a trend towards raw material conservation and tool curation. This is somewhat counterintuitive, as the Chattahoochee River drainage is an area known to be replete with hig hquality Coastal Plain chert. Aucilla Suwannee hafted bifaces Blade length and notch height had the highest values for skewness (1.28 and 1.05, respectively) among the sample of hafted bifaces from the AucillaS uwannee River drainage. B etween notch width had a relatively high skewness (1.01) which is just below the skewness value for notch height (1.05). These three attributes also have the highest kurtosis values (blade length =4.72, notch height = 1.31, and betweennotch width = 1.64). T his suggests that a higher percentage of hafted bifaces from the AucillaSuwannee River drainage have a blade length, notch height, and betweennotch width near the minimum values. These three attributes are also the least varied in distribution among the attributes recorded for this sample. These data indicate a leptokurtic distribution for blade length values since data are highly concentrated around the mean (40.71 mm) due to less variation in the sample. Coastal Plain chert was locally available throughout this river drainage; numerous chert quarry clusters have been identified in this immediate area (refer to Figure 4 10). Some haft stability is indicated in these da ta for notch height and between notch width. It is likely that many of the specimens exa mined in the Aucilla Suwannee River drainage represent tools that were never reworked and resharpened. T he abundance of nearby and highquality raw material would not have necessitated curation of hafted bifaces in this area. It is important to note that the majority of hafted bifaces examined from the Aucilla Suwannee River drainage were collected from submerged contexts. This is a different taphonomy than seen elsewhere in the study area, where artifacts were collected in

PAGE 162

162 terrestrial contexts. Two of the haft ed bifaces from this sample had possibly hypertrophic blade lengths of 108 mm and 114 mm, which suggests that these were not necessarily functional tools but were perhaps ceremonial or symbolic (after Morse 1997a) (Figure 54). The ritual significance of water to Early Archaic groups has been well documented, as in the case of underwater burials at the Windover site (Tomczak and Powell 2003). Perhaps the oversized bifaces recovered from the AucillaSuwannee had been ritually deposited. Tampa Bay hafted bifaces Among the analyzed sample from Tampa Bay, shoulder length (1.29) and notch height (1.08) have the highest skewness values. These variables also have the highest kurtosis values (3.26 and 3.3, respectively). A greater percentage of hafted bifaces h ere, with regard to shoulder length and notch height, are nearer the minimum values for those attributes. They are also the least varied in their distribution among recorded metric attributes, such that they have leptokurtic distributions and are highly concentrated near the mean due to less variation. These data indicate some standardization of the haft element of bifaces in the Tampa Bay region. Simila r patterning was observed with respect to the SavannahOgeechee River drainage. As is true in the Savanna hOgeechee River drainage, the T ampa Bay region is home to highquality Coastal Plain chert; three high quality chert sources h ave been identified within the immediate area ( Figure 4 10) Ready access to hig h quality raw material may have afforded tool mak ers the luxury of implementing a somewhat standardized design in the manufacture of bifaces. Discussion for Blade Portion of Hafted Bifaces Here, I provide some discussion regarding patterned variation in blade length, shoulder width, and thickness. These continuous variables all reflect aspects of the blade portion of hafted

PAGE 163

163 bifaces, and relevant data are available for all seven river drainages. As a result of raw material availability and variability, and tool use an d resharpening, hafted biface b lade lengths exhibit a high degree of variation across the Southeastern Coastal Plain. Table 54 present s comparative data for blade lengths in each of the seven river drainages that produced bifaces for analysis. Only comple te blades are included in these data; totals for each river drainage reflect this distinction. Table 5 4. Summary of blade leng th (mm) data for hafted bifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 60 30.77 19.49 61.8 8.95 0.29 Savannah Ogeechee 378 33.03 12.2 69.01 10.04 0.3 Ocmulgee 57 27.9 12.64 48.74 9.5 0.34 Flint 69 31.27 15.65 59.14 8.18 0.26 Chattahoochee 274 38.54 17.54 81.69 11.24 0.29 Aucilla Suwannee 375 40.71 16.51 113.52 12.31 0.3 Tampa Bay 51 42.84 26.35 67.27 10.41 0.24 Total 1,264 As Table 5 4 illustrates, the highest mean values for blade length are found in the Chattahoochee, AucillaSuwannee, and Tampa Bay drainages. Blade length for hafted bifaces is seen to increase incrementally to the west and south of the Flint River drainag e, as the greatest mean values for this attribute are observed in the Aucilla Suwannee and Tampa Bay drainages. Figure 5 5 is a percentile plot that contains all raw sample data for biface blade length, allowing comparison of numerous data sets on the same graph. Percentile plots are especially well suited to analysis of variation in side notched hafted bifaces. Historically, these artifacts have been divided into distinct types (Taylor, Bolen, and Big Sandy) (Bullen 1958, 1975; Kneberg 1956; Michie 1966), and I expected any subregional variations that exist to be graphically expressed.

PAGE 164

164 In all percentile plots conducted for this dissertation, xi represents the percent probability that a data point exists. The calculation is performed against all sorted data using the following equations: )(100iiPx n i Pi/ ) 5 0 ( where i is the index of the sorted data, and n is the number of values in the sample. For instance, using the Santee Cooper blade length sample in Table 5 4, y is the ratio scale attribute (i.e., blade length), the first data point is i with a value of 1, the last data point is i with a value of 60, and n is the last number for the sample, which is 60. For the entire sequence of samples for which blade length was measured, n =1,264. In general, data from all of the drainages (abbreviated for all percentile plots) fit to one of two groupings for hafted biface blade length. Samples from the Ocmulgee, Santee Cooper, Flint, and Savannah Ogeechee River drainages are more skewed to the left of n (where n is the sample of all drainage data) with a wider and broader trend, while data from the Chattahoochee, Aucilla Suwannee, and Tampa Bay drainages tend to cluster to the right with their higher overall values and more distinct trend. T he possibly hypertrophic bifaces previously noted with respect to the Aucilla Suwannee sample can be seen at the top right of the percentile plot. Coefficient of variation data indicate that the widest range of variation for hafted biface blade lengths is in the Ocmulgee R iver drainage (c.v.= 0.34). Meanwhile the least amount of variation is found in data from Tampa Bay ( c.v.= 0.24). The SavannahOgeechee and AucillaSuwannee drainages display the same degree of variation in biface blade length as one anothe r ( c.v.= 0.3), as do the Santee Cooper and Chattahoochee drainages ( c.v.= 0.29).

PAGE 165

165 An examination of the statistical trends within blade length data allows for more intensive analysis of these gross trends. In the SanteeCooper River drainage, most blade lengt hs are nearer the minimum values and blade length is among the least varied of the metric attributes for this sample. Data for blade length for hafted bifaces from the SanteeCooper River drainage are less concentrated around the mean due to a large degree of variation. In the Ocmulgee River drainage, blade length is among the most varied attributes in distribution. This trend is also borne out by coefficients of variation, as previously discussed. The negative kurtosis value for blade length in the Ocmulgee River drainage may reflect tool conservation; numerous hafted bifaces in various stages of resharpening and reuse are found in this sample, which account for the flatter distribution of blade lengths. A relatively high skewness value was found for blade lengths of hafted bifaces from the Chattahoochee River drainage, indicating that a greater percentage of hafted bifaces here have blade lengths nearer the minimum value. In the Aucilla Suwannee River drainage, data for hafted biface blade length resulted in high values for both skewness and kurtosis. This indicates that a greater percentage of hafted bifaces from the Aucilla Suwannee River drainage have blade lengths near the minimum value, and that blade length is among the least varied in distribution of the attributes recorded for this sample. Additionally, blade length data from the Aucilla Suwannee River drainage have a leptokurtic distribution; these values are highly concentrated around the mean (40.71 mm) due to less variation in the sample. It is w orth noting that the lowest minimum value for hafted biface blade length is seen for the Savannah Ogeec hee River drainage ( Table 5 4). It would be difficult to argue that this relatively small size blade is a reflection of economizing behavior, as it was recovered in close

PAGE 166

166 proximity to sources of highquality raw material. Perhaps this and similar examples of relatively short blades r eflect long distance movements both to and from the Santee Cooper or Ocmulgee R iver drainages, where chert sources are absent. Alternatively, perhaps examples of small b lades in the SavannahOgeechee R iver drainage are a residue of retooling at the Allenda le chert quarries, by groups living in this drainage, or traveling into it. This is but one of many fine examples of equifinality in this study of Early Archaic hunter gatherers. Since the percentile plot shows distributions that are not exactly normal, and the samples have varying means, I conducted a Levenes test to determine whether the differences in sample variances for biface blade length are statistically significant (Gastwirth et al. 2009; Levene 1960). Table 5 5 presents the results of the Levene s test for blade length varianc e, as well as all other ratio scale attributes. As Levenes Statistic (W) is greater than the critical value (F) obtained for these data, the null hypothesis can be rejected; the inequality of variances for blade length data is statistically significant. Table 55. Results of Levene's tests conducted for ratio scale data Ratio scale Attributes (mm) Levene's Statistic (W) Critical Value (a=0.05) (F) Blade length 3.18401999 2.105781361 Shoulder width 3.620521401 2.106618596 Maximum thickness 0.907633739 2.104132095 Shoulder length 13.54198384 2.1040816 Between notch width 5.410834234 2.10405835 Notch height 22.27016129 2.104132095 Base width 3.361457698 2.104734877 Another measurement of the blade portion of side notched hafted bifaces is shoulder width. I did not know what patterning to expect for shoulder width, since I could not anticipate the level to which this attribute might covary with blade length. Table 5 6 displays comparative

PAGE 167

167 data for shoulder width measurements of hafted bifaces recovered from all seven river drainages. Only those hafted bifaces which have two complete (unbroken) shoulders are included in these data. Table 5 6. Summary of shoulder width (mm) data for hafted bifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 58 21.63 15.84 31.31 3.58 0.17 Savannah Ogeechee 399 22.05 17.46 36.68 3.55 0.16 Ocmulgee 56 22.09 14.87 36.52 4.87 0.22 Flint 34 21.28 16.42 32.15 4.18 0.2 Chattahoochee 207 26.08 15.92 44.87 4.69 0.18 Aucilla Suwannee 319 26.29 16.81 41.07 3.96 0.15 Tampa Bay 60 25.2 17.15 34.89 4.01 0.16 Total 1,133 T he widest range of variation for hafted biface shoulder widths is in the Ocmulgee River drainage (c.v.= 0.22 ), while the least variance is found in the Aucilla Suwannee sample (c.v.= 0.15). The same c.v. ( 0.16) was obtained for data from the Savannah Ogeechee and Tampa Bay drainages ; these data share the same degree of variance. A s previously stated, a higher coefficient of variation (compared to other drainages) also was obtained for blade length data from the Ocmulgee River drainage. Collectively, these data point to more variance for the blade portion of hafted bifaces from the Ocmulgee River drainage than what was observed for other Coastal Plain drainages. As Table 5 6 shows, the highest mean values for shoulder width are seen in hafted bifaces from the Chattahoochee, AucillaSuwannee, and Tampa Bay drainages. This trend mirrors what I observed for blade leng th in bifaces from the same drainages. Of course, this mirroring trend is correlative, since these data do not necessarily indicate a cause and effect relationship. A more detailed analysis of trends for shoulder width is needed here, for

PAGE 168

168 comparative purposes. In the Ocmulgee and Flint River drainages, the majority of hafted bifaces have shoulder widths near the minimum values. Also, shoulder width is among the least varied in distribution among the ratio scale attributes measured for hafted bifaces from th e Ocmulgee and Chattahoochee River drainages. This runs somewhat contrary to what might otherwise have been assumed for the Ocmulgee River drainage given the coefficient of variation analysis provided above. Figure 5 6 presents a percentile plot that cont ains all raw sample data for biface shoulder width. Samples from all of the drainages generally fit to one of two groupings for shoulder width. Samples from the Flint, Ocmulgee, Santee Cooper, and SavannahOgeechee River drainages are more skewed to the left of n (again, where n is the sample of all drainage data), and Tampa Bay Chattahoochee, and Aucilla Suwannee samples are seen to cluster to the right with their higher overall values The trends for both segments of these data are broad and wide with numerous data points from the Ocmulgee and Tampa Bay samples spanning the difference in shoulder width groupings greater than (approximately) 25 mm. As this plot shows non normal distributions, and the samples have varying means, I conducted a Levenes test to determine whether the differences in sample variances for biface shoulder width are statistically significant (Gastwirth et al. 2009; Levene 1960). The results of the Levenes test for biface shoulder width variance are displayed in Table 55. As Levenes Statistic (W) is greater than the critical value (F) obtained for these data, the null hypothesis can be rejected; the inequality of variances for shoulder width data is statistically significant. A comparison of the ratio of mean blade length to mean s houlder width for hafted biface data from all seven drainages, seems to reflect both raw material availability and tool curation.

PAGE 169

169 Table 5 7 illustrates broad trends in these ratios, in relation to the proximity of Coastal Plain chert. The SanteeCooper and Ocmulgee River drainages are the only drainages in this study without local access to Coastal Plain chert. These river drainages produced the lowest ratios of mean blade length to mean shoulder width, with values ranging from 1.26 in the Ocmulgee River dr ainage, to 1.42 in the Santee Cooper River drainage. These lower ratios may reflect curation or tool maintenance strategies. Such economizing behavior may have been necessary for groups facing a paucity of locally available high quality raw material. The o ther five drainages, all of which had locally available Coastal Plain chert, produced higher ratios of mean blade length to mean shoulder width, ranging from 1.47 in the Flint River drainage, to 1.7 in the Tampa Bay drainage. While the difference between t he ratio calculated for hafted bifaces from the Flint and the SanteeCooper River drainages is only 0.05, it is noteworthy that all drainages with locally available Coastal Plain chert produced higher ratios than those in chert poor regions. Table 5 7. Ratios of mean blade length to mean shoulder width (mm) for side notched hafted bifaces Drainage Mean Blade Length (MBL) Mean Shoulder Width (MSW) Ratio of MBL:MSW Local chert availability Santee Cooper 30.77 21.63 1.42 No Savannah Ogeechee 33.03 22.05 1.5 Yes Ocmulgee 27.9 22.09 1.26 No Flint 31.27 21.28 1.47 Yes Chattahoochee 38.54 26.08 1.48 Yes Aucilla Suwannee 40.71 26.29 1.55 Yes Tampa Bay 42.84 25.2 1.7 Yes Hafted biface maximum thickness is the final blade element attribute to be discussed with respect to the seven examined Coastal Plain river drainages. Table 5 8 presents comparative data

PAGE 170

170 collected for this attribute. The highest mean values for maximum thickness are seen for bifaces from the Chattahoochee, Aucilla Suwannee, and Tampa Bay drainages. This trend echoes patterned variation observed in measured blade length and shoulder width for bifaces from these same Coastal Plain drainages. Mean values for all three measured portions of the biface blade element, which include blade length, shoulder width, and maximum thickness, are consistently highest in the Chattahoochee, Aucilla Suwannee, and Tampa Bay drainages. Blade length and shoulder width may be take n as proxy evidence for blade size; maximum thickness also can be taken as an additional measurement of overall blade size. Table 5 8. Summary of maximum thickne ss (mm) data for hafted bifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 66 6.5 4.16 9.65 1.04 0.16 Savannah Ogeechee 570 6.54 3.89 10.95 0.97 0.15 Ocmulgee 72 6.75 3.43 9.75 0.98 0.14 Flint 97 6.59 4.84 9.17 0.94 0.14 Chattahoochee 319 7.22 3.57 10.54 0.95 0.13 Aucilla Suwannee 440 7.4 5.04 11.19 1.04 0.14 Tampa Bay 74 7.42 5.38 9.99 1.08 0.15 Total 1,638 I constructed a percentile plot that contains all raw sample data for maximum biface thickness (Figure 57). For the most part, data points from all of the drainages fit to one of two groupings for maximum thickness; samples from the Santee Cooper, SavannahOgeechee, and Flint River drainages are more skewed to the left of n. Meanwhile, data from the Chattahooch ee, Aucilla Suwannee, and Tampa Bay drainages tend to cluster to the right with higher overall values. While the trends of both groupings can be described as relatively distinct, the left skewed data grouping can be seen to share a more similar trend M any data points from the Ocmulgee River drainage are found between the major data groupings, spanning the difference between the

PAGE 171

171 trends at and below the median (at the 50th percentile). At higher values for thickness, Ocmulgee River drainage data can be seen to generally share the same trend as data from the Savannah Ogeechee River drainage. As seen in Table 5 8, coefficient of variation data indicate that the widest range of variation for maximum thickness is in the Santee Cooper Ri ver drainage (c.v.=0.16); t he lowest degree of variation is found in the Chattahoochee sample (c.v.= 0.13). However, coefficients of variation obtained for all samples are broadly similar, and data from the Ocmulgee, Flint, and Aucilla Suwannee River drainages produced the same stati stical result ( c.v.=0.14), as did the Savannah Ogeechee and Tampa Bay drainage samples ( c.v.=0.15). These interpretations are therefore based on comparisons of degrees of variance; in general, all examined assemblages display low variance for maximum biface thickness. As the percentile plot described above show s distributions that are not exac tly normal and the samples have varying means, I conducted a Levenes test of equality of variances to determine whether the differences in sample variances for maximum biface thickness are statistically significant (Gastwirth et al. 2009; Levene 1960). Table 55 presents the results of the Levenes test for variance in maximum biface thickness. Levenes Statistic (W) is not greater than the critical value (F) obt ained for these data. Therefore, the null hypothesis cannot be rejected; the inequality of variances for maximum thickness data is not statistically significant. This finding is not unexpected given the broad similarity of coefficients of variation for max imum thickness described above. Figure 5 8 illustrates trends for hafted biface blade length, shoulder width, and maximum thickness across the Southeastern Coastal Plain Figure 59 illustrates the same information, in

PAGE 172

172 the context of the Coastal Plain, and with reference to specific measured values. As Figure s 5 8 and 59 show the highest mean blade lengths and shoulder widths are seen for hafted bifaces from the Chattahoochee, Aucilla Suwannee, and Tampa Bay drain ages. The highest mean thicknesses for hafted bifaces also are seen for hafted bifaces from those drainages, although that is rather more difficult to discern from the histogram in Figure 5 9. As previously discussed, mean blade lengths for hafted bifaces from the SanteeCooper, SavannahOgeechee, Ocmulgee, and Flint River drainages range from 27.9 to 33.03 mm. Mean blade lengths for hafted bifaces from the Chattahoochee, Aucilla Suwannee, and Tampa Bay drainages range from 38.54 to 42.84 mm. Mean shoulder widths for hafted bifaces from the SanteeCooper, SavannahOgeechee, Ocmulgee, and Flint River drainages range from 21.28 to 22.09 mm. I recorded mean shoulder widths for hafted bifaces from the Chattahoochee, Aucilla Suwannee, and Tampa Bay drainages ranging f rom 25.2 to 26.29 mm. Mean maximum thicknesses for hafted bifaces from the SanteeCooper, SavannahOgeechee, Ocmulgee, and Flint River drainages range from 6.5 to 6.75 mm; these values range from 7.22 to 7.42 mm for the sample from the Chattahoochee, Aucil la Suwannee, and Tampa Bay drainages. As detailed above, percentile plots allowed for comparison of various blade portion attributes from all of the examined drainages. With regard to blade length and shoulder width, data generally fit to one of two geogra phical groupings; one including the SanteeCooper, Savannah Ogeechee, Ocmulgee, and Flint River drainages, and one including the Chattahoochee, Aucilla Suwannee, and Tampa Bay drainages. For maximum hafted biface thickness, many data points from the Ocmulgee River drainage are found between the major regional data groupings, spanning the difference between their trends at and below the median. Ocmulgee River drainage

PAGE 173

173 data can be seen to generally share the same trend as data from the Savannah Ogeechee River drainage at higher values obtained for maximum biface thickness. Recognition of these apparent subregional technological traditions relating to the blade portion of hafted bifaces was based on both statistical and graphical analyses. However, I am obliged to qualify discussions of the blade portions of these tools in light of the numerous alterations blades undergo as a result of resharpening and reworking during tool use life. Overall, the hafted biface s I examined throughout Florida have (on average) sig nificantly longer blades, wider shoulders and greater maximum thickness than I observed elsewhere. This patterning is arguably linked to the subregional technological tradition known as Bolen, typically applied to Florida expressions of the Early Side Not ched Horizon. As such, it may be a reflection of a social boundary comprised of the peninsular Gulf Coast, as well as north and northcentral Florida. These trends in hafted biface blades also may be related to specific environmental constraints, as well a s to a different mobility pattern along the Florida peninsula. Two constraints in particular are vital to consider with respect to the Early Holocene; these are the availability of potable water, and the availability of highquality raw material. As was di scussed in Chapter 3, these constraints were a speci fic challenge to Florida hunter gatherers, as they dealt with the Bolen Drought during the Early Archaic period (Dunbar 2002). Water had to be obtained from aquifer fed springs, or f rom lakes perched on limestone, as the water table was some seven meters lower than contemporary levels. Fortuitously, c herts found in Tertiary karsts are among the most homogeneous and highest quality raw materials available in Florida. These cherts also may have come in larg er original package sizes (when first obtained at chert sources) than what

PAGE 174

174 was available in other portions of the Coastal Plain, leading ultimately to larger sized blades for side notched hafted bifaces. Regardless of original raw material package size, on e obvious advantage of using a larger initial blade is that a larger blade remains to be resharpened following breakage (cf. Bleed 1986; Christenson 1997; Kelly 1988). This last point may well have been a consideration for Early Archaic groups in Florida. Floridas Early Archaic hunter gatherers may have been somewhat tethered to specific Tertiary karsts for both water and raw mat erial (cf. Sassaman 2003). Perhaps they were more closely tied to specific locations for raw material and for water, as well as f or settlement and social interaction, than what has been proposed for their contemporaries in ot her portions of the Southeast. I explore the implications of this patterned variation in Chapter 6. It is also possible that more general characteristics of the Florida landscape, including those relating to topography and vegetation, differed from environments in other portions of the Coastal Plain, such that particular technological choices relating to hafted biface size were appropriate. Ethnoarchaeological research suggests that larger projectile technology is highly correlated to cover structure. Specifically, a more open cover structure may be amenable to small hunting gear, while thick vegetative cover may require larger projectiles to ensure greater shot distance. Thick cover also can provide concealment of larger sized weapons from the wildlife being hunted (Bartram 1997). Discussion for Haft Element of Hafted Bifaces As previously discussed, metric variables relating to hafted biface haft element in clude: shoulder length, between notch width, notch height, and base width. These measurements were taken for all bifaces. The haft element of a biface, situated as it is inside a haft, is presumed to be

PAGE 175

175 the most stable portion of a tool during use life (Austin a nd Mitchell 2010; Bissett 2003; Keeley 1982). As such, the shoulder length (the distance from the top of the side notch to the base of the tool) is presumed to be one of the most stable portions of a hafted biface. Shoulder lengths varied across the examin ed river drainages, but certain broad trends became evident in the data. Table 59 presents shoulder length data for each of the river drainages that produced hafted bifaces. Table 5 9. Summary of shoulder leng th (mm) data for hafted bifaces Drainage n Me an Min Max s.d. c.v. Santee Cooper 65 10.14 7.2 15.67 1.79 0.18 Savannah Ogeechee 586 10.33 1.53 26.15 1.93 0.19 Ocmulgee 70 10.82 6.8 15.87 1.89 0.17 Flint 96 10.38 6.9 17.14 1.84 0.18 Chattahoochee 318 10.43 6.14 16.76 1.66 0.16 Aucilla Suwannee 443 12.4 6.31 22.24 2.71 0.22 Tampa Bay 75 12.09 8.21 22.2 2.49 0.21 Total 1,653 Data for patterned variation in biface shoulder length for the AucillaSuwannee and Tampa Bay drainages are suggestive of a subregional trend for this attribute, beginning at the Aucilla River. As Table 5 9 shows, mean values for biface shoulder length are significantly higher in the Aucilla Suwannee and Tampa Bay drainages (12.4 and 12.09 mm respectively). This is in contrast with shoulder length data for bifaces from the other five examined drainages (a range of 10.14 to 10.82 mm). Figure 5 10 presents a percentile plot that contains all raw sample data for biface shoulder length. Samples from all of the drainages generally fit to one of two groupings for shoulde r length Samples from the SanteeCooper, SavannahOgeechee, Flint, Chattahoochee, and Ocmulgee River drainages are more skewed to the left of n, with a relatively distinct (i.e., shared) trend. Notably, data from these drainages can be seen to share a sim ilar trend at and

PAGE 176

176 below the median value (at the 50th percentile) In the second major grouping of data points, Aucilla Suwannee and Tampa Bay drainage samples are seen to cluster to the right with their higher overall values N umerous data points from the Tampa Bay sample (above the median value) span the difference between the left skewed data grouping and the Aucilla Suwannee sample. Shoulder length data for the Ocmulgee River drainage also can be seen to span the difference between the two major groupin gs at data points greater than the median value, to the right of the left skewed distribution. In general, all of the examined drainages have low variance for biface shoulder length; these interpretations therefore are based on relative variance in the stu dy area. The widest range of variation for hafted biface shoulder lengths is in the Aucilla Suwannee River drainage (c.v.= 0.22); this coefficient of variation is slightly higher than that obtained for data from Tampa Bay (c.v.=0.21). As the mean values obt ained for these samples also are similar, the difference in variance is not significant. Meanwhile, the lowest degree of variation is found for samples from the Chat tahoochee River drainage (c.v.= 0.16 ). I obtained t he same c.v. ( 0.18) for data from the SanteeCooper and Flint River drainages; these data share the same degree of variance for biface shoulder length. Reexamining certain other statistical trends first discussed with respect to individual river drainages is of utility here, in teasing out so urces of patterned variation for shoulder length data throughout the Coastal Plain. For example data from the SanteeCooper Ri ver drainage indicate that a greater percentage of hafted bifaces there are near the minimum value for this attribute. Additional ly, biface shoulder length is among the least varied of the metric attributes analyzed in the SanteeCooper region; these data are less concentrated around the mean due to a high degree

PAGE 177

177 of variation within observed values Observed patterned variation in biface shoulder length data may indicate a trend toward tool curation and raw material conservation in the Santee Cooper River drainage. Such a trend is somewhat intuitive, as the nearest sources of high quality Coastal Pl ain chert are located some 97 k m to the west near the Savannah River ( Figure 5 3). Data for biface shoulder length in the SavannahOgeechee River drainage also indicate a greater percentage of tools near the minimum value for this attribute. These data are highly concentrated around the mea n and are among the least varied of the metric attributes for this river drainage. This patterning for biface shoulder length is broadly dissimilar to that witnessed in the SanteeCooper River drainage, where data were significantly less concentrated aroun d the mean. I submit that in the SavannahOgeechee River drainage, these data are highly suggestive of a standardized haft length. In the Flint River drainage, shoulder length data are among the least varied in distribution among measured attributes. This may be indicative of some level of standardization in haft element length in this drainage, similar to the patterning observed for the SavannahOgeechee River drainage. Hafted biface data from Tampa Bay are broadly similar to those from the Savannah Ogeech ee River drainage, with respect to shoulder length. Specifically, Tampa Bay data indicate a greater percentage of tools near the minimum value for this attribute. Also, these data are concentrated around the mean value. As was suggested with respect to the Savannah Ogeechee River drainage, patterning for shoulder length data in Tampa Bay may be suggestive of a standardized shoulder length.

PAGE 178

178 As percentile plots for biface shoulder length show nonnormal distributions, and the samples have varying means, I co nducted a Levenes test to determine whether the differences in sample variances for biface shoulder length are statistically significant (Gastwirth et al. 2009; Levene 1960). The results of the Levenes test for variance in shoulder length are displayed i n Table 5 5. As Levenes Statistic (W) is greater than the critical value (F) obtained for these data, the null hypothesis can be rejected; the inequality of variances for shoulder length data is statistically significant. Between notch width is another me asurement taken of the biface haft element. Similar to shoulder length, this attribute is presumed to be one of the most stable portions of a hafted biface, as it is assumed to have been protected inside a haft during tool use life (Austin and Mitchell 201 0; Bissett 2003; Keeley 1982). Table 510 presents comparative data for between notch width measurements taken for hafted bifaces from all seven examined river drainages. Table 5 10. Summary of betweennotch width (mm) data for hafted bifaces Drainage n M ean Min Max s.d. c.v. Santee Cooper 66 15.73 11.11 22.23 2.32 0.15 Savannah Ogeechee 581 15.4 10.22 25.32 2.47 0.16 Ocmulgee 69 15.66 10.65 24.34 3.3 0.21 Flint 97 14.84 10.2 19.24 1.71 0.12 Chattahoochee 320 15.62 11.27 26.86 2.51 0.16 Aucilla Suwannee 450 15.8 10.05 25.2 2.53 0.16 Tampa Bay 77 15.75 9.72 25.45 2.97 0.19 Total 1,660 Figure 5 11 presents a percentile plot that contains all raw sample data for between notch width. For the most part, data points from all of the drainages are characterized by a singular grouping below the 50th percentile, indicating that the same general trend is shared by data in all examined drainages. However, above the median, Flint River drainage data can be seen to

PAGE 179

179 diverge markedly from data obtained for other drainages, as between notch width values for this sample were among the lowest observed. The trend observed for Flint River drainage data is remarkable in these data, and must be distinguished from patterning observed for Ocmulgee R iver drainage data, where low frequencies of outliers skew the sample to the right beginning below the 90th percentile. As Table 5 10 demonstrates, data from the Ocmulgee River drainage produced the highest coefficient of variation (0.21) of any Coastal Pl ain data set, illustrating a high degree of variation for biface between notch width. It bears repeating that the SanteeCooper and Ocmulgee River drainages are located away from high quality chert raw material. Perhaps some level of raw material conservat ion is reflected in this statistical trend. Meanwhile, a lower range of variation compared to other assemblages is found for samples from the Flint River drainage (c.v.= 0.12). T he same c.v. ( 0.16) value was obtained for data from the Savannah Ogeechee, Chattahoochee, and Aucilla Suwannee River drainages; these data share the same degree of variance for biface between notch width. Other specific st atistical trends first discussed with respect to individual river drainages have great relevance here in understanding patterned variation for between notch width data collected using hafted bifaces throughout the study region. Three river drainages (the Ocmulgee, the Chattahoochee, and the Aucilla Suwannee) all produced statistical evidence to be reviewed here, with r espect to patterning in between notch width. In the Ocmulgee River drainage, a greater percentage of hafted bifaces are nearer the minimum value for between notch width. This may be another indication of tool curation or raw material conservation i n this river drainage. Meanwhile, the majority of hafted bifaces from both the Chattahoochee and Aucilla Suwannee

PAGE 180

180 River drainages also p roduced data indicating between notch widths nearer the minimum values. Significantly, in these more western and souther n riv er drainages, between notch width also is among the least varied in distribution among continuous variables. This low degree of variability may be indicative of a relatively high degree of standardization for between notch width for bifaces from the C hattahoochee and AucillaSuwannee River drainages. Patterned variation in data for betweennotch width also can be examined in light of the larger toolkit. The manufacture and maintenance of consistent haft areas would have ensured that newly produced bifa ces would fit existing hafts (Austin and Mitchell 2010). Perhaps in the SanteeCooper and Ocmulgee R iver drainages, a wider variety of haft sizes was employed as compared to elsewhere in the Coastal Plain. This is especially intrigui ng in the case of the O cmulgee R iver drainage, which has been interpreted as a key aggregation location for regional groups during the Early Archaic (Blanton and Snow 1986, 1989). Groups from throughout the region may well have brought a variety of haft sizes when they traveled to the Ocmulgee environs. Since the percentile plots show distributions that are not exactly normal, and the samples have varying means, I conducted a Levenes test to determine whether the differences in samp le variances for biface between notch width are statistically significant (Gastwirth et al. 2009; Levene 1960). Table 5 5 presents the results of the Levenes test for variance in biface between notch width. As Levenes Statistic (W) is greater than the critical value (F) obtained for th ese data, the null hypothesis can be rejected; the inequality of variances for between notch width data is statistically significant.

PAGE 181

181 Notch height data were collected for hafted bifaces from all seven Coastal Plain drainages; Table 5 11 presents these data for comparative purposes. As displayed in Table 511, mean values for biface notch height are significantly higher in the Aucilla Suwannee and Tampa Bay drainages (7.07 and 6.8 mm, respectively). The more northern drainages produced lower mean values for notch height, ranging from an overall minimum in the Santee Cooper River drainage (4.88 mm), to a high value in the Chattahoochee River drainage (5.7 mm). While c oefficient of variation data indicate that the widest range of variation for hafted biface not ch height is in the Aucilla Suwannee River drainage (c.v.=0.29), in general, all of the examined drainages have low variance for hafted biface notch height; this discussion focuses on relative variance. The least amount of variance for this measured attrib ute is found for samples from the SanteeCooper and Chatt ahoochee River drainages (c.v.= 0.19). The SavannahOgeechee and Tampa Bay drainages display an equal amount of variance in biface notch height (c.v.= 0.24). Table 5 11. Summary of notch heig ht (mm) data for hafted bifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 65 4.88 3.05 8.11 0.92 0.19 Savannah Ogeechee 577 5.36 2.7 13.67 1.28 0.24 Ocmulgee 71 5.62 3.34 8.92 1.25 0.22 Flint 93 5.41 3.2 8.93 1.15 0.21 Chattahoochee 317 5.7 2.99 9.07 1.1 0.19 Aucilla Suwannee 440 7.07 3.6 15.41 2.08 0.29 Tampa Bay 75 6.8 3.6 13.41 1.61 0.24 Total 1,638 I constructed a percentile plot that contains all raw sample data for biface notch height (Figure 5 12). Samples from all seven drainages generally fit to one of two groupings for notch height. Data from the Santee Cooper, SavannahOgeechee, Ocmulgee, Flint, and Chattahoochee River drainages are m ore skewed to the left of n with a relatively distinct (i.e., shared) trend. In

PAGE 182

182 the second major grouping of data points, Aucilla Suwannee and Tampa Bay drainage samples are seen to cluster to the right with their higher overall values. Numerous data points from the Tampa Bay sample (above the median value) span the differen ce between the left skewed data grouping and the Aucilla Suwan nee sample, similar to the patterning observed for shoulder length data. I examined all individual drainages that produced hafted bifaces for this research, to determine if other statistical tr ends were evident for notch height. Five drainages produced evidence to be further discussed below; the Santee Cooper, SavannahOgeechee, Flint, Aucilla Suwannee, and Tampa Bay. Hafted bifaces from the SanteeCooper River drainage have notch heights less c oncentrated around the mean value. This may indicate some degree of tool conservation in this region. In the Savannah Ogeechee sample, most hafted bifaces have notch heights nearer the minimum value, and data are highly concentrated around the mean value. Notch height is among the least varied in distribution among measured attributes. Patterned variation in notch height for hafted bifaces from the Flint, Aucilla Suwannee, and Tampa Bay drainages is broadly similar to that observed for the SavannahOgeechee River drainage. As the plots in Figure 512 show nonnormal distributions, and the samples have varying means, I conducted a Levenes test to determine whether the differences in sample variances for biface notch height are statistically significant (Gas twirth et al. 2009; Levene 1960). The results of the Levenes test for variance in biface notch height are displayed in Table 55. As Levenes Statistic (W) is greater than the critical value (F) obtained for these data, the null hypothesis can be rejected; the inequality of variances for notch height data is statistically significant.

PAGE 183

183 Biface base width is the final haft element attribute to be discussed with respect to the seven examined Coastal Plain river drainages. Table 5 12 presents comparative data collected for this attribute. In general, all of the examined drainages have low variance for hafted biface base width; this discussion necessarily focuses on comparative degrees of variance. The widest range of variation for hafted biface base width is in the Ocmulgee River drainage (c.v.= 0.17). As previously discussed, the negative kur tosis value ( 0.56) for base width in the Ocmulgee River drainage further demonstrates that base width is among the most varied in distribution among measured attributes. Meanwhile, the least amount of variance is found in the Chattahoochee sample (c.v.= 0.12). The Flint and Aucilla Suwannee River drainages display an equal degree of variation in biface base width (c.v.= 0.13). Table 5 12. Summary of base wid th (mm) data for hafted bifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 52 22.16 17.25 30.13 3.1 0.14 Savannah Ogeechee 542 22.2 10.57 39.68 3.48 0.16 Ocmulgee 52 23.3 16.05 32.62 3.97 0.17 Flint 73 23.25 16.89 30.89 2.99 0.13 Chattahoochee 276 23.72 17.86 33.88 2.84 0.12 Aucilla Suwannee 416 24.27 16.11 36.32 3.08 0.13 Tampa Bay 67 23.75 16.03 32.94 3.67 0.15 Total 1,478 I constructed a percentile plot that contains all raw sample data for biface base width (Figure 5 13). In general, data points for all seven drainages can be seen to share a similar trend, with a spread ranging from approximately 10 to almost 40 mm. Despite this notable range, data for base width remain largely clustered into a singular broad trend, with the exceptions of outliers at the lower and higher ends of the percentile axis.

PAGE 184

184 Since the percentile plots show distributions that are nonnormal, and the samples have varying means, I conducted a Levenes test to determine whether the differences in sample variances for biface base width are statistically significant (Gastwirth et al. 2009; Levene 1960). Table 5 5 presents the results of the Levenes test for variance in biface base width. As Levenes Statistic (W) is greater than the critical value (F) obtained for these data, the null hy pothesis can be rejected; the inequality of variances for hafted biface base width data is statistically significant. Figure 5 14 shows all mean notch heights, betweennotch widths, shoulder lengths, and base widths for hafted bifaces examined from all sev en Coastal Plain river drainages. As this chart illustrates the haft elements of side notched hafted bifaces throughout the Southeast are quite similar. This broad patterning alone should raise doubts as to the utility of making rigid distinctions between these artifacts when they are found in different pa rts of this region. However, three attributes, notch height, shoulder length, and base width, stand out slightly on this chart, as the means of these tend to be greater in samples from the Aucilla Suwanne e and Tampa Bay t han what I observed elsewhere. As expressed in Figure 5 14, the highest mean notch heights are seen for hafted bifaces from the Chattahoochee, Aucilla Suwannee, and Tampa Bay drainages. Mean notch heights for hafted bifaces from the SanteeCooper, Savannah Ogeechee, Ocmulgee, and Flint River drainages range from 4.88 to 5.62 mm, as previously noted. Mean notch heights for hafted bifaces from the Chattahoochee, AucillaSuwannee, and Tampa Bay drainages range fr om 5.7 to 7.07 mm. Mean between notch widths for hafted bifaces do not exhibit similar patterned variation in t hat high values are not restricted to a particular geographical region. I observed

PAGE 185

185 values ranging fr om 14.84 to 15.8 mm for betweennotch width. The highest mean values for shoulder length are seen for hafted bifaces from the AucillaSuwannee and Tampa Bay drainages. Mean shoulder lengths for hafted bifaces from the Santee Cooper, Savannah Ogeechee, Ocmulgee, Flint, and Chattahoochee River drainages range from 10.14 to 10.82 mm. Mean shoulder lengths for hafted bifaces from the Aucilla Suwannee and Tampa dr ainages range from 12.08 to 12.4 mm. As Figure 5 14 illustrates, the highest mean values for base width are seen for hafted bifaces from the Chattahoochee, AucillaSuwannee, and Tampa Bay drainages (23.72, 24.27, and 23.75 mm, respectively). Mean values for base width from all other river drainages examined range from 22.16 to 23.3 mm. As previously discussed, percentile plots allowed me to compare various haft element attributes from all of the examined drainages. With respect to two portions of the haft e lement, shoulder length and notch height, da ta generally fit to one of two geographical groupings; one tied to the Santee Cooper, SavannahOgeechee, Ocmulgee, Flint, and Chattahoochee River drainages, and one tied to the Aucilla Suwannee and Tampa Bay drai nages. Hafted bifaces in peninsular Florida appear to be especially well suited for use as knives. In general, a longer haft element (measured here as shoulder length) is better able to manage stresses applied through lateral pressure in cutting tasks, which can cause the biface to twist in its haft. Additionally, lengthening the notches (measured here as notch height) can help to reduce stress from lateral pressure (cf. Gunn and Rovner 2003:7172). Previous research in the Aucilla Suwannee River drainage has used biface edge angle data to establish that most sidenotched bifaces were used as knives; impact fractures (indicative of use of bifaces as projectiles) were notably rare in that study (Austin and Mitchell 2010). Similarly, I identified impact frac t ures on

PAGE 186

186 fewer than 4% of a sample ( n=9 out of 268) from the Aucilla Suwannee River drainage. Previous researchers have noted that function is a key factor in understanding differences in biface form, and that certain forms are better suited to functions (s uch as cutting) than others (cf. Austin and Mitchell 2010; Bissett 2003; Gunn and Rovner 2003). I would argue that most evidence supports consideration of side notched hafted bifaces (whatever their distribution in the Coastal Plain) as multipurpose tools (cf. Binford 1979; Hayden et al. 1996; Kelly 1988; Odell 1996). For between notch width, Flint River drainage data can be seen to diverge measurably from data obtained for other drainages above the median, a s between notch width values for this sample were among the lowest observed. Detection of these apparent subregional technological traditions relating to the haft element was a result of statistical and graphical analyses of quantitative information for hafted bifaces. Side N otched Unifaces across Draina ges The uniface data set consists o f 255 Early Archaic specimens. The majority (188, or approximately 74% ) of these were found in the SavannahOgeechee River drainage. Four unifaces from the Santee Cooper River drainage were examined. I also examined sidenotched unifaces from the Ocmulgee ( n=9), the Aucilla Suwannee ( n=39), and the Tampa Bay drainages ( n=14). However, no side notched unifaces were found in the Flint drainage, and only one was found in the Chattahoochee; discussions of those two drainages are exempted here. As only one excavated site (the Big Pine Tree site in Allendale County, South Carolina) in my studies produced hafted unifaces ( n=4), the majority of the sc rapers (more than 99% ) analyzed were collected by avocationalists.

PAGE 187

187 The results of patterned variation observed for both nominal and ratioscale attributes among side notched unifaces are presen ted below, with respect to their associated drainages. Then, comparisons of patterned variation among the data collected from the six river drainages are made. Statistical analyses of select attributes that highlight broad similarities and differences among these data sets are also presented. Mapping Nominal Data As discussed in Chapter 4, and as was done with hafted bifaces, my research sample was restricted to Coastal Plain chert artifacts; raw material selection is not considered a variable. Among the eleven nominal attributes, eight (condition, manufacturing condition, fracture type, notch type, basal grinding, cross section, beveling, and later al condition) are a rtifacts of attrition, design, and use. Basal configuration and handedness are the nominal attributes that more likely reflect technological choices made as part of daily practice. Handedness was determined for all examined unifaces, wit h the assumption that the primary tool edge was held away from the tool users dominant hand. For instance, a right handed individual would be expected to use unifacial tools with a primary edge located on the left lateral side of the tool; a lefthanded t ool user would be expected to do the opposite. Trends observed in uniface handedness and basal con figuration are discussed below. Santee Cooper unifaces All four of the unifaces e xamined from the Santee Cooper R iver drainage have a primary working edge along the left lateral edge, indicating that they were all intended for use by right handed individuals. These unifaces also all had a recognizable basal configuration (Appendix D) Half of these unifaces ( n =2, 50% ) have an excu rvate base. Straight bases comprise the other half

PAGE 188

188 of this sample ( n=2). No incurvate bases were observed among unifaces in the Santee Cooper R iver drainage. The small sample size for unifaces in the SanteeCooper River drainage adversely affected certain comparative analyses, as discussed further below. Savannah Ogeechee unifaces A total of seven of the 188 hafted unif aces (approximately 4% ) examined from the Savannah Ogeechee River drainage had a primary working edge along the right lateral edge. This indicates a small percentage of left handed unifaces among the sample. I recorded 175 unifaces that have a recognizable basal configuration from the Savannah Ogeechee River drainage (Appendix D) Thirteen unifaces from this sample did not have a recognizable base shape. Straight bases were the most frequent among the sample ( n=85, approximately 48% ). The second most frequent base shape was incurvate ( n=50, approximately 29% ) Excurvate bases were the le ast frequently represented ( n =40, approximately 23% ) among the Savannah Ogeechee unifaces. Ocmulgee unifaces Within the Ocmulgee R iver drainage, a ll of the examined unifaces have a primary working edge along the left lateral edge, indicating that they were all intended for use by right handed individuals. Among this data set, all nine of the examined unifaces have an identifiable base shape (Appendix D). Excurvate is the most frequent basal configuration ( n=8, approximately 89% ) in this sample. The second most frequent basal configuration is incurvate ( n=1, approximately 11 % ). Straight bases were absent from the examined unifaces in the Ocmulgee River drainage sample.

PAGE 189

189 Aucilla Suwannee unifaces All 39 of the unifaces examined from the Aucilla Suwannee R iver drainage were intended for use by right handed individuals since the primary working edge is located along the left lateral edge. A mong the sample I analyzed from th is river drainage, all hafted unifaces have a recognizable basal configuration (Appendix D) The majority of these unifaces ( n=22, approximatel y 56% ) have an excurvate base. Straight bases were the second most frequent ( n=15), comprising approximately 39% of the sample. Incurvate bases were the least represented among the Aucilla Suwannee sample; two bases (approximately 5% ) were incurvate. Tampa Bay unifaces All 14 of the unifaces examined from the Tampa Bay area have a primary working edge along the left lateral edge. Thirteen (approximately 93% ) of the 14 unifaces from this region have an identifiable base shape (Appendix D). Straight bases ( n= 6, approxim ately 46% ) are found m ost frequently in this sample. The second most frequent basal configuration is excurvate ( n=5, approximately 39% ). Incurvate bases were the least frequent base shape among unifaces in the Tampa Bay area sample, with two (ap proximately 15% ) such examples observed. Comparisons across Drainages I examined s ide notched unifaces ( n =255) from six river drainages in the Coastal Plain. In order to make some inferences about past social boundaries, I compare the nominal data collected for the two attributes that likely reflect patterns of daily practice (basal configuration and handedness ) Here, I compare these attributes among the river drainages in an attempt to determine if any trend s can be identified spatially.

PAGE 190

190 Firs t, a brief discussion of handedness is relevant here. Only seven of the 255 unifaces examined from throughout the Coastal Plain have a primary edge on the right lateral side of the tool. These examples all were from the Savannah Ogeechee River drainage, wh ich notably produced the highest proportion of unifaces in the entire study sample ( n=188, approximately 74% of the study sample). According to thes e data, less than 3% of all examined unifaces (and approximately 4% of unifaces from the Savannah Ogeechee R iver drainage) were used by left handed people. To the n suggest that 96 to 97% of the Early Archaic population was right handed would probably be unwise. Globally and throughout time, 90% of the population is thought to be comprised of right handed people (Coren and Porac 1977). Significantly higher proportions of left handed individuals have been documented for certain hunting and fishing societies, where environmental and cultural constraints on handedness may be more relaxed (Dawson 1974; Marrion 1986; S assaman and Rudolphi 2001). Frequencies of handedness in Early Archaic communities likely were affected by residential patterns, as well as a combination of cultural and genetic factors. Recent ethnographic research suggests that the frequency of left handedness may actually have been highly variable in the ancient past, even in relation to different tasks being performed, and in positive correlation to the prevalence of face to face combat (Faurie et al. 2005). I used data which had been aggregated to Southeastern drainage (as detailed above), and recorded the apparent trends in each area, to attempt to map all patterning in basal configuration for side notched unifaces across the study area. These patterns are displayed in Table 513. Figure 5 15 displays the information from Table 5 13 in terms of percentage. Figure 516 presents the same information, and also allows for the graphic representation of differentiation in

PAGE 191

191 Table 5 13. Absolute and relative frequencies of h afted unifaces by drainage and basal configuration Drainage Incurvate Straight Excurvate Total n % n % n % Santee Cooper 0 0 2 50 2 50 4 Savannah Ogeechee 50 28.57 85 48 .57 40 22.86 175 Ocmulgee 1 11 .11 0 0 8 88.89 9 Chattahoochee 0 0 0 0 1 100 1 Aucilla Suwannee 2 5 .13 15 38.46 22 56 .41 39 Tampa Bay 2 15 .38 6 46 .15 5 38.46 13 Total 55 108 78 241 basal configuration for hafted unifaces throughout the Coastal Plain. In the Santee Cooper Ri ver drainage, no unifaces with incurvate bases were examined, and equal proportions of straight and excurvate bases were observed. Very few unifaces were available for study from the SanteeCooper region ( n =4). Due to this small sample size, it is difficult to describe trends in basal configuration for this river drainage. To the south, in the SavannahOgeechee Riv er drainage, there is a strong tendency for side notched unifaces to be made with straight bases ( Figure 5 16). Slightly less than 49% ( n=85) of the sample had straight bases. Meanwhile, approximately 29% ( n =50) and 23% ( n=40) of this sample have incurvate and excurvate bases, respectively. This pattern changes to the south, with predominantly excurvate bases in the Ocmulgee and AucillaSuwannee r i ver drainages of south Georgia and north Florida. In the southernmost r egion of the study area, Tampa Bay, there appears to have been no clear preference for ei ther straight or excurvate bases. Almost equal frequencies of Tampa Bay unifaces were classified either as straight or excurvatebased.

PAGE 192

192 There is a high degree of inter nal conformity to particular base types in specific regions. Straight bases are most common for unifaces in the SanteeCooper and SavannahOgeechee River drainages. This pattern shifts at the Ocmulgee R iver drainage, as it does with respect to side notched haft ed biface basal configuration. In the Ocmulgee R iver drainage, excurvate bases are most common for unifaces, and incurvate bases are second most common. No straight based unifaces were found in the Ocmulgee artifact assemblages. In the AucillaSuwanne e R iver drainage, excurvatebased unifaces are again most common, but the second most frequently oc curring base type is straight. This pattern changes again in Tampa Bay, where straight based unifaces again are most common, and where excurvate bases occur second most frequently ( Figure 5 16). Based on regional patterning of basal configuration alone, three clusters can be seen within the Coastal Plain; one extending from the Santee Cooper Ri ver drainage, west to the Savannah Ogeechee Ri ver drainage; one in the Ocmulgee Ri ver drainage; and one extending from the A ucilla Suwannee R ive r drainage, south to Tampa Bay. In order to test the significance of these apparent trends in basal configuration throughout the Coastal Plain, I performed a chi square test, creating contingency tables that compared basal configuration for unifaces, to specifi c drainage location. Table 514 s hows the observed number of unifaces of different basal configuration from the six examined river drainages; Table 515 shows the expected number of unifaces of different basal configuration from these regions. The chi square value that resulted is 37.96825, with 10 degrees of freedom and a probability of <.0001. This probability indicates that type of uniface basal configuration and drainage location are not independent variables. Uniface base shape appears to be directly correlated to geography

PAGE 193

193 within the South Atlantic Slope. This same finding was made with respect to the basal configuration of hafted bifaces. Table 5 14. Hafted uniface basal configuration (Observed) Drainage Incurvate Straight Excurvate Total Santee Cooper 0 2 2 4 Savannah Ogeechee 50 85 40 175 Ocmulgee 1 0 8 9 Chattahoochee 0 0 1 1 Aucilla Suwannee 2 15 22 39 Tampa Bay 2 6 5 13 Total 55 108 78 241 Table 5 15. Hafted uniface basal configuration (Expected) Drainage Incurvate Straight Excurvate Total Santee Cooper 0.912 1.792 1.296 4 Savannah Ogeechee 39.9 78.4 56.7 175 Ocmulgee 2.052 4.032 2.916 9 Chattahoochee 0.228 0.448 0.324 1 Aucilla Suwannee 8.892 17.472 12.636 39 Tampa Bay 2.964 5.824 4.212 13 Total 54.948 107.968 78.084 241 Mapping RatioS cale Data As related in Chapter 4, attributes recorded for side notched hafted unifaces largely are the same as those recorded for hafted bifaces. In addition to those, measurements also were taken of the length and angle of the primary tool edge. Since sidenotched unifaces typically are asymmetrical, I recorded both sides of each example separately. This resulted in a total of ten ratio scale attributes collected among the Early Archaic hafted unifaces. As is true for hafted bifaces, the distinction between the bl ade portions of unifaces, versus their haft elements, is key to meaningful analysis. As these tools were used and resharpened, their blade lengths and shoulder widths would have incrementally decreased. Utilization of these

PAGE 194

194 unifaces ultimately would have r esulted in decreased primary edge angles. However, where maintenance of a certain edge angle was desired, this could have been achieved by directed resharpening, resulting in decreased primary blade length and shoulder width. To summarize, while almost eve ry portion of uniface blades was subject to change during the tools use life, an efficacious primary edge angle could have been maintained through reworking of the blade itself. Such a strategy might be expected in regions where Coastal Plain chert could not easily be obtained. Meanwhile, I expected that the haft element of unifaces would exhibit the least evidence of use alteration, and that this portion of the tool was most likely to reflect initial tool design (after Bissett 2003; Goodyear 1995 [1974]; Goodyear et al 1980). The metric variables I studied relating to the uniface haft element include: left and right shoulder length, betweennotch width, left and right notch height, and base width. These measurements were taken for all unifaces (Appendix E) Appendix F presents detailed data regarding the range of variation observed for all metric attributes collected for unifaces, at the individual river drainage level of analysis. Santee Cooper unifaces The highest skewn ess among the attributes of unifaces from the Santee Cooper River drainage was for right notch height (1.0). It would appear that a greater percentage of unifaces here have right notch heights nearer the minimum val ue. Left edge angle and between notch wid th have the highest kurtosis values (4.0 and 2.66, respectively). This suggests t hat left edge angle and between notch width are the least varie d in distribution among ratioscale attributes. The distribution of values for left edge angle in particular is leptokurtic and highly concentrated around the mean with low variation. Left notch height had a negative kurtosis value ( 4.56),

PAGE 195

195 indicating a high degree of variation for this attribute. Taken together, these data seem to suggest that unifaces in the Sante e Cooper R iver drainage were curated and their use was conservative. Since the nearest high quality Coastal Plain chert source is located some 97 km to the west the reserved use of these unifacial tools may be expected ( Figure 5 3). However, it is worth r estating that the small uniface sample size ( n=4) for the SanteeCooper region creates difficulties in describing meaningful trends in ratio scale data for this river drainage. Savannah Ogeechee unifaces Among the unifaces examined from the SavannahOgeec hee River drainage, left shoulder length (1.91) and right shoulder length (0.97) have the highest skewness values. These attributes also have the highest kurtosis values of 8.93 and 2.6, respectively. These data indicate that a greater percentage of unifaces here have right and left shoulder lengths nearer the m inimum values. Shoulder lengths are the least varied in distri bution among metric attributes. Left shoulder length values have a leptokurtic distribution and are heavily concentrated near the mean. T his patterning suggests a somewhat standardized length of haft elemen t for unifaces here, and mirrors evidence collected for hafted bifaces in this river drainage. It bears repeating that the Savannah Ogeechee River d rainage is the location of high quality Coastal Plain chert sources. Several attributes observed for the blade elements (left edge angle, left blade length, right blade length, and shoulder width) of unifaces from the SavannahOgeechee River drainage have a platykurtic distribution ( 0.37, 0.2, 1.09, and 0.52, respectively). These data suggest that the blade portions of unifaces from this region are highly varied. This high variation may be due to the convenience of plentiful and local raw material in the region. These unifaces may reflect

PAGE 196

196 vario us stages of use, and the high variability for the blade portions of these tools may be a function of flintknappers choosing to replace rather than to conserve their unifacial tools. Ocmulgee unifaces The highest skewness among the attributes of unifaces from the Ocmulgee River drainage was for right blade length (0.55) and right notch height (0.46). These data suggest that a greater percentage of unifaces here have right blade length and right notch height nearer the minimum values. Negative skewness val ues were observed for seven attributes; right shoulder length and left edge angle were the attributes with the highest negative skewness ( 1.15 and 1.08, respectively). This suggests that a greater percentage of the data for these seven attributes are nearer the maximum values. The distribution for each of the eleven attributes is platykurtic, indicating a high degree of variation for all of the attributes. However left blade length and between notch width had the highest kurtosis values (1.48 and 1.63, r espectively), which suggests that these two attributes are the least varied in distribution among the metric attributes. Negative kurtosis values were observed for left shoulder length, shoulder width, right notch height, and left notch height ( 1.46, 0.46, 1.28, and 1.00, respectively). These four attributes are the most varied in distribution among this sample. With regard to the blade elements of Ocmulgee River drainage unifaces, we see that right blade length values are closer to the minimum, and me asurements for left blade length are among the least varied values. Shoulder width values are highly varied, with most values near the maximum. All of the unifaces in the Ocmulgee River drainage had a left lateral working edge (the left blade length measur es this working edge). Here, the length of the lateral working edge and the shoulder width are not seen to covary. Throughout the lifespan of unifaces from this

PAGE 197

197 drainage, the tool width (l ocated at the shoulders where they meet the base of the working edge), is being reduced (more than is the blade portion of the tool) during resharpening and use. Taken together, these data suggest that the shoulders of the unifaces are being reduced to maintain the desired working edge angle. The high variability observed among the values for left shoulder length suggests that the distance from the tool base to the left shoulder decreases as a resul t of tool use and maintenance. There appears to be a covariant relationship between the left shoulder length and the shoulder width; as the left shoulder length decreases, s o too does the shoulder width. The goal for flintknappers here is to maintain the desired working edge angle. Therefore, this could be an example of tool conservation and curation for unifaces in the Ocmulgee R iver drainage. This patterning is broadly similar to that seen among ha fted bifaces from the Ocmulgee R iver drainage. As previously sta ted, the nearest source of high quality Coastal Plain chert is located some 83 km to the northwe st (Blanton and Snow 1986). Aucilla Suwannee unifaces Left notch height (1.44) and left shoulder length (0.74) have the highest skewness values among the unifaces examined in the Aucilla Suwannee River drainage. This suggests that a greater percentage of unifaces have these attr ibutes nearer the minimum values. Left notch height and betwee nnotch width have the highest kurtosis values of 3.77 and 1.44, respectively. These two attributes are the least varied among the sample in this drainage. Left notch height has a leptokurtic di stribution, indicating that the majority of data is highly concentrated near the mean due to lower variations. Six of the attributes (left blade length, right blade length, right shoulder length, shoulder width, and base width) have negative kurtosis value s. This suggests that these attributes are highly varied within the sample. Meanwhile, some level of standardization of the

PAGE 198

198 haft element (with the exception of base width) may be indicated by these data, as many attributes of the haft element exhibit a low degree of variation. Tampa Bay unifaces The skewness for left notch height (3.08), right notch height (2.97), and left shoulder length (2.06) are the highest among the attributes recorded for hafted unifaces examined in the Tampa Bay area. These attributes also have the highest kurtosis values (10.08, 9.42, and 5.12, respectively) among the sample examined. These data indicate that a greater percentage of unifaces here have left notch heights, right notch heights, and left shoulder lengths nearer the m inimum values. The recorded values for these three attributes all have leptokurtic distributions and are heavily concentrated near the mean due to few observed variations. This patterning suggests a standardized length of haft element for unifaces in the Tampa Bay area, which is consistent with the evidence previously discussed for hafted bifaces in this area. As previously mentioned, the Tampa Bay area is located within high quality Coastal Plain chert sources. Measured blade elements (left and right blade len gth, and shoulder width) had negative kurtosis value ( 1.52, 0.32, and 1.17, respectively), indicating high variation for these attributes in this region. The blade portions of unifaces in the Tampa Bay region may reflect varying stages of use. The prima ry edge angle (left edge angle for all of the Tampa Bay examined unifaces) also exhibits a high degree of variation; a negative kurtosis value ( 1.14) was obtained for the primary edge angle. This pattern likely is due to the fact that tools could easily b e replaced, since high quality raw materials could readily be obtained. Taken together, trends in data collected for blade elements of unifaces in the Tampa Bay sample suggest there was no need for area flintknappers to meticulously preserve desired blade elements since tools could simply be replenished.

PAGE 199

199 Discussion for Blade Portion of Unifaces Continuous variables relating to the blade elements of hafted unifaces include left and right blade length, shoulder width, and thickness. Blad e lengths for hafted unifaces vary widely across the Southeastern Coastal Plain, and the reasons for this surely are closely tied to use and associated attrition. Through use, the primary edge becomes worn, and the blade necessarily becomes smaller. While no use wear analyses were performed as part of this research, indications of use, including step fracturing and polish, are evident to the naked eye on many examples Table 5 16 present s gross comparative data for uniface left blade lengths. Although the tools are asymmetrical, and right blade length also was recorded for all examples, the left lateral blade edge typically was the obvious primary working edge. Therefore, this measurement reflects the length of the primary working edge, and there is great utility in comparisons of this measurement across the five river drainages that produced more than one example (the Chattahoochee R iver drai nage is exempted from this discussion since it produced only one uniface). Table 516. Summary of left blade length (mm) data for hafted u nifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 4 38.43 19.74 52.58 13.66 0.36 Savannah Ogeechee 144 38.6 16.46 63.6 8.69 0.23 Ocmulgee 7 41.25 30.61 52.89 6.8 0.16 Aucilla Suwannee 38 47.18 35.28 59.11 6.5 0.14 Tampa Bay 12 49.27 39.05 58 6.97 0.14 Total 205 As Table 5 16 illustrates, the lowest mean values for left blade length are found in the SanteeCooper and SavannahOgeechee River drainages. Left blade length for unifaces is seen to

PAGE 200

200 increase incrementally to the south, as the greatest mean values for this attribute are observed in the Aucilla Suwannee and Tampa Bay drainages. D ata from the SanteeCooper River drainage produced the highest coefficient of variation (0.36) of any Coastal Plain data set, illustrating that the most variance for left blade length is found in this sample. This drainage notably also produced the smallest sample size for analysis ( n=4). Meanwhile, the lowest degree of variation is found for samples from the Aucilla Suwannee and Tampa Bay drainages (c.v.=0.14). The range of variance is a significant difference of 0.22, indicating a high degree of dissimilarity in variation from the northernmost and southernmost assemblages in this analysis. I conducted a Levenes test to determine whether the differences in sample variances for uniface left blade length are statistically significant (Gastwirth et al. 2009; Levene 1960). As Levenes Statistic (W) is not greater than the critical value (F) obtained for these data, the null hypothesis cannot be rejected; the inequality of variances for uniface left blade length data is not statistically s ignificant. In fact, no ratioscale uniface data passed Levenes test for equality of variance; observed differences in variance therefore are not significant at a 95% co nfidence interval. The examined sidenotched hafted unifaces all generally fit the type description (Michie 1968) of the Edgefield Scraper, and subregional differences within this artifact class were not necessarily anticipated. Nonetheless, I constructed percentile plots to allow for comparison of numerous data sets on the same graphs Figure 517 contains all raw sample data for uniface left blade length. In general, these data (abbreviated for all percentile plots) fit to one of two groupings; samples fr om the SavannahOgeechee River drainage are more skewed to the left of n (where n is the sample of all drainage data) and i n the second major grouping of data points,

PAGE 201

201 Aucilla Suwannee and Tampa Bay drainage samples are seen to cluster to the right with hi gher overall values for left blade length Several data points from the Ocmulgee River drainage sample span the difference between the left skewed Savannah Ogeechee data and the AucillaSuwa nnee sample. I examined the statistical trends within left blade length data for more intensive analysis of these gross trends in unifaces Data from the Ocmulgee and Tampa Bay drainages exhibit patterned variation warranting additional discussion. In the Ocmulgee River drainage, left blade length is among the least varied in distribution among measured attributes. This is borne out by the relatively high kurtosis value (1.48) previously discussed with respect to this drainage. Left blade length data for Tampa Bay unifaces indicate a high degree of variation for this attr ibute. In fact, these data resulted in the lowest negative kurtosis value ( 1.52) obtained for any measured attribute in this drainage. Table 5 17 presents right blade length data for the five Coastal Plain drainages that produced more than one uniface. I observed the highest mean values for uniface right blade lengths in the Aucilla Suwannee and Tampa Bay drainages. Data from these drainages also resulted in the lowest values for standard deviation. Table 5 17. Summary of right blade lengt h (mm) data for hafted unifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 4 26.98 12.96 42.23 11.98 0.44 Savannah Ogeechee 142 28.42 7.13 60.98 10.21 0.36 Ocmulgee 7 26.19 13.98 31.96 9.49 0.36 Aucilla Suwannee 37 43 29.38 60.02 7.4 0.17 Tampa Bay 12 40.22 25.61 56.79 8.95 0.22 Total 202

PAGE 202

202 Analysis of data for coefficient of variation suggests that the highest degree of variation for right blade length data may be seen in ( n=4) examples from the SanteeCooper River drainage (c.v.=0.44). The least amount of variance is found for samples from the Aucilla Suwannee River drainage (c.v.=0.17). This range of variance is a difference of 0.27, and is indicative of a degree of dissimilarity from north to south previously reported with respect to left blade length. The Savannah Ogeechee and Ocmulgee River drainages display an equal degree of variation in right blade length (c.v.=0.36) despite obvious disparities in minimum and maximum recorded measurements I constructed a percentile plot that contains all r aw sample data for uniface right blade length (Figure 5 18). As was the case for left blade length, these data generally fit one of two groupings, with SavannahOgeechee data points more left skewed relative to n, and Aucilla Suwannee and Tampa Bay samples more skewed to the right with higher overall values. Numerous data points from the Tampa Bay sample can be seen to span part of the difference between the right and left skewed data points, while trending more closely with right skewed data. A reexaminat ion of other specific statistical analyses reveals interesting trends in data from the Ocmulgee and Tampa Bay drainages with respect to right blade length. In the Ocmulgee River drainage, most recorded values for this attribute are nearer the minimum value This is based on a high skewness value (0.55) relative to other measured attributes for unifaces in the Ocmulgee River drainage. Right blade length data for Tampa Bay unifaces reflect a high degree of variation for this continuous variable. A negative kurtosis value ( 0.32) for right blade lengths in Tampa Bay provides evidence for this trend.

PAGE 203

203 Shoulder width is another measurement of the blade portion of side notched hafted unifaces. I expected that this attribute would exhibit similar patterning to lef t blade length, since as that value decreases as a result of tool use and attrition, shoulder width also would decrease. Table 5 18 displays comparative data for shoulder width measurements of unifaces from the five drainages that produced more than one uniface. Table 5 18. Summary of shoulder widt h (mm) data for hafted unifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 4 32.99 23.42 42.09 7.79 0.24 Savannah Ogeechee 141 33.13 20.39 54.43 6.2 0.19 Ocmulgee 8 34.82 26.82 40.83 5.03 0.14 Aucilla Suwannee 32 36.51 26.55 45.01 4.41 0.12 Tampa Bay 12 39.11 28.84 46.95 6.43 0.16 Total 197 As expected, patterned variation in shoulder width is broadly similar to that observed for left blade length. The lowest mean shoulder width values are seen for unifaces from the Santee Cooper and SavannahOgeechee River drainages; the highest mean values for this attribute are seen in the Aucilla Suwannee and Tampa Bay drainages. Coefficient of variation data indic ate that the widest range of variation for uniface shoulder widths is in the SanteeCooper River drainage (c.v.=0.24). As previously stated, the most relative variance also was observed for right and left blade length data from the SanteeCooper River drai nage. These data point to a high degree of variation for the blade portion of unifaces from the Santee Cooper River drainage, although the sample size in consideration is small ( n=4). As both the Santee Cooper and the Ocmulgee River drainages are sit uated away from sources of high quality Coastal Plain cherts, this patterned variation may reflect efforts to curate tools that could not be easily replenished. In fact, it is rather surprising that unifaces from

PAGE 204

204 the Ocmulgee River drainage do not display even more variation in terms of this particular continuous variable. The least amount of variance for uniface shoulder widths is found in the Aucilla Suwannee sample (c.v.=0.12). Figure 5 19 is a percentile plot c ontaining all raw sample data for uniface shoulder width. In general, data from the five drainages display a broad distribution; most samples from the Savannah Ogeechee River drainage are more skewed to the left of n, and Aucilla Suwannee and (a majority o f) Tampa Bay drainage samples are seen to cluster to the right with progressively higher overall values Data collected for the Ocmulgee River drainage are seen to share a trend with Aucilla Suwannee data at and below the median, although these data are ch aracterized by a considerable spread. I reexamined some of the statistical analyses first discussed with respect to individual drainages, to learn more about patterned variation in uniface shoulder width. D ata from two drainages were found worthy of furthe r discussion; the Ocmulgee and Tampa Bay. These same drainages also had been marked for special consideration with respect to left blade length. A high degree of variation was found to exist for uniface shoulder width in the Ocmulgee River drainage. In fac t, this attribute is among the most varied in distribution of metric attributes, a claim supported by a negative kurtosis value ( 0.46). Most values for uniface shoulder width are nearer the maximum value in the Ocmulgee River drainage. Unifaces from Tampa Bay also had a high degree of variation for shoulder width, indicated by a negative kurtosis value ( 1.17). Another attribute reflective of the blade portion of unifaces is the primary edge angle. This continuous variable, measured in degrees, was recorde d for unifaces from six river drainages. Table 5 19 presents data collected from those drainages that produced more than one

PAGE 205

205 uniface, for comparative purposes. Data presented in Table 519 are only for left primary edge angle, as all of these river drainag es produced examples of right handed unifaces. For right handed unifaces ( n =231), t he widest range of variation for uniface primary edge angle is in the Ocmulgee River drainage (c.v.=0.27), while the least amount of variance is found in the SanteeCooper s ample (c.v.= 0.1). This range of variance is a difference of 0.17, which is particularly notable as it was observed for equally chert poor drainage locations. Table 5 19. Summary of primary edge angl e (degrees) for hafted unifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 4 52.5 45 55 5 0.1 Savannah Ogeechee 166 51.84 20 85 13.59 0.26 Ocmulgee 8 41.25 20 50 10.94 0.27 Aucilla Suwannee 39 62.56 40 85 9.45 0.15 Tampa Bay 14 57.86 40 75 12.04 0.21 Total 231 It should be noted that the Ocmulgee sample ( n=8) is comprised solely of unifaces recovered from the Feronia site complex in Coffee County, Georgia. It seems that the Feronia site complex defies some of the assumptions made by researchers who study the Ear ly Archaic period. The sites are located approximately 83 km a way from high quality raw material, yet they clearly were intensively occupied, and, based on the relatively high frequency of hafted unifaces recovered at the locale, a number of specific activ ities were performed ther e (after Blanton and Snow 1986, 1989) Discussion of the apparent attraction for hunter gatherers to Feronia, in spite of the distance they had to travel to acquir e decent quality raw material, is provided in Chapter 6. For the purposes of the present discussion, I suggest that the high degree of variation in primary edge angles observed for Ocmulgee River drainage (i.e., Feronia ) unifaces belies the expected result s of that distance. I observed the lowest mean and maximum values for primary edge angle

PAGE 206

206 in the Ocmulgee sample, which may indicate an effort on the part of Feronia toolmakers to conserve what chert they had, and to use the tools more intensively than was evidenced at other locales (such as the Savannah Ogeechee and AucillaSuwan nee River drainages, where high quality cherts wer e readily available). As noted above, primary edge angle data for the SanteeCooper River drainage are indicative of a low degree of variation (c.v.=0.1) compared with data for other drainages. A closer inspection of data for primary edge angles in the SanteeCooper River drainage reveals that values for this attribute are highly concentrated around the mean, with a range of 45 to 55 Much like the Ocmulgee River drainage, the SanteeCooper region is located far (approximately 97 km ) from high quality chert raw material ( Figure 53). More variation might therefore have been expected for samples from this region. However apparent trends observed for primary edge angles in the SanteeCooper River drainage may be a function of the small sample size examined for this river drainage ( n =4). The Savannah Ogeechee River drainage is the only region that produced examples ( n=7) of left handed unifaces, where the primary working edge is on the right lateral tool edge. Table 520 displays descriptive statistics for lefthanded unifaces from the SavannahOgeechee River drainage. Left handed unifaces displayed moderate variation (c.v.=0.16) compared to all drainages, but notably less variation than right handed examples from the same (Savannah Ogeechee River) drainage where they were examined. Table 5 20. Summary of primary edge angle (degrees) for left handed hafted unifaces Drainage n Mean Min Max s.d. c.v. Savannah Ogeechee 7 70 45 75 11.55 0.16

PAGE 207

207 Maximum thickness is the final blade element attribute to be discussed with respect to the five Coastal Plain river drainages that produc ed multiple unifaces. Table 5 21 presents comparative data collected for this attribute. Coefficient of variation data indicate that the widest range of variation for uniface thickness is in the SanteeCooper River drainage (c.v.=0.25). As discussed above, I also obtained high er coeffici ents of variation for right and left blade length and shoulder width data from the Santee Cooper River drainage compared to other drainages Collectively, these data point to a high degree of variation for the blade portion of Santee Cooper River unifaces, and may reflect efforts to curate tools in a chert poor locale. Here, it is worth repeating that it is difficult to describe meaningful trends in ratio scale data with such a small uniface sample size ( n=4). Table 5 21. Summary of maximum thicknes s (mm) data for hafted unifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 4 10.6 7.09 12.96 2.66 0.25 Savannah Ogeechee 186 8.78 5.05 13.2 1.79 0.2 Ocmulgee 9 9.34 5.39 11.79 2.05 0.22 Aucilla Suwannee 39 9.31 5.59 11.79 1.41 0.15 Tampa Bay 14 9.89 7.12 13.7 2.04 0.21 Total 252 The least amount of variance for maximum uniface thickness is found in the Aucilla Suwannee sample (c.v.=0.15). As I did for all other blade element attributes, I attempted to locate additional statistical trends in data for uniface thickness. None were readily apparent, in terms of skewness and kurtosis tests performed for this attribute. I constructed a percentile plot that contains all raw sample data for maximum uniface thickness (Figure 520). The distribution of data from all five drainages is broad, with a majority of samples from the SavannahOgeechee River drainage skewed to the left of n, and most

PAGE 208

208 Aucilla Suwannee and Tampa Bay drainage samples skewing to the right with higher overall values Above the 70t h percentile, several Aucilla Suwannee data points can be seen to share the same trend as Savannah Ogeechee data. Meanwhile, data collected for the Ocmulgee River drainage are characterized by a considerable spread; at and above the median these data are s kewed to the right of the Aucilla Suwannee River drainage sample. Figure 5 21 illustrates trends for uniface blade length, shoulder width, and maximum thickness across the Southeastern Coastal Plain. Figure 5 22 illustrates the same information, in the con text of the Coastal Plain, and with reference to specific measured values. As Figure s 5 21 and 522 show, the highest mean left and right blade lengths and shoulder widths are seen for unifaces from the Aucilla Suwannee and Tampa Bay drainages. As previous ly discussed, mean left blade lengths for hafted unifaces from the SanteeCooper, Savannah Ogeechee, and Ocmulgee River drainages range from 38.43 to 41.25 mm. Mean left blade lengths for unifaces from the Aucilla Suwannee and Tampa Bay drainages range fro m 47.18 to 49.27 mm. Mean right blade lengths for unifaces from the Santee Cooper, SavannahOgeechee, and Ocmulgee River drainages range from 26.19 to 28.42 mm. Mean right blade lengths for unifaces from the Aucilla Suwannee and Tampa Bay drainages range f rom 40.22 to 43.0 mm. Mean shoulder widths for unifaces from the SanteeCooper, SavannahOgeechee, and Ocmulgee River drainages range from 32.99 to 34.82 mm. I recorded mean shoulder widths for hafted unifaces from the Aucilla Suwannee and Tampa Bay draina ges ranging from 36.51 to 39.11 mm. I observed an interesting shift for data that occurs between the Ocmulgee and the Aucilla Suwannee River drainages. South of the Ocmulgee River drainage, mean values for right blade length are greater than those record ed for mean shoulder width ( Figure 5 21). Meanwhile,

PAGE 209

209 maximum thicknesses for hafted unifaces from throughout the Coastal Plain display different patterned variation compared to the other portions of the blade examined for this research, in that high mean val ues are not restricted to a particular geographic region. I observed mean values for maximum thickness ranging from 8.78 mm in the SavannahOgeechee sample, to 10.6 mm in the SanteeCooper sample. Hafted unifaces I examined from the Aucilla Suwannee and Ta mpa Bay drainages have higher means for left and right blade length and shoulder width than I observed elsewhere These trends mirror those seen for hafted bifaces in this study, and may be interpreted as a further expression of a social boundary located a long Floridas Gulf Coast, as well as throughout north and northcentral Florida. Trends in hafted uniface blades also may be related to environmental constraints, as well as to a different mobility pattern in Florida than elsewhere. As previously di scusse d, potable water and highquality raw material likely were important criteria in Early Archaic settlement strategies. Early Archaic Floridians dealt with a significantly reduced water table as a result of the Bolen Drought (Dunbar 2002). Additionally, cher ts found in Floridas Tertiary karsts may have come in larger original package sizes as compared to elsewhere in the Coastal Plain, resulting in comparatively larger sized blades for side notched hafted unifaces produced in the Florida peninsula. In Chapte r 6, I explore the implications of this patterned variation for modeling Early Archaic settlement and social interactions. Discussion for Haft Element of Unifaces As previously discussed, t he metric variables I studied relating to uniface haft el ement include: left and right shoulder length, betweennotch width, left and right notch height, and base width. As the Flint River drainage produced no unifaces, and the Chattahoochee River drainage

PAGE 210

210 produced only one uniface, those regions are not discuss ed below with respect to uniface haft element trends. Comparative data for uniface left shoulder length are presented below, in Table 5 22. The highest mean values for left shoulder length are seen for unifaces from the Ocmulgee, Aucilla Suwannee, and Tampa Bay drainages (18.22, 16.77, and 17.2 mm, respectively). I observed the lowest mean values for left shoulder length for unifaces from the SanteeCooper and SavannahOgeechee River drainages (13.78 and 14.79 mm, respectively). Table 5 22. Summary of left shoulder length (mm) data for hafted unifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 4 13.78 8.69 17.66 7.77 0.56 Savannah Ogeechee 161 14.79 4.88 45.51 4.97 0.34 Ocmulgee 8 18.22 12.52 22.28 3.73 0.2 Aucilla Suwannee 31 16.77 12.69 22.51 2.51 0.15 Tampa Bay 12 17.2 12.21 31.12 5.09 0.3 Total 216 Figure 5 23 is a percentile plot of uniface left shoulder length. At and below the median, Savannah Ogeechee River drainage data are left skewed, and the second major grouping consists of data from the Tampa Bay, Aucilla Suwannee, and Ocmulgee drainages, which can be seen to share a similar trend with higher values. Above the median, Ocmulgee River drainage data are the most rightskewed, while data points from the Tampa Bay, Aucilla Suwannee, and Savannah Ogeechee drainages largely share a similar trend below the 80th percentile. Analysis of data for coe fficient of variation suggests that the highest degree of variation for left shoulder length data may be seen in examples ( n=4) from the Santee Cooper River drainage (c.v.=0.56). The least amount of variance among the examined assemblages is found for samp les from the Aucilla Suwannee River drainage (c.v.=0.15). The difference between

PAGE 211

211 coefficients of variation for these assemblages is 0.41, and is indicative of a degree of dissimilarity previously reported with respect to aspects of uniface blades from thes e assemblages. A majority of the drainages that produced more than one uniface for this study are remarkable for specific statistical trends respective to left shoulder length. A greater percentage of unifaces from the SavannahOgeechee River drainage are nearer the minimum value for this continuous variable. Additionally, left shoulder length is among the least varied in distribution among measured attributes for this drainage. Left shoulder length data for unifaces from the Savannah Ogeechee River drainag e are heavily concentrated around the mean. Statistical analyses of data for unifaces from the Ocmulgee River drainage reveal that left shoulder length is among the most varied in distribution of measured attributes. As is true of all metric data for unifa ces from the Ocmulgee River drainage, a high degree of variation is evident for left shoulder length data. A greater percentage of uniface left shoulder lengths from the Aucilla Suwannee and Tampa Bay drainages are nearer the minimum values for that attrib ute. Additionally, in Tampa Bay, these data are heavily concentrated around the mean, with few observed variations. Table 5 23 presents right shoulder length data for each of the river drainages that produced multiple unifaces. I observed the highest mean values for right shoulder length for unifaces from the Ocmulgee, AucillaSuwannee, and Tampa Bay drainages. These data range from 17.51 mm for the Aucilla Suwannee unifaces, to 18.51 mm for the Ocmulgee unifaces. The lowest mean values are seen for right s houlder lengths in unifaces from the Santee Cooper and Savannah Ogeechee River drainages (16.82 and 15.64 mm, respectively ). These trends are very

PAGE 212

212 similar to those observed with respect to left shoulder length data throughout the examined Coastal Plain dra inages. Table 5 23. Summary of right shoulder lengt h (mm) data for hafted unifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 3 16.82 11.52 24.96 4.78 0.28 Savannah Ogeechee 155 15.64 6.94 39.76 4.95 0.32 Ocmulgee 8 18.51 11.38 21.78 3.76 0.2 Aucilla Suwannee 36 17.51 12.37 24.35 2.84 0.16 Tampa Bay 12 17.71 13.86 27.17 3.93 0.22 Total 214 I constructed a percentile plot that contains all raw sample data for right shoulder length. Figure 5 24 presents two major data groupings at and below the median; SavannahOgeechee River drainage data are left skewed, while Tampa Bay and Aucilla Suwannee River drainage data can be seen to share a similar trend with higher values. Above the 30th percentile, the major ity of right shoulder length data for the Ocmulgee River drainage are right skewed, with higher values compared to data from AucillaSuwannee. Two data points from the Tampa Bay sample can be seen at the top right of the percentile plot for right shoulder length. As Table 5 23 demonstrates, data from the Savannah Ogeechee River drainage produced the highes t coefficient of variation (0.32) of any Coastal Plain data set, illustrating a high degree of variation for right shoulder length Meanwhile, the least amount of variance is found for samples from the Aucilla Suwannee River drainage (c.v.=0.16). The Savannah Ogeechee and the Ocmulgee River drainages both produced interesting results in response to other statistical tests performed for right shoulder length data. In the Savannah Ogeechee River drainage, a greater percentage of unifaces are nearer the minimum value for right shoulder length. Right shoulder length is among the least varied in distribution

PAGE 213

213 among measured attributes in the SavannahOgeechee River drainage. These claims are supported by high values for both skewness and kurtosis for right shoulder length for unifaces from this drainage. Coefficient of variation tests alone had indicated that right shoulder length for unifaces from the SavannahOge echee River drainage was characterized by relatively high variation; additional statistical tests allowed me to more fully describe the extent of that variation, which is best understood as a concentration of data nearer the minimum for this particular att ribute. Patterned variation of right shoulder length data for unifaces from the Savannah Ogeechee River drainage may reflect some level of haft element standardization, or the presence of a high percentage of unifaces n ear the end of their use lives. Data from the Ocmulgee R iver drainage indicate that, while the majority of uniface right shoulder lengths are nearer the maximum value, a high degree of variation is evident for this continuous variable; this is because the distribution of all metric attributes taken for Ocmulgee R iver drainage unifaces is pla t ykurtic. These trends likely reflect an interest in tool curation and raw material conservation for unifaces in the Ocmulgee River drainage. Additionally, it may be possible that relatively large (with res pect to haft element) unifaces were brought to the Ocmulgee River drainage as part of a hunter gatherers standard toolkit, in preparation for spending any period of time in an area devoid of highquality chert raw material. Perhaps intensive use of unifac es was neither anticipated nor common practice by visitors to the Ocmulgee River drainage, who typically made short term and temporary trips to and through the area. Further suggestions about the social boundaries that may be reflected by the Ocmulgee sidenotched artifacts are provided in Chapter 6.

PAGE 214

214 Table 5 24 presents between notch width data for each of the river drainages that produced multiple unifaces. The highest mean values for between notch width are seen for unifaces from the Santee Cooper and Ocm ulgee River drainages (24.14 and 24.73 mm, respectively). It bears repeating that these two drainages are located at some distance from high quality chert raw material. Perhaps Early Archaic visitors to these drainages arrived with unifaces manufactured wi th wide haft elements, in anticipation of the need for tool curation and raw material conservation during their travels. Table 5 24. Summary of between notch width (mm) data for hafted unifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 4 24.14 15.82 28.89 5.76 0.24 Savannah Ogeechee 182 21.81 11.32 40.05 4.82 0.22 Ocmulgee 9 24.73 15.72 31.45 4.44 0.18 Aucilla Suwannee 38 20.39 13.56 30.38 3.3 0.16 Tampa Bay 14 22.61 14.79 28.69 4.91 0.22 Total 247 Figure 5 25 is a percentile pl ot illustrating uniface between notch width; s amples from all five drainages genera lly fit to one of two groupings. Data from the Savannah Ogeechee and Aucilla Suwannee River drainages are more skewed to the left of n with a relatively distinct (i.e., shar ed) trend; several Tampa Bay data points also are part of this shared trend. In the second major grouping of data, a majority of Ocmulgee, Tampa Bay and Santee Cooper drainage samples are seen to cluster to the right with their higher overall values near and above the median Coefficient of variation data indicate that the widest range of variation for uniface between notch width is in the Santee Cooper River drainage (c.v.=0.24). The lowest relative variance for uniface between notch width is found in the Aucilla Suwannee sample (c.v.=0.16).

PAGE 215

215 Meanwhile, the same coefficient of variation value (0.22) was obtained for data from the Savannah Ogeechee and Tampa Bay drainages. Once again, it is important to recall that the sample size for SanteeCooper unifaces is relatively small ( n=4), lest its relative variance be unduly overemphasized. A reexamination of other specific statistical analyses indicates that the SanteeCooper, Ocmulgee, and Aucilla Suwannee River drainages all exhibit patterne d variation for uniface between notch width that requires some additional discussion. In these three river drainages, between notch width is among the least varied in distribution among measured attributes. This is borne out by the high kurtosis values (2.66, 1.63, and 1.05, respectively) for between notch width data in these river drainages. Gross patterned variation in left notch height is presented in Table 5 25. The highest mean values for left notch height are seen for unifaces from the Santee Cooper, Aucilla Suwanne e, and Tampa Bay drainages (9.29. 9.49, and 9.65 mm, respectively). The lowest mean and minimum values for this measured attribute are seen for the SavannahOgeechee and Ocmulgee River drainages. Table 5 25. Summary of left notch height (mm) data for ha ft ed unifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 4 9.29 6.62 12.76 3.11 0.33 Savannah Ogeechee 148 8.05 2.99 17.91 2.5 0.31 Ocmulgee 8 8.04 5.64 11.24 1.95 0.24 Aucilla Suwannee 34 9.49 6.29 17.08 2.13 0.22 Tampa Bay 12 9.65 6.36 23.24 4.42 0.46 Total 206 Figure 5 26 is a percentile plot that contains all raw sample data for uniface left notch height, allowing comparison of numerous data sets on the same graph. In general, all of the

PAGE 216

216 drainages fit to one of two groupings at and below the median; data from the SavannahOgeechee and Ocmulgee River drainages are more skewed to the left of n, while data from the Tampa Bay and AucillaSuwannee drainages are right skewed with higher values. At and above the 70th percentile, Tampa Bay, Ocmulgee, Savannah Ogeechee, and Aucilla Suwannee data points largely share the same trend for left notch height. Analysis of data for coefficient of variation suggests that the highest degree of variation for left notch height data may be seen in examples from Tampa Bay (c.v.=0.46). The lowest degree of variation is found for samples from the Aucilla Suwannee River drainage (c.v.=0.22). This range of variance is a difference of 0.24, and is notable due to the similar mean values obtained for these samples, as well as the geographical proximity of the assemblages. Additional trends in left notch height data were revealed by tests for skewness and kurtosis, first discussed with respect to individual drainages. Four drainages produced patterned variation for l eft notch height in response to these tests; the Santee Cooper, the Ocmulgee, the Aucilla Suwannee, and Tampa Bay. A high degree of variation for left notch height is indicated for unifaces from the SanteeCooper River drainage. This claim is supported by a low negative kurtosis value ( 4.56) for this attribute. In the Ocmulgee River drainage, left notch height is among the most varied in distribution among measured attributes. Left notch height data for unifaces from the Aucilla Suwannee and Tampa Bay drai nages pattern much differently. In those drainages, a greater percentage of unifaces are nearer the minimum value for this attribute. Also, the majority of left notch heights recorded for unifaces from the Aucilla Suwannee and Tampa Bay drainages are highl y concentrated near the mean values.

PAGE 217

217 Table 5 26 presents comparative data for right notch height in unifaces recovered from five Coastal Plain drainages. I observed the highest mean right notch heights for unifaces in the Ocmulgee, Aucilla Suwannee, and Ta mpa Bay drainages. The lowest relative mean, minimum, and maximum values all are seen for right notch heights in the Santee Cooper River drainage. Table 5 26. Summary of right notch height (mm) data for hafted unifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 4 8.12 3.93 13.83 4.19 0.52 Savannah Ogeechee 163 8.29 4.02 16.44 2.55 0.31 Ocmulgee 8 9.18 6.34 12.76 2.43 0.26 Aucilla Suwannee 38 9.67 5.87 16.51 2.33 0.24 Tampa Bay 12 9.53 7.55 19.17 3.18 0.33 Total 225 I constructed a percentile plot that contains all raw sample data for right notch height (Figure 5 27). Samples from the five drainages generally f it to one of two modes or peaks; the Savannah Ogeechee River drainage sample is left skewed with lower relati ve values, and the Aucilla Suwannee data are more right skewed. A majority of Santee Cooper River drainage data points share a similar trend with the SavannahOgeechee. At and above the median, several Tampa Bay data points share the same trend as data fro m the Savannah Ogeechee River drainage. Meanwhile, most data samples from the Ocmulgee River drainage span the difference between the Savannah Ogeechee and AucillaSuwannee River drainage data. Coefficient of variation data indicate that the most relative variance for uniface right notch height is in the Santee Cooper River drainage (c.v.=0.52). The least amount of variation for this measured attribute is found in the Aucilla Suwannee sample (c.v.=0.24). This range of variance is a difference of 0.28, and is indicative of a degree of dissimilarity from the northernmost assemblage in this study (in an area devoid of local highquality chert), to one of

PAGE 218

218 the southernmost samples in consideration, where cherts were locally available. Similar patterning was prev iously reported with respect to trends in uniface left and right blade length. In the Santee Cooper River drainage, a greater percentage of unifaces have right notch heights nearer the minimum value. This trend also is seen for data from the Ocmulgee and T ampa Bay drainages. Right notch height data for unifaces from Tampa Bay are heavily concentrated around the mean due to few observed variations. These claims are supported by statistical tests for skewness and kurtosis for right notch height data from all examined unifaces throughout the Coastal Plain. Table 5 27 presents comparative data for base width. Patterned variation for unifaces from five river drainages is included in these data. Mean values for uniface base width in the five examined drainages ran ge from a minimum (29.81 mm) in the Aucilla Suwannee River drainage, to a maximum (32.93 mm) in Tampa Bay. This relatively tight clustering of mean base width values across the Coastal Plain is belied by the broad range of recorded values, ranging from a m inimum (14.64 mm) to a maximum (57.1 mm), both values of which were recorded for unifaces in the Savannah Ogeechee River drainage. Table 5 27. Summary of base width (mm) data for hafted unifaces Drainage n Mean Min Max s.d. c.v. Santee Cooper 4 31.62 20.64 38.83 7.81 0.25 Savannah Ogeechee 160 29.88 14.64 57.1 7.36 0.25 Ocmulgee 8 32.93 20.26 44.32 7.39 0.22 Aucilla Suwannee 32 29.81 22.08 38.25 4.43 0.15 Tampa Bay 11 32.45 23.22 39 5.43 0.17 Total 215 As Table 5 27 demonstrates, data from the Santee Cooper and Savannah Ogeechee River drainage s produced the highest coefficient s of variation (0.25) of any Coastal Plain data set,

PAGE 219

219 exhibiting the most relative variance for uniface base width. As discussed above, I also obtained high er coefficients of variation for left shoulder length and right notch height data from the SanteeCooper River drainage compared to other assemblages Collectively, these data point to a high degree of variation for the haft element of SanteeCooper River unifaces ( n=4) and may reflect efforts to curate tools in a chert poor locale. Alternatively, these trends may be a function of low sample size for the SanteeCooper River drainage ( n =4). Meanwhile, the lowest degree of variation is found for samples from the Aucilla Suwannee River drainage (c.v.=0.15). Other specific statistical analyses of data for uniface base width (including tests for skewness and kurtosis) did not yield meaningful results for the purposes of this discussion. Figure 5 28 presents a percentile plot of all raw sample data for uniface base width. In general, all data points can be seen to share a similar trend beginning around the 30th percentile. Beginning above the 70th percentile, data points for the Aucilla Suwannee River drainage become more left skewed relative to the SavannahOgeechee River drainage trend. The clustering of uniface base width data into a single broad trend between the 30th and 70th percentiles, is similar to the patterning previou sly observed for biface base width. Figure 5 29 shows all mean left a nd right notch heights, betweennotch widths, left and right shoulder lengths, and base widths for unifaces from throughout the Coastal Plain. As Figure 529 shows, the highest mean left and right notch heights are seen for hafted un ifaces from the Aucilla Suwannee and Tampa Bay drainages. As previously discussed, mean left notch heights for unifaces from the SanteeCooper, Sa vannah Ogeechee, and Ocmulgee Ri ver drainages range from 8.04 to 9.29 mm. Mean left notch heights for unifaces from the Aucilla Suwannee and Tampa Bay drainages range from 9.49 to 9.65 mm. Mean right notch heights for unifaces from

PAGE 220

220 the SanteeCooper, Sa vannahOgeechee, and Ocmulgee Ri ver drainages range from 8.12 to 9.18 mm. Mean right notch heights for unifaces from the Aucilla Suwannee and Tampa Bay drainages range from 9.53 to 9.67 mm. Mean between notch widths for unifaces do not exhibit similar patterned variation in that high values are not restricted to a particu lar geographical region. I observed values ranging from 20.39 to 27.73 mm for between notch width, recording the highest mean values for data from the Santee Cooper and Ocmulgee River drainages. The highest mean values for left and right shoulder length ar e seen for hafted unifaces from the Ocmulgee and Tampa Bay drainages. Mean left shoulder lengths for hafted unifaces from the SanteeCooper, SavannahOgeechee, and Aucilla Suwannee R iver drainages range from 13.78 to 16.77 mm. Mean left shoulder lengths for hafted uni faces from the Ocmulgee and Tampa Bay dr ainages range from 17.2 to 18.22 mm. Mean right shoulder lengths for hafted unifaces from the SanteeCooper, SavannahO geechee, and Aucilla Suwannee Ri ver drainages range from 15.64 to 17.51 mm. I recorde d mean values of 18.51 and 17.71 mm for right shoulder length in unifaces from the Ocmulgee and Tampa Bay drainages. The highest mean base widths are seen for unifaces from the Ocmulgee and Tampa Bay drainages (32.93 and 32.45 mm, respectively). Mean base widths for unifaces from the other examined drainages ranged from 29.81 to 31.62 mm. Comparing Side N otched Hafted Bifaces to Edgefield Scrapers across Drainages The primary research question in this dissertation relates to the identification of social boundaries that may be manifested in patterned variation in Early Side Notched Horizon material culture. Early Archaic sidenotched technology was a social production, and thus a reflection of the varying social contexts of its production and reproduction (after Lemonnier 1992). As such, I

PAGE 221

221 would argue that the subregional technological traditions practiced by Early Archaic groups are reflective of shared knowledge, as well as interaction. I used patterned variation in two classes of Coastal Plain chert side notched artifacts from seven physiographic regions to address this research question. In this section, I relate the numerous comparisons I have been able to make be tween these two classes of data. I examined patterned v ariation in nominal and ratio scale attributes, and m apping trends in basal configuration for all side notched artifacts (hafted bifaces and unifaces) led to some interesting findings. I rather expected that uniface base shapes would mirror those seen for hafted bifaces in the same respective drainages. However, this patterning is markedly different than what I observed for side notched hafted bifaces. Table 528 shows these data. It bea rs repeating here that only one uniface was e xamined from the Chattahoochee R iver drainage, and no unifaces were examined from the Flint River drainage. Table 5 28. Trends in basal configuration for hafted bifaces and unifaces Drainage Most Common Basal C onfiguration Second Most Common Basal Configuration Hafted Bifaces Hafted Unifaces Hafted Bifaces Hafted Unifaces Santee Cooper Incurvate Straight Straight Excurvate Savannah Ogeechee Incurvate Straight Straight Incurvate Ocmulgee Straight Excurvate Incurvate Incurvate Flint Straight N/A Incurvate N/A Chattahoochee Straight Excurvate Incurvate N/A Aucilla Suwannee Straight Excurvate Excurvate Straight Tampa Bay Straight Straight Excurvate Excurvate What I found cannot easily be explained; the Tampa Bay region is the only place in my study area where basal configuration of unifaces is the same as that of side notched hafted bi faces. In Tampa, straight followed by excurvate, was the most common base form for both

PAGE 222

222 types of side notched tools. In all other drainages I examined, the base forms of unifaces can be seen to differ significantly from those of side notched hafted bifaces. The key finding with respect to basal configuration of side notched artifacts has been that base shape appears to be directly correlated to geography throughout the Coastal Plain. This is an original contribution to Southeastern archaeology, and it applies to the basal configuration of both hafted bifaces and unifaces. Describing these trends in basal configuration is instrumental to placing patterned variation in Early Archaic material culture in geographical context. Eight portions of the data presented in Table 5 28 can be parsed out to illustrate the division of basal configuration data into spatial categories that cr oss cut contiguous Coastal Plain drainages. These categories are demarcated by the intersection of the Savannah Ogeechee and Ocmulgee River drainages, and the intersection of the Aucilla Suwannee River drainage with the Ocmulgee, Flint, and Chattahoochee R iver drainages. The specific natures of these intersections are detailed below. I identified eight instances of drainage contiguity throughout the Coastal Plain ( Figure 4 3): (1) Santee Cooper and Savannah Ogeechee; (2) Savannah Ogeechee and Ocmulgee; (3) Ocmulgee and Flint; (4) Flint and Chattahoochee; (5) Ocmulgee and Aucilla Suwannee; (6) Flint and Aucilla Suwannee; (7) Chattahoochee and Aucilla Suwannee; and (8) Aucilla Suwannee and Tampa Bay. I then examined data presented in Table 5 28 to determine what data for basal configuration could be aggregated into multi drainage subregions. Three traits are shared by the SanteeCooper and SavannahOgeechee River drainages. The most common base shape for hafted bifaces is incurvate, and the most common base sha pe

PAGE 223

223 for unifaces is straight. The second most common base shape for hafted bifaces in the Santee Cooper and SavannahOgeechee River drainages is straight. The Ocmulgee River drainage shares one trait with the Savannah Ogeechee, and two traits with the Flin t and Aucilla Suwannee River drainages. The second most common type of base shape observed for unifaces in the Ocmulgee and Savannah Ogeechee River drainages is incurvate. Straight is the most common base shape for hafted bifaces in the Ocmulgee, Flint, and Aucilla Suwannee River drainages, and excurvate is the most common base shape for unifaces in the Ocmulgee and AucillaSuwannee River drainages. Incurvate is the second most common type of basal configuration for hafted bifaces in the Ocmulgee and Flint River drainages. Two traits are shared by the Flint and Chattahoochee River drainages. In both drainages, straight is the most common base shape for hafted bifaces, and incurvate is the second most common type of biface basal configuration. One common trai t is found in the Flint and Aucilla Suwannee River drainages. Once again, it is biface base shape, which is predominantly straight. As previously stated, no unifaces from the Flint River drainage were examined. Given that the highest frequencies of unifaces from adjacent river drainages (the Ocmulgee, Chattahoochee, and Aucilla Suwannee) have excurvate basal configuration, it seems very likely that unifaces recovered from the Flint River also would tend to have excurvate basal configuration. The Chattahoochee and AucillaSuwannee River drainages share two traits. In both drainages, the most common base shape for bifaces is straight, and the most common type of basal configuration for unifaces is excurvate. As previously discussed, the Chattahoochee drainage also shares common traits with the Flint River drainage.

PAGE 224

224 Finally, the Aucilla Suwannee River drainage shares two traits with Tampa Bay. The most common basal configuration for hafted bifaces in both the Aucilla Suwannee and Tampa Bay drainages is straight. This trait also is shared with hafted bifaces from the Ocmulgee River drainage, as previously noted. The second most common type of basal configuration for hafted bifaces from both the Aucilla Suwannee and Tampa Bay drainages is excurvate. As displayed in Table 5 28, 28 entries were needed, and it is possible to group 21 (75% ) of these into multidrainage subregions. Three of the 28 entries are N/A; these can effectively be subtracted, bringing the total to 25 out of 28 (approximately 89% ) entries that c an be grouped into multidrainage subregions. Data that cannot be grouped at the multi drainage level are limited to the following: most common basal configuration for unifaces from Tampa Bay, and second most common base shape for unifaces from the Santee Cooper, Aucilla Suwannee, and Tampa Bay drainages. I assigned an arbitrary numerical value (1 4) to each category constructed for Table 5 28. Given the fact that bifaces are much more common than unifaces in the Early Archaic toolkit, I assigned more weigh t to the basal configuration of hafted bifaces, whether it was the most or the second most com mon type of basal configuration. The most common type of biface basal configuration in a multidrainage subregion was assigned the value 4; the second most common base type for bifaces was assigned the value 3; the most common type of uniface basal configuration in a particular multidrainage subregion was assigned the value 2; and the second most common base type for unifaces was assigned the value 1. This informa tion generated the values presented in Table 529.

PAGE 225

225 Table 5 29. Trends in basal configuration for hafted bifaces and unifa ces by multidrainage subregion Multi drainage Subregion Most Common Basal Configuration Second Most Common Basal Configuration Hafted Bifaces Hafted Unifaces Hafted Bifaces Hafted Unifaces Total Santee Cooper/Savannah Ogeechee 4 2 3 9 Savannah Ogeechee/Ocmulgee 1 1 Ocmulgee/Flint 4 3 7 Flint/Chattahoochee 4 3 7 Ocmulgee/Aucilla Suwannee 4 2 6 Flint/Aucilla Suwannee 4 4 Chattahoochee/Aucilla Suwannee 4 2 6 Aucilla Suwannee/Tampa Bay 4 3 7 The ultimate goal of this particular analytical tool is to gauge the level of cross drainage transhumance using basal configuration as a proxy for social group affiliation. I provide discussion of that effort in Chapter 6. For the purposes of the present discussion, it is most important to note the numerical values generated for each of the eight multi drainage subregions (seen in the Total column). The highest t otal values in Table 529 each are a function of the sum of assigned (weighted) values for specific shared traits. For example, the highest total value (9) is seen for the SanteeCooper/Savannah Ogeechee subregion. This reflects that the most common base s hape for bifaces in both drainages is incurvate. This shared trait has a value of 4. The most common base shape for unifaces in both drainages is straight, resulting in a value of 2. Finally, the second most common type of basal configuration for bifaces i n both drainages is straight, which is given a value of 3. The Santee Cooper/SavannahOgeechee subregion is remarkable for

PAGE 226

226 the presence of shared traits for biface basal configurations, and data from this multi drainage subregion are represented by the hig hest total seen for any examined data class. The next highest total value (7) is seen for trends in basal configuration for the Ocmulgee/Flint, Flint/Chattahoochee, and Aucilla Suwannee/Tampa Bay multi drainage subregions. A slightly lower total value (6) resulted for the Chattahoochee/Aucilla Suwannee and Ocmulgee/AucillaSuwannee subregions. The Flint/Aucilla Suwannee multi drainage subregion resulted in a still lower total value (4). I observed the lowest total value (1) for basal configuration trends in multidrainage subregions for the SavannahOgeechee/Ocmulgee, where the only shared trait is the second most common base shape for unifaces. Table 5 29 is a useful heuristic device in determining which of the eight examined multi drainage subregions exhibit the most and least similarities for trends in side notched artifact basal configuration. The highest degree of similarity is seen for the Santee Cooper/SavannahOgeechee. The Ocmulgee/Flint, Flint/Chattahoochee, and AucillaSuwannee/Tampa Bay subregions exhibit somewhat less similarity for basal configuration, but to the same extent as one another. The least similarity in terms of trends in basal configuration is seen for the SavannahOgeechee/Ocmulgee subregion. This is due to the fact that these draina ges share the second most common base shape for unifaces, and this category was assigned the lowest weight of the four arbitrary numerical possibilities (1). The Aucilla Suwannee and Tampa Bay drainages are the only physiographic regions examined for this study where incurvate was not the most or even the second most common type of basal configuration for side notched hafted bifaces or unifaces. To the north of the Aucilla Suwannee River drainage, incurvate is the second most common base shape for bifaces i n three

PAGE 227

227 river drainages; the Ocmulgee, Flint, and Chattahoochee. Additionally, incurvate is the second most common type of basal configuration for unifaces in the Ocmulgee River drainage. North of the Ocmulgee, incurvate is the most common type of basal configuration for bifaces in both the SanteeCooper and SavannahOgeechee River drainages, and it is the second most common base shape for unifaces in the SavannahOgeechee River drainage. It is important to consider why incurvate bases are in the minority f or all side notched tools in the Aucilla Suwannee and Tampa Bay drainages. Perhaps these Early Archaic groups elected not to make high frequencies of their side notched tool bases incurvate because they sought to minimize the amount of tool failure that can result from endthinning. However, since high quality cherts were so abundant in many parts of the reg ion, it is not clear why hunter gatherers might have been conserving lithic resources These trends in basal configuration also may be related to diff e rences in hafting techniques. While no hafted side notched tools have yet been recovered, some experts in primitive technology suggest that incurvate bases are appropriate to haft tools into hardwood, bone, or antler, while straight or excurvate bases may help to stabilize the haft element into material such as rivercane (Bissett 2003; Scott Jones and Kenn Steffy, pers onal communication 20032005). As such, there is a strong possibility that many of the side notched forms recovered in northcentral and Tamp a Bay, Florida were designed to be fit into rivercane hafts, while hardwood hafts may have been much more common in the Santee Cooper and Savannah Ogeechee River drainages. While more comparative paleoecological data are needed to follow up on these ideas, patterning in basal configuration clearly is tied to geographic locations and should be further investigated.

PAGE 228

228 Notably, no single type of basal configuration is exclusive to any Coastal Plain river drainage. A ll three types of basal configurations (incurva te, straight, and excurvate) occur in each of the seven river drainages tha t were defined for this study. Patterns of basal configurations certainly occur in differing proportions throughout this broad region, and that finding has perhaps been the most int riguing aspect of my study. Differenti al patterning in basal configura tion across the large expanse defined as the Southeastern Coastal Plain, suggests varying levels of social interaction throughout t he region. These interactions likely took many forms. P ossible examples of social interaction at the regional (read: extra drainage) level include, but certainly are not limited to such activities as mobility, trade, aggregatio n, gifting, and mate exchange. Furthermore, it seems that, at the local level, relat ed males who shared hunting histories and territories would learn how to make bifaces from their ancestors and other elders. Therefore, a sort of generational learning pattern may be evident in the collected data, as I am fully cognizant that the palimpses t we call the Early Side Notched Horizon took several centuries to form (cf. Morse 1996). I also compared hafted bifaces and unifaces as to the numerous attributes measured for both classes of side notched tools. The goal of this comparison was to identify those drainages where high mean values for specific attributes are typical, and to identify those attributes most known to exhibit certain other patterns of variation. For the purposes of this comparative discussion, it is worth repeating that, as no unif aces were examined from the Flint River drainage, and only one uniface was examined from the Chattahoochee River drainage, those drainages are considered here only with respect to patterned variation in hafted bifaces.

PAGE 229

229 First, I considered mean values for s even continuous variables examined for each tool class. With respect to the blade portion of all tools, I considered blade length, shoulder width, and maximum thickness. With respect to the haft element of all tools, I considered shoulder length, notch hei ght, base width, and betweennotch width. Measurements were taken for both the left and right sides of unifaces for blade length, shoulder length, and notch height. I observed the highest mean blade length values for hafted bifaces in the Chattahoochee, A ucilla Suwannee and Tampa Bay drainages. The highest mean values for uniface blade length are seen in the AucillaSuwannee and Tampa Bay drainages. With respect to shoulder width, I observed the highest mean values for hafted bifaces in the Chattahoochee, Aucilla Suwannee, and Tampa Bay drainages. The highest mean value for shoulder width is seen for Aucilla Suwannee hafted bifaces. Meanwhile, I observed the highest mean values for shoulder width in unifaces in the Aucilla Suwannee and Tampa Bay drainages. Mean values for maximum thickness for hafted bifaces increase from the Chattahoochee to the Aucilla Suwannee River drainage, with the highest mean values for hafted biface thickness seen in Tampa Bay. This is the same patterning observed with respect to bl ade length for hafted bifaces. Among unifaces from the five drainages that produced a sample greater than one, maximum thickness appears to vary according to geographical location to a lesser degree than do other elements of the blade portion of the tool c lass. The highest means for uniface thickness are seen in the Santee Cooper and Tampa Bay drainages. Overall, the blade portion of side notched hafted bifaces can be seen to increase to the west and south of the Flint River drainage. Mean values for blade length and thickness increase incrementally from the Flint to the Chattahoochee, and south to the Aucilla Suwannee, with the

PAGE 230

230 highest mean value for these continuous variables seen for hafted bifaces in Tampa Bay. Shoulder width for hafted bifaces is seen t o increase from Tampa Bay to the Chattahoochee, with the highest mean value for this measured attribute seen for the Aucilla Suwannee River drainage. These trends also are apparent in percentile plots constructed as part of this analysis. The blade portion of unifaces shares similar patterned variation to that seen for hafted bifaces. Mean values for blade length and shoulder width are highest in the Aucilla Suwannee and Tampa Bay drainages. Meanwhile, uniface maximum thickness follows no definable geograph ical patterning, with the lowest mean value seen for the Savannah Ogeechee River drainage, and the highest values observed in the Tampa Bay and Santee Cooper drainages. I also compared measured portions of the haft elements of bifaces and unifaces. These portions include shoulder length, between notch width, notch height, and base width. As previously stated, I expected the haft elements of both tool classes to exhibit the most stability in terms of patterned variation across the examined portions of the Coastal Plain. Two haft element attributes seem to su pport this expectation; betweennotch width and base width. Mean values for these two attributes do not appear to be significantly higher in the Aucilla Suwannee and Tampa Bay drainages for hafted bifaces or unifaces. Once again, these trends also are apparent in percentile plots. Mean values for notch height are highest in the Aucilla Suwannee and Tampa Bay drainages, for both hafted bifaces and unifaces. The only metric attribute relating to haft element that seems to warrant additional discussion with respect to relative mean values, is shoulder length. I observed the highest mean values for hafted biface shoulder length in the Aucilla Suwannee and Tampa Bay drainages. Shoulder length patterning in unifa ces looks decidedly different. I observed the highest mean

PAGE 231

231 values for uniface shoulder length in the Ocmulgee Ri ver drainage. The next highest mean values are seen for the Tampa Bay and Aucilla Suwannee drainages. This patterning holds true for both left a nd right shoulder length data and it is the only instance of the Ocmulgee River drainage producing the highest mean value for a measured attribute. This patterned variation for uniface shoulder length in the Ocmulgee River drainage may be best understood as a reflection of tool curation strategies employed at the time of tool manufacture. Perhaps Early Archaic visitors to the Feronia sites, who made and used all of the unifaces examined from the Ocmulgee River drainage, intentionally brought tools that wer e designed for long term use, since tool replacement was known to be unpredictable in an area with no high quality chert raw material. On a related note, while I observed the highest mean values for uniface notch heights (both left and right) in the Aucill a Suwannee and Tampa Bay drainages, the third highest mean values are seen in the Santee Cooper (for left notch height) and the Ocmulgee (for right notch height). As previously discussed, the Santee Cooper and Ocmulgee River drainages are devoid of hi gh quality chert raw material. Notch height, along with shoulder length, is apparently an aspect of the uniface haft element which can be interpreted in light of tool curation strategies. As presented above with respect to both hafted bifaces and unifaces, I a pplied statistical tests to all examined metric attributes. Using coefficient of variation as a measure of relative variance across the various drainages in this study, it is possible to make some gross comparisons for these two artifact classes. The value of the coefficient of variation is that it compares the degree of variation from one data series to another, even if the means are drastically different. Once again, I am parsing discussion into attributes relating to the blade portions of tools, and thei r respective haft elements.

PAGE 232

232 With respect to blade length, the most variation is seen for hafted bifaces from the Ocmulgee River drainage; I also observed a high degree of variation for left and right blade lengths of unifaces from the Santee Cooper River drainage ( n=4). The lowest coefficients of variation (indicative of the lowest degree of variation) were obtained for hafted biface blade length in Tampa Bay samples, and for uniface blade length (both left and right, respectively) in samples from Aucilla S uwannee and Tampa Bay. The greatest degree of variation with respect to shoulder width is seen for hafted bifaces from the Ocmulgee River drainage and for unifaces from the SanteeCooper River drainage. The least amount of variance for shoulder width data, for both hafted bifaces and unifaces, may be seen in examples from the Aucilla Suwannee River drainage. For maximum thickness for both hafted bifaces and unifaces, data from the SanteeCooper River drainage produced the highest coefficient s of variation ( 0.16 and 0.25, respectively ) of any Coastal Plain data set, illustrating a high degree of variation for this measured attribute in this particular location regardless of tool class. The lowest degree of variation for maximum thickness was identified in haf ted biface data from the Chattahoochee River drainage, and in uniface data from the AucillaSuwannee River drainage. Analysis of data for coefficient of variation also suggests that the highest degree of variation for primary edge angle data may be seen in examples of unifaces from the Ocmulgee River drainage (c.v.=0.27). For hafted bifaces, most haft element variation was observed in samples from the Ocmulgee and AucillaSuwannee River drainages. I observed the most relative variance for shoulder length and notch height in Aucilla Suwannee samples; Ocmulgee samples yielded comparatively high coefficients of variation for bet ween notch width and base width. Coefficient of variation data for hafted unifaces indicate that the widest range of variation for left shoulder

PAGE 233

233 length, between notch width, right notch height, and base width is found in samples from t he SanteeCooper River drainage. These data also indicate that the least relative variance for biface haft elements is evidenced in the Chattahoochee (for s houlder length, notch height, and base width); for uniface haft elements it is evidenced in the Aucilla Suwannee River drainage for all six measured attributes. It bears repeating that the inequality of vari ances for hafted biface ratio scale data (with th e exception of maximum thickness) is statistically significant, based on a series of Levenes tests (Gastwirth et al. 2009; Levene 1960). Meanwhile, the inequality of variances for uniface data is not statistically significant with respect to any ratio scale attribute. Finally, i t is interesting to note the ratios of the two tool types (side notched hafted bifaces and unifaces) throughout the Coastal Plain. The closest ratios of the tools can be found in the Savannah Ogeechee River drainage, where there is one uniface for every thre e side notched hafted bifaces. North of that point, in the Santee Coope r drainage, the ratio is 1:17. The Ocmulgee River drainage produced a ratio 1:8. As previousl y stated, no unifaces were available for examination in the Flint River drainage. S imilarly, in the Chattahoochee R iver drainage, the ratio is extremely low; 1:384. In the Aucilla Suwannee R ive r drainage, the ratio is 1:12. Then, in Tampa Bay, the ratio is 1:6, making that region the place with the second closest ratio of unifaces to side notched hafted bifaces in the Coastal Plain. These data can be seen in Table 530. Perhaps the best explanation for these ratios is that, while hafted bifaces were in perennial use, Edgefield Scrapers were a seasonal technology (after Goodyear et al. 1980) According to some researchers, we might expect to see higher frequencies of unifaces when

PAGE 234

234 Table 5 30. Ratios of hafted bifaces to unifac es throughout the Coastal Plain Drainage Total Unifaces Total Bifaces Ratio of Unifaces to Bifac es SanteeCooper 4 66 1:17 Savannah Ogeechee 188 603 1:3 Ocmulgee 9 74 1:8 Flint 0 97 0:97 Chattahoochee 1 384 1:384 Aucilla Suwannee 39 464 1:12 Tampa Bay 14 77 1:6 Total 255 1,765 1:7 groups were g earing up in advance of winter, for which they would need to process a lot of hides. That explanation may work well for hafted endscrapers, and is supported by ethnoarchaeological studies (after Cable 1996). However, a s previously stated, it is unlikely that Edgefield Scrapers were used for the processing of hides, since their design is poorly suited for that purpose. The original type description for the Edgef ield Scraper suggests that the tool form may have been used for working bone and wood, and subsequent researchers have continued to operate on that assumption (Goodyear et al. 1980 ; Michie 1968). Zones of Patterned Variation across Drainages Throughout this chapter, I have explicated the p atterned variation in Early Archaic sidenotched tools for the S outheastern Coastal Plain. In explaining the genesis of that variation, I have relied heavily on conventional wisdom about how these tools were made and resharpened. Geographical trends in sources for potable water and Coastal Plain chert informed many of my interpretations about variation in continuous variables, relating to both the blade and haft elements of bifaces and unifaces (cf. Bamforth 1986, Goodyear 1979; Kelly 1988). The social

PAGE 235

235 context of side notched tools in daily practice must now be further investigated as my research question relates to the social boundaries that may be reflected in material culture variation It is possible to view the patterned variation detailed in this chapter as a product of macroregional social interaction rather than differentiation. In building a framework for the Early Side Notched Horizon from an anthropological perspective, I must return to concepts including universal kinship and large scale social networks, as documented in lessons from ethnographic literature first introduced in Chapter 2. This research has broad implications for our understanding of social boundaries in highly mobile populations. Social group affil iation for Early Archaic hunter gatherers may have been much more f luid than has traditionally been presumed by archaeologists considering materiality at a smaller regional scale. These concepts will be further developed in Chapter 6. For the present discussion, I consider patterning within the Early Side Notched Horizon with respect to the relatively discrete geographical zones where sidenotched biface types were defined, comparing variation in the centers of these zones with the boundary area where said zones can be seen to intersect. As described in Chapter 4, I desig nated the Strong Collection locality ( n =411 hafted bifaces) in Allendale County, South Carolina (part of the Savannah Ogeechee River drainage) as the approximate center of the Taylor Zone. I determined that the Hendrix Collection locality ( n=268 hafted bif aces) from the Santa Fe and Suwannee Rivers in Florida (part of the Aucilla Suwannee River drainage) was the approximate center of the Bolen Zone. I was not able to examine large frequencies of hafted bifaces from the center of the Big Sandy Zone. Therefor e, I compared variation in the respective centers of the Taylor and Bolen Zones, with variation observed in a boundary area in Coffee County, Georgia ( n=74) (part of the

PAGE 236

236 Ocmulgee River drainage ) Tables 5 31 and 532 present descriptive statistics for the centers of the Taylor and Bolen Zones, respectively. Table 5 31. Descriptive statistics for the center of the Taylor Zone n Mean Min Max s.d. c.v. Blade length 212 30.31 12.2 69.01 9.6 0.32 Shoulder width 261 21.71 12.9 35.57 3.25 0.15 Maximum thickness 385 6.43 3.89 9.35 0.95 0.15 Shoulder length 398 10.18 6.66 19.8 1.8 0.18 Between notch width 393 15.4 10.22 25.32 2.55 0.16 Notch height 390 5.35 2.7 13.67 1.28 0.24 Base width 368 21.78 10.57 39.68 3.64 0.17 Table 532. Descriptive statistics for the center of the Bolen Zone n Mean Min Max s.d. c.v. Blade length 238 41.27 18.99 113.52 12.72 0.31 Shoulder width 213 25.74 16.91 41.07 4.15 0.16 Maximum thickness 268 7.61 5.31 11.19 1.08 0.14 Shoulder length 268 13.43 7.89 22.24 2.64 0.2 Between notch width 268 16.24 10.05 25.2 2.73 0.17 Notch height 267 7.84 4.16 15.41 2.15 0.27 Base width 254 24.32 16.11 36.32 3.26 0.13 Overall, the coefficients of variation obtained for ratioscale data from the centers of the Taylor and Bolen Zones are quite similar A majority (four, or 57%) of coefficients of variation obtained for the measured attributes display a difference of 0.01. Similar relative variance was observed for blade length, shoulder width, maximum thickness, and between notch width. Meanwhile, differences of 0.02, 0.03, and 0.04 were observed for coefficients of variation obtained for shoulder length, notch height, and base width, respectively. There is evidence for

PAGE 237

237 higher relative variance for the haft elements of bifaces from the centers of the Taylor and Bolen Zones, compared to aspects of the blade. However, in general biface data from the centers of these zones can be seen to display low variance. This patterning is to be expected if t hese l ocalities in fact represent the centers of respective social group range s More variation would then be expected for inferred boundary areas. Descriptive statistics for the boundary area of the Ocmulgee River drainage are presented in Table 5 33. I c ompared the relative variance indicated by these data, to the same obtained for the centers of the Taylor and Bolen Zones discussed above. The coefficients of variation obtained for hafted biface data from the Ocmulgee boundary area are indicative of highe r variance compared to the centers of the Taylor and Bolen Zones, for both the blade portions and haft elements of these tools. A majority (five, or 71 % ) of coefficients of variation obtained for the measured attributes display a difference of 0.02 or grea ter. The greatest disparity is found in comparing coefficients of variation obtained for shoulder width (a difference of 0.060.07) and between notch width (a difference of 0.040.05). Table 533. Descriptive statistics for the Ocmulgee boundary area n Mean Min Max s.d. c.v. Blade length 57 27.9 12.64 48.74 9.5 0.34 Shoulder width 56 22.09 14.87 36.52 4.87 0.22 Maximum thickness 72 6.75 3.43 9.75 0.98 0.14 Shoulder length 70 10.82 6.8 15.87 1.89 0.17 Between notch width 69 15.66 10.65 24.34 3.3 0.21 Notch height 71 5.62 3.34 8.92 1.25 0.22 Base width 52 23.3 16.05 32.62 3.97 0.17

PAGE 238

238 I created percentile plots of all ratio scale data from the centers of the Taylor and Bolen Zones, to allow comparison with data from the Ocmulgee boundary area on the same graph s. These plots are presented in Figures 5 30 through 536. For certain metric attributes (including maximum thickness, shoulder length, and notch height), Taylor Zone data are left skewed relative to n (where n is the sample of all data), an d Bolen Zone data are right skew ed, while data from the Ocmulgee boundary area span the difference between the two samples. Trends for base width pattern similarly, except that above the median, Ocmulgee boundary area data largely follow the same trend as data from the Bolen Zone (Figure 36). With respect to blade length, Ocmulgee boundary area data points largely share the Taylor Zone trend, albeit with lower measurement values (Figure 5 30). For shoulder width, Ocmulgee boundary zone data are left skewed relative to the Taylor Zone data at and below the median; above the median value, Ocmulgee boundary area data points span the difference between the Taylor and Bolen Zone samples (Figure 5 31). Patterning for between notch width stands out among the percentile plots created for zone boundary relationships (Figure 534). Here, data from the Taylor and Bolen Zones share similar trends. Meanwhile, below the median, Ocmulgee boundary area data are left skewed relative to Taylor Zone data. At the median value, O cmulgee boundary area data largely share the same trend as the Taylor Zone. Beginning at approximately the 70th percentile, data from the Ocmulgee boundary area share the same trend as Bolen Zone data, ultimately skewing to the right of these data above t he 80th percentile with numerous hig h value data points for between notch width. The high variance observed for shoulder width and betweennotch width data from

PAGE 239

239 the Ocmulgee boundary area (relative to the Taylor and Bolen Zones) also is reflected in the ma rked disparities for coefficients of variation, as previously discussed. Since the percentile plots show distributions that are not exactly normal, and the samples have varying means, I conducted a Levenes test to determine whether the differences in samp le variances for ratio scale data from the zone boundary test case are statistically significant (Gastwirth et al. 2009; Levene 1960). Table 534 presents the results of th e Levenes test for all ratio scale attributes. With respect to all measured attribu tes other than maximum thickness, Levenes Statistic (W) is greater than the obtained critical value (F). Therefore, the null hypothesis can be rejected in six out of seven instances; the inequality of variances for blade length, shoulder width, shoulder l ength, between notch width, notch height, and base width data is statistically significant. Meanwhile, the null hypothesis cannot be rejected with respect to maximum thickness in the zone boundary test case; the inequality of variances for maximum thicknes s is not statistically significant. Table 5 34. Results of Levenes tests conducted for ratio scale data (zoneboundary test case) Ratio scale Attributes (mm) Levene's Statistic (W) Critical Value (a=0.05) (F) Blade length 4.212050054 3.013609416 Shoulder width 8.437136018 3.012826237 Maximum thickness 1.791543358 2.30994408 Shoulder length 20.50941968 3.0080091 Between notch width 3.321610439 3.008110701 Notch height 35.24771551 3.008144942 Base width 2.743274246 2.310504688 As anticipated in Chapters 3 and 4, given the likelihood that Early Archaic group ranges were characterized by permeable social boundaries, I expected that the most variance for nominal and ratioscale artifact attributes would be evidenced at the zonal bo undaries of those regions. This expected patterning was substantiated by analyses of relative variance for metric

PAGE 240

240 attributes, as previously described. N ext I considered patterning for basal configuration in these sample areas. Table 5 35 displays frequencies of hafted b ifaces by sample and basal configuration. Table 535. Absolute and relative frequencies of hafted bifaces by sample and basal configuration Sample Incurvate Straight Excurvate Total n % n % n % Taylor Zone 267 66 .25 102 25 .31 34 8.44 403 Boundary Area 29 43.94 31 46.97 6 9 .09 66 Bolen Zone 72 27 .07 86 32 .33 108 40.6 266 Total 368 219 148 735 As shown in Table 535, relative frequency obtained for base shape in the Taylor Zone is identical to that obtained for the Savann ah Ogeechee River drainage ( Table 5 1 and Figures 51 and 52). A majority ( n =267, or 66% ) of bifaces from the center of the Taylor Zone is incurvate. Straight and excurvate forms are found less frequently in this sample. In the center of the Bolen Zone t here is a strong tendency for hafted bifaces to be made with excurvate bases. Specifically, slightly less than 41% (108 out of 266) of this sample have excurvate bases. The Bolen Zone sample produced lower frequencies of straight and incurvate bases [32% ( n=86) and 27% ( n=72) respectively]. This patterning is notably different from that observed for the Aucilla Suwannee River drainag e, where a majority (48% ) of hafted bifaces was straight based, followe d by excurvatebased (34% ). Variation in basal configuration for the Ocmulgee boundary area was previously discussed, with respect to the Ocmulgee River drainage. It bears repeating that straight and incurvate bases occur almost equally in the boundary area sample. That the boundary area produced the most variation in basal configuration relative to the Taylor and Bolen Zones is suggestive of a coalescence of various cultural and technological traditions in this

PAGE 241

241 area. At mi nimum, variation in basal configuration in the Ocmulgee boundary samples may be interpreted in light of the permeability of social boundaries. Perhaps this patterning also lends support to the notion that groups throughout the region aggregated along the O cmulgee River. Pilot Study for Qu arry Cluster Sourcing of Florida Bifaces The selection of particular raw material sources over others is at once a technological choice and a social production. For instance, a comparison of the raw materials available to E arly Archaic groups with the materials they actually selected may reveal information about knowledge shared among members of specific band or macroband groups (after Lemonnier 1992). In this sense, raw material selection may be seen as a function of social ization, and as such it is a function of group structure and organization, which I would argue was kinbased during the Early Archaic. Therefore, it is relevant here to consider the results of a pilot study for the specific quarry cluster sourcing of numer ous side notched hafted bifaces from two portions of Florida; the Aucilla Suwannee and Tampa Bay drainages. As was related in Chapter 4, w hile conferring with various collectors in the Tampa Bay region, I was able to analyze high frequencies of side notche d artifacts from both that region, and environs to the northwest (one more northern locale is known as Lake Bird, and detailed discu ssion of that particular site also was provided in Chapter 4). Robert Austin of Southeastern Archaeological Research, Inc. i n Riverview, Florida, generously assisted me by microscopically analyzing the petrology of numerous ( n=118) side notched hafted bifaces from sever al collections. Austin (1997; Austin and Mitchell 1999) has conducted similar studies throughout F lorida for a lmost two decades. The findings of Austins recent analyses on my behalf are described in detail below. They well illustrate the value of such studies, and their implications

PAGE 242

242 for understanding the movement of ancient artifacts and people, throughout the ea rliest portion of the Early Archaic period. Artifact collectors who reside in St. Petersburg, Florida allowed me to examine a total of 151 side notched hafted bifaces. Their collections were made from the Lake Bir d site in the Aucilla Suwannee R iver draina ge (where I analyzed a total of 74 examples), and from throughout the Tampa Bay region ( n =77). In the Tampa Bay region, the collectors had separated their finds according to the three distinct archaeological sites of Kepler in Pinellas County, the Harris G rove and Harney Flats site location in Hillsborough County, and the Spring Hill site in southern Pasco County. They provided more general proveniences for certain other artifacts, such as the St. Petersburg area, the Tampa Bay area, Pinellas County, and th e Lake Seminole area (which is in Pinellas County). For the purposes of my study, I aggregated all locations in the Tampa Bay region, to one unit of anal ysis (the Tampa Bay drainage). However, in order to reflect Austins raw material analyses, here I prov ide specific data according to Austins distinctions, which were made based on provenience information provided by the local collectors. Robert Austin typed a total of 118 of the Early Archaic artifacts collected by individuals res iding in the Tampa Bay re gion. T his rep resents approximately 78% of the total 151 pie ces I analyzed He examined 37 (or, 50% ) of the pieces I analyzed from Lake Bird, and 79 of the pieces I analyzed for the Tampa Bay region [two of the pieces he typed later were determined to be corner notched, resulting in my total of 77 pie ces analyzed for this region]. Austins analyses were geared toward the identification of raw material sour ce used in biface manufacture. Using 1070 power microscopy with a Bausch and Lomb Stereozoom 7TM, Aust in was able to identify

PAGE 243

243 the majority of pieces he examined, to specific Florida chert formation and major quarry clust er (after Upchurch et al. 1982). An adjustable 150 watt light source through movable fiber optic cab les facilitated this analysis. The spe cific findings of Austins research are detailed below. Refer to Figure 4 10 for the locations of Florida sites in the context of chert quarry clusters. Lake Bird Of the 37 side notched hafted bifaces from the Lake Bird site that were examined as to raw material source, 13 appear to be manufactured from chert found in the Wacissa quarry cluster; an additional nine bifaces were possibly manufactured using c hert from this quarry cluster. Suwannee cherts and silicified corals are common in this quarry cluster ( Robert Austin personal communication 2005). As shown in Figure 410, the Wacissa quarry cluster is the nearest known source of chert to the location of Lake Bird, as it is located approxima tely 14.5 km west of the site. Two side notched hafted bifaces from Lake Bird appear to be manufactured of chert from the Lower Suwannee quarry cluster, whic h is located approximately 48 .0 km to the south and southea st of the archaeological site. Table 5 36 presents data for quarry cluster sourcing for many of the side notched hafted bifaces from Lake Bird (Robert Austin, personal communication 2005). Table 5 36. Quarry cluster sourcing for side notched hafted bifaces from Lake Bird Quarry Cluster Total artifacts Distance from site (km) Hillsborough River/Bay Bottom? 2 257 Lower Suwannee 2 48 Unidentified 11 Wacissa 13 14.5 Wacissa? 9 14.5 Total 37

PAGE 244

244 Additionally, the raw material of two side notched hafted bifaces was described as having characteristics of a chert known as Bay Bottom, associated with the Hillsborough River quarry cluster. This finding is especially interesting, since that quarry cluster is located approximately 257 km south of Lake Bird. While Austin was unable to provide a definitive association between these tools and cherts from the Hillsborough River quarry cluster, he noted the presence of pelodial packstone in the material; a feature common to the cherts from that vi cinity. The possibility that Early Archaic hunter gatherers traveled such great distances, whether as part of a semi annual round or for other reasons, is very real, and should not be dismissed out of hand. Ethnographic correlates for such movement are known to exist, and similar raw material sourcing studies based on the archaeological record also lend support to the notion of broad regional transhumance (after Anderson and Hanson 1988; Goodyear 2010). A total of 11 pieces from Lake Bird could not be typed to quarry cluster; one was too patinated for a definitive identification, and ten were described as having too few fossils to permit more than a tentative designation to the Wacissa quarry cluster ( Robert Austin personal communication 2005). In summary, a majority of the Lake Bird hafted bifaces that could be sourced to specific quarry cluster (24 out of a total of 26) were m ade of locally available cherts. Evidence for the use of extra local raw materials was identified for two hafted bifaces from Lake Bird. Tampa Bay Robert Austin examined a total of 79 side notched hafted bifaces from the Tampa Bay region, typing them to r aw material source and specific quarry cluster where possible. Austin organized these artifacts according to localities described by the artifact collectors I interviewed

PAGE 245

245 in St. Petersburg. Subsequently, for my own analytical purposes, I aggregated these data into one grouping; Tampa Bay. For the purposes of this pilot study, Austins (by virtue of the collectors) regional distinctions are maintained within four subsections; (1)Tampa Bay Area, (2) St. Petersburg Area, (3) Pinellas County (including greater Pinellas County, as well as the Lake Seminole area and the Kepler site), and (4) Hillsborough County (including the Harris Grove and Harney Flats sites). Austin did not analyze artifacts from the Spring Hill site in Pasco County. Tampa Bay area Austin exa mined a total of nine hafted bifaces from a general provenience known simply as the Tampa Bay a rea. Table 5 37 displays the quarry cluster sourcing information for these artifacts ( Robert Austin, personal communication 2005). Three of these could be sour ced to chert from the Hillsborough River quarry cluster, which extends to the east and northeast of Tampa Bay; one additional biface was described as possibly made of cherts from the Hill sborough River quarry cluster. Austin was able to trace the raw mater ial of one biface to the Upper Withlacoochee quarry cluster, located to the northwest of Tampa Bay. Chert used to produce one biface found in the Tampa Bay area was thought to originate from either the Hillsborough River or the Upper Withlacoochee quarry c lusters, as certain characteristics of the silicified cora l in these cherts are similar. One biface was sourced to the Bay Bottom cherts of the Hillsborough River quarry cluster. Austin tentatively identified one tool as being manufactured of chert from th e La ke Panasoffkee quarry cluster. This is of special note since that quarry cluster is located more than 100 km (sp ecifically, approximately 147 k m ) northeast of Tampa Bay, a region whe re high quali ty raw material was plentiful. As at Lake Bird, it is possible to infer a high degree of mobility for Early Archaic groups in the Tampa Bay region, despite not knowing the exact

PAGE 246

246 provenience of these particular artifacts. The final hafted biface from the Tampa Bay area could not be ident ified as to quarry cluster of origin ( Robert Austin, personal communication 2005). Table 5 37. Quarry cluster sourcing for side notched hafted bifaces from the Tampa Bay area Quarry Cluster Total artifacts Hillsborough River 3 Hillsborough River or Upper Withlacoochee 1 Hillsborough River/Bay Bottom 1 Hillsborough River? 1 Lake Panasoffkee? 1 Unidentified 1 Upper Withlacoochee 1 Total 9 In summary, a majority of the hafted bifaces from the Tampa Bay area that could be sourced to specific quarry cluster (seven out of a total of nine) were ma de of locally available cherts. Evidence for the use of extra local raw materials was identified for one hafted biface from the Tampa Bay area. St. Petersburg area Austin examined a total of six hafted bifaces from a general provenience known as the St. Petersburg a rea. He described one of these tools as likely originating from cherts in the Caladesi quarry cluster, located to the northwest of Tampa Bay. One tool was identified as being manufactured fr om Bay Bottom cherts in the Hillsborough River quarry cluster, located to the east and northeast of Tampa Bay. One biface was determined to be manufactured of chert from either the Hillsborough River or the Caladesi quarry cluster. Additionally, Austin tra ced the cherts of two hafted bifaces to the Brooksville quarry cluster, located approxima tely 30 km north of Tampa Bay. It was not possible to determine the quarry cluster of origin for the final hafted bifa ce from the St. Petersburg area. Table 5 38 prese nts data for quarry cluster sourcing for side -

PAGE 247

247 notched hafted bifaces from the St. Petersburg area (Robert Austin, personal communication 2005). Table 5 38. Quarry cluster sourcing for side notched hafted bifac es from the St. Petersburg area Quarry Cluster Total artifacts Brooksville 2 Caladesi? 1 Hillsborough River/Bay Bottom 1 Hillsborough River/Caladesi 1 Unidentified 1 Total 6 In summary, locally available raw materials were utilized for the manufacture of all St. Petersburg area hafted bifaces that could be sour ced to specific quarry cluster. No evidence for the use of extralocal raw materials was identified for hafted bifaces from the St. Petersburg area. Pinellas County A total of 18 hafted bifaces collected from a general provenience of Pine llas County was examined by Austin. Table 5 39 displays the quarry cluster sourcing information for these artifacts ( Robert Austin, personal communication 2005). The raw material of the majority of these pieces ( n=10) was traced to the Hillsborough River quarry cluster, located to the east and northeast of Tampa Bay; one of these ten was manufactured of Bay Bottom cherts. The cherts of two tools were sourced to the Caladesi quarry cluster, located to the northwest of Tampa Bay, and one artifact was traced t o the Brooksville quarry cluster, which begins approxima tely 30 km north of Tampa Bay. Two bifaces were determined to be manufactured of chert from the Upper Withlacoochee quarry cluster, which is found to the northeast of Tampa Bay, and abuts the northern extent of the Hillsborough River quarry cluster. A total of three hafted bifaces from the

PAGE 248

248 general provenience of Pinellas County could not be definitively traced to quarry cluster, and this largely was due to patination on the artifacts. Table 5 39. Quar ry cluster sourcing for side notched haft ed bifaces from Pinellas County Quarry Cluster Total artifacts Brooksville 1 Caladesi 2 Hillsborough River 9 Hillsborough River/Bay Bottom 1 Unidentified 3 Upper Withlacoochee 2 Total 18 Austin analyzed two hafted bifaces that were collected in the Lake Sem inole area of Pinellas County. One was made of chert from the Hillsborough River quarry cluster, and the other was typed to t he Brooksville quarry cluster. These quarry clusters both were acces sible wit hin approximately 30 km of the Lake Seminole area. The chert of one hafted biface from the Kepler Site in Pinellas County was sourced t o the Caladesi quarry cluster. This quarry cluster was access ible within 20 km of the site. Interestingly, the Turtlecrawl Point quarry cluster, located along the southern extent of Pinellas County, was not identified as the likely source of chert for any of the hafted bifaces recovered in Pinel las County. Turtlecrawl Point quarry cluster cherts would have been par ticularly convenient for transport to the Kepler site, which is locat ed to the immediate northwest. Perhaps other locally available raw material sources, such as those in the nearby Hillsborough River, Caladesi, and Upper Withlacoochee quarry clusters, wer e preferred. In summary, locally available raw materials were utilized for the manufacture of all Pinellas County hafted bifaces that could be sour ced to specific quarry cluster. No evidence for the use of extralocal raw materials was identified for hafte d bifaces from Pinellas County.

PAGE 249

249 Hillsborough County, Harris Grove Collection Austin examined a total of 40 side notched hafted bi faces from the Harris Grove Collection in Hillsborough County. The majority ( n=27) of these tools was traced to the Hillsboroug h River quarry cluster, the most locally available raw materia l source. The Harris Grove Collection locality is located within the Hillsborough River quarry cluster. Three hafted bifaces were traced to the Upper Withlacoochee quarry cluster, located northeast of the site; one of these bifaces is made of chert from the Rock Ridge portion of the Upper Withlacoochee quarry cluster. Austin determined that three bifaces were manufactured from a rather indeterminate source containing silicified coral; eithe r the Hillsborough River or the Upper Withlacoochee quarry clusters may have been the source for this material. Two additional hafted bifaces were traced to probable quarry clusters; one to the Caladesi, to the northwest of the Harris Grove Collection loc ality ; a nd one to the Lake Panasoffkee. The Lake Panasoffkee quarry cluster is located approximately 126 km northea st of the collection locality Four bifaces were too patinated to determine their quarry cluster of origin ( Robert Austin, personal communica tion 2005). Table 540 shows data for quarry cluster sourcing of side notched hafted bifaces from the Harris Grove Collection. In summary, a majority of the Harris Grove Collection hafted bifaces that could be sourced to specific quarry cluster (35 out of a total of 36, approximately 97% ) were made of locally available cherts and many of these cherts were available in the immediate vicinity Evidence for the use of extra local raw materials was identified for one hafted biface from the Harris Grove Collec tion.

PAGE 250

250 Table 5 40. Quarry cluster sourcing for side notched hafted bifaces f rom the Harris Grove Collection Quarry Cluster Total artifacts Caladesi? 1 Hillsborough River 27 Hillsborough River or Upper Withlacoochee 3 Hillsborough River/Bay Bottom 1 Lake Panasoffkee? 1 Unidentified 4 Upper Withlacoochee 2 Upper Withlacoochee/Rock Ridge 1 Total 40 Hillsborough County, Harney Flats Austin examined five hafted bifaces from the Harney Flat s site in Hillsborough County. Four of these pieces were manufactured from Hillsborough River quarry cluster material. This is not surprising, as the archaeological site is situated at that particular quarry cluster. One biface from the Harney Flats site was too patinated for a definiti ve quarry cluster association. Locally available raw materials were utilized for the manufacture of all Harney Flats hafted bifaces that could be sourced to specific quarry cluster. No evidence for the use of extra local raw materials was identified for ha fted bifaces from Harney Flats. Summary of Pilot Study As previously discussed, Early Holocene hunter gatherers likely designed their settlement strategies in part to ensure proximity t o potable water as well as high quality lithic raw material. It is impo rtant to note some exceptions to these trends in site lo cation. One apparent exception is the Lake Bird site in northern Taylor County, which is located between two quarry clusters (the Wacissa and the Lower Suwannee). While this location was not absolutel y ideal, in that it is situated away from chert sources, it likely was an important source of freshwater to site visitors.

PAGE 251

251 Lake Bird apparently was a popular place for many Early Archaic peoples, as hundreds of side notched tools have been recovered there, 74 of which I was allowed to examine, and others of which I am aware from published reports (Bissett 2003; Gramly 1994). Very few of the samples examined by specialists in chert sourcing have identified the use of extra local raw materials in side notched tool manufacture throu ghout Floridas Coastal Plain. This may be a reflection of the Bolen Drought, which may have constrained hunter gatherer mobility. Evidence of Florida Indians using local cherts during the Early Archaic comes from at least three locales, two of which are located in north Florida, and one of which is located in central Florida. These locales include the Lake Bird (8TA143) site in Taylor County, the Jeanies Better Back (8LF54) site in Lafayette County, and the Tampa Bay re gion. In sepa rate studies, Robert Austin examined numerous individual side notched artifacts from these locales (Austin and Mitchell 1999; Robert Austin personal communication 2005). In all but six instances, two of which were at the Jeanies Better Back site, two of which were at the Lake Bird site, one of which was from the Tampa Bay area, and one of which was at the Harris G rove Collection locality researchers have sourced those side notched artifacts to local quarry clusters. The Lake Bird site in Taylor County i s typical of Early Archaic samples from throughout Florida, as a majority (22 out of a total of 26 that could be sourced to quarry cluster) of side notched hafted bifaces from the site likely were made from chert obtained from the Wacissa quarry clusters, located approxim ately 14.5 km west of the site. This is not terribly surprising, since the most homogenous and highest quality raw materials found in Florida are the cherts found in Tertiary karsts. However, this is a very different pattern than others have observed

PAGE 252

252 throughout the Southeast, where side notched tools commonly are made from extra local raw materi al, and where high quality raw material is scarcer. As archaeologists commonly use data for distance from raw material source as a proxy for group mobility, some statements can be made here r egarding the mobility of hunter gath erers in Early Archaic Florida. Some degree of movement between the AucillaSuwannee and Tampa Bay drainages is indicated by the presence of a total of six hafted bifaces made from extralocal cherts. Table 5 41 displays the estimated maximum mobility data indicated by preliminary studies of quarry cluster sourcing for side notched hafted bifaces from both the Aucilla Su wannee and Tampa Bay drainages. Table 5 41. Estimated maximum mobility using quarry cluster sourcing data Drainage Provenience Northern movement (km) Southern movement (km) Total artifacts Aucilla Suwannee Jeanie's Better Back 173 150 2 Aucilla Suwannee Lake Bird 257 2 Tampa Bay Tampa Bay Area 147 1 Tampa Bay Harris Grove 126 1 Total artifacts 3 3 6 The Southern movement (km) column for the first two rows of Table 5 41 corresponds directly to apparent evidence for cross drainage movement between the AucillaSu wannee and Tampa Bay drainages. This movement is indicated by the presence of one biface from the Jeanies Better Back site that was sourced to the Brooksville quarry cluster (a distance of approximately 150 km ), and two bifaces from the Lake Bird site that were sourced to Bay Bottom che rts from the Hillsborough River quarry cluster (a distance of approximately 257 km ).

PAGE 253

253 This apparent evidence for cross drainage movement relates to a proposed model for Early Archaic social boundaries that is further developed in Chapter 6. Summary In this chapter, I presented the results of a broad regional study of patterned variation in side notched hafted bifaces and unifaces from throughout the Coastal Plain. First, I presented a detailed analysis of hafted bifaces, considering patterned variat ion in both nominal and ratioscale attributes. These data were introduced with respect to individual drainage systems, from the SanteeCooper River drainage in South Carolina, south to Tampa Bay, and west to the Chattahoochee River drainage on the border of Alabama with Georgia. Next I discussed the patterning for the blade portions and haft elements of all hafted bifaces considered for this study. The same level of analysis and discussion then was provided with respect to side notched hafted unifaces, which were found to occur in much lower frequencies and in fewer drainage systems than were sidenotched hafted bifaces. I then compared patterned variation for hafted bifaces and unifaces throughout the Coastal Plain, and noted specific geographic zones of apparent regionalization rel ating to blade size and basal configuration, among other attributes. One key finding with respect to basal configuration is that base shape appears to be directly correlated to geography throughout the Coastal Plain for both side notched hafted bifaces an d unifaces. This chapter also included comparison of variation in hafted bifaces from the centers of the Taylor and Bolen Zones, with examples from the inferred boundary area in the Ocmulgee River drainage where those zones intersect. Using percentile plot information and a variety of statistical tests, I demonstrated that hafted bifaces from the Ocmulgee boundary area display more variance compared to samples

PAGE 254

254 from the centers of the Taylor and Bolen Zones. This patterning is indicative of subregional techn ological traditions within the Early Side Notched Horizon. Finally, I presented the results of a pilot study for quarry cluster sourcing of hafted bifaces from throughout Florida. The implications of all of these analyses for understanding social boundarie s during the earliest portion of the Early Archaic are addressed in Chapter 6.

PAGE 255

255 Figure 5 1. Hafted biface basal config uration by percentage ( n =1,723)

PAGE 256

256 Figure 5 2. Differentiation in basal configuration for hafted bifaces throughout the Coastal Plain

PAGE 257

257 Figure 5 3. Major drainage systems and the distribution of Coastal Plain chert throughout the study region

PAGE 258

258 Figure 5 4. Possibly h ypertrophic hafted bifaces recovered from the Waccasassa and Santa Fe Rivers (Artifact IDs 505 [top] and 506 [bottom]) (Photos courtesy of the author)

PAGE 259

259 Figure 5 5. Normal distribution percentile plot for hafted biface blade length

PAGE 260

260 Figure 5 6. Normal distribution p ercentile plot for hafted biface shoulder width

PAGE 261

261 Figure 5 7. Normal distribution percentile plot for hafted biface thickness

PAGE 262

262 Figure 5 8. Trends for blade length, shoulder width, and maximum thickness for hafted bifaces throughout the Coastal Plain

PAGE 263

263 Figure 5 9. Variation in the blade portion of hafted bifaces throughout the Coastal Plain

PAGE 264

264 Figure 5 10. Normal distribution percentile plot for hafted biface shoulder length

PAGE 265

265 Figure 5 11. Normal distribution p ercentile plot for hafted biface between notch width

PAGE 266

266 Figure 5 12. Normal distribution percentile plot for hafted biface notch height

PAGE 267

267 Figure 5 13. Normal distribution p ercentile plot for hafted biface base width

PAGE 268

268 Figure 5 14. Variation in the haft element of hafted bifac es throughout the Coastal Plain Figure 5 15. Hafted uniface basal conf iguration by percentage ( n =241)

PAGE 269

269 Figure 5 16. Differentiation in basal configuration for hafted unifaces throughout the Coastal Plain

PAGE 270

270 Figure 5 17. Normal distribution percentile plot for uniface left blade length

PAGE 271

271 Figure 5 18. Normal distribution p ercentile plot for uniface right blade length

PAGE 272

272 Figure 5 19. Normal distribution p ercentile plot for uniface shoulder width

PAGE 273

273 Figure 5 20. Normal distribution p ercentile plot for uniface thickness

PAGE 274

274 Figure 521. Trends for blade length, shoulder width, and maximum thickness for hafted unifac es throughout the Coastal Plain

PAGE 275

275 Figure 5 22. Variation in the blade portion of hafted unifaces throughout the Coastal Plain

PAGE 276

276 Figure 5 23. Normal distribution percentile plot for uniface left shoulder length

PAGE 277

277 Figure 5 24. Normal distribution percentile plot for uniface right shoulder length

PAGE 278

278 Figure 5 25. Normal distribution percentile plot for uniface betweennotch width

PAGE 279

279 Figure 5 26. Normal distribution percentile plot for uniface left notch height

PAGE 280

280 Figure 5 27. Normal distribution p ercentile plot for uniface right notch height

PAGE 281

281 Figure 5 28. Normal distribution percentile plot for uniface base width

PAGE 282

282 Figure 5 29. Variation in the haft element of hafted unifaces throughout the Coastal Plain

PAGE 283

283 Figure 5 30. Normal distribution percentile plot for hafted biface blade length (zoneboundary test case)

PAGE 284

284 Figure 5 31. Normal distribution percentile plot for hafted biface shoulder width (zone boundary test case)

PAGE 285

285 Figure 5 32. Normal distribution p ercentile plot for hafted biface thickness (zoneboundary test case)

PAGE 286

286 Figure 5 33. Normal distribution percentile plot for hafted biface shoulder length (zone boundary test case)

PAGE 287

287 Figure 5 34. Normal distribution p ercentile plot for hafted biface betweennotch width (zone boundary test case)

PAGE 288

288 Figure 5 35. Normal distribution percentile plot for hafted biface notch height (zone boundary test case)

PAGE 289

289 Figure 5 36. Normal distribution percentile plot for hafted biface base width (zoneboundary test case)

PAGE 290

290 CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS Hunter Gatherers during the Early Archaic Many histories have effectively been hidden by archaeologists apparent preoccupation with finding social entities, which are pres umed to be reflected by culturehistorical patterns in the archaeological record (after Sassaman 2001a; Wolf 1982). This preoccupation is a direct historical consequence of the reliance on twentieth century models of hunter g atherers as isolates (Sassaman 2011; Woodburn 1995) As a result, observed similarities in material culture have been interpreted as reflective of social connections and sameness, while patterns of divergence in material culture long have been treated as products of social boundaries and differentiation. Ancient hunter gatherers long have been perceived as if they existed apart from processes of history and change. Such notions cannot be sustained in reconstructions of their histories (after Wolf 1984). It is feasible that people who regularly interact (such as members of a cooperative hunting group) would create a certain level of patterning, in that the attributes of their artifacts might be similar (after Deetz 1968, who actually gave as his example, an assemblage of sidenotched incurvate bases made by the Chumash). According to this example, we might expect members of the same hunting group to all produce incurvate based tools, while other hunting groups from neighboring communities might produce excurvate or straight ba sed tools. It is also the case that members of interacting group s will deliberately mark differences in order to render unambiguous the social obligat ions and privileges among them (cf. Wies sner 1982, 1983). I t is possible to view such patterns as products of social interactions, rather than differenti ation and isolation (Sassaman 2011). This is because social interactions commonly are asymmetrical in nature, and can result in differences in material expression as well as lifeways.

PAGE 291

291 These differences have a strong potential to be archaeologically visible, especially if viewed on a macroregional scale (Sassaman 2001a). Such an approach is in direct contrast to the culture historical and processual notions that cultures can be viewed as bounded entities (after Barrett 1994; Jones 1997; Wolf 1984). Greater homogeneity in material culture might then be expected at the centers of social ranges, while heterogeneity might be expected at the geographical boundaries of those ranges, especially in cases where social boundaries were permeable and social group affilia tion was fluid. I suggest that what we see in the archaeological record of the Early SideNotched Horizon is a result of multiscalar social and cultural process The subregional technological traditions evident in the Early Side Notched Horizon arguably tr ace ancestry to post Clovis Paleoindian cultural traditions and reflect movement both within and between habitual use areas. Historically, artifact types (including Taylor, Bolen, and Big Sandy) have been proposed ( Bullen 1958, 1975; Kneberg 1956; Michie 1966) to describe at once the regional expression of side notching, and the subregional differences in artifact form across the Southeast. This dissertation empirically documents patterned variation within the Early Side Notched Horizon, with the ultimate goal of identifying Early Archaic social boundaries, and the processes that shaped and transformed them. Hunter gatherers who lived in the Coastal Plain some one hundred centuries ago likely practiced cooperative hunting and fishing, relied on alliances in times of drought and environmental instability, and traded for lithic raw materials and other products of material culture, as well as for the purpose of socialization (cf. Ember 1975; Fitzhugh et al. 2011; Service 1962, 1966; Whallon 2006). These process es all were part of culture building, and can be obscured or rendered unknowable by the normative use of typology (Hawkes 1954; Sassaman 2001a). I would suggest that the best way to investigate Early Archaic

PAGE 292

292 social boundaries and interactions is to conside r the variations within one or more classes of material culture throughout this region, both locally at the level of specific sites and counties, and macroregionally, within and across the drainages of large rivers. These rivers surely were conduits for tr ansportation and trade, and had served as focal points for settlement and aggregation since Paleoindian colonization of the region (Anderson 1990; Miller 2011). I attempted to apply a macroregional approach to the linkages of patterned variation in materia l culture to social identity and affiliation, tacking back and forth between the local and the broader regional landscapes. The results of this approach have been illuminating, both with resp ect to the Early Archaic hunter gatherers whose lifeways led to t he palimpsest we can study today, and in terms of the ways we do archaeology in the Southeast. I am referring here to the ways we are wedded to culture historical artifact types, so much so that hafted biface type names are used as metonyms in daily practi ce. This dissertation uses some 2,020 artifacts from the Early Side Notched Horizon as a case study for analysis of t he complex relationship between artifact variation and cultural affiliation On the basis of ethnographic analogy, the material culture of the Early Side Notched Horizon may reflect various aspects of cultural identity throughout the lower Southeast. For instance, the entire domain of hunting and associated technology may have been controlled by initiated males, who may have been the only group members to have access to certain raw materials or other equipment (after Gould 1969; Goodyear 1995 [1974]; Potter 2004). Specific types of raw material used in the manufacture of side notched tools may have reflected shared knowledge about the landscape, or may have been acquired through large scale social networks or intergroup institutions of gift giving and exchange (after Cooney 1999; Hendon 2000; Whallon 2006; Wiessner 1982). Additionally, these raw materials may reflect direct procurement

PAGE 293

293 by distinct clans ( after Gould 1971, 1978). Early Archaic side notched tools may also have implicitly communicated information about social boundaries, their variation in form referenc ing other people and places (after Hendon 2000; Potter 2004; Tilley 1994; Wiessner 1982; Williams 1982). Unfortunately, it is not possible to empirically test these various ideas at this time, and this case study considers broader spatial and temporal scales than those typically described in ethnographies. However, much of the ethnographic literature leads me to posit the existence of social boundaries during the earliest portion of the Early Archaic. Below, I provide broad interpr etations of the results of this study, reviewing empirical evidence for the existence of a minimum of three subregional traditions within the Early Side Notched Horizon. P atterned variation in the Early Side Notched Horizon may be a reflection of distinct and contemporary social groups with relatively permeable social boundaries. Alternatively, this patterning may be a result of time transgressive trends in group mobility from certain drainage locations to others Specifically, variants of side notched haft ed bifaces throughout the lower Southeast may be the long term consequence of range relocation by groups who moved south to north as the climate warmed (cf. Gunn and Rovner 2003). Subregional Trends in the Early Side Notched Horizon As related throughout Chapter 5, I identified evidence for subregional technological traditions for two classes of side notched flake d stone tools; hafted bifaces and unifaces. Empirical evidence for this claim comes from pe rcentile plots for ratio scale hafted biface and unif ace data, numerous analyses of relative variance for metric variation and tests of association for hafted bifaces and uniface basal configuration. A majority of these lines of evidence point to subregional traditions within the Early Side Notched Horizon, especially with respect to hafted

PAGE 294

294 bifaces, the variation for which historically has been referenced through the use of distinct artifact types (including Taylor, Bolen, and Big Sandy). In discussion of the results of this research, I indicated where the i nequality of variances for certain ratio scale data are not statistically significant. These instances are limited to the following: maximum thickness for hafted bifaces, and any of the ten measured attributes for unifaces. I expected that the most varianc e for nominal and ratio scale artifact attributes would be evidenced for hafted bifaces from inferred zonal boundary areas, relative to the centers of zones where side notched hafted biface types were defined ( Bullen 1958, 1975; Kneberg 1956; Michie 1966). Hafted biface data from the Ocmulgee boundary area, situated at the interface of the Atlantic and Gulf watersheds, and between the Taylor and Bolen Zones, produced the most variance in the zone boundary test case presented in Chapter 5. While a great deal of this variation may be a product of the lack of local highquality raw material, other lines of evidence support interpretation of the Ocmulgee River as archaeologically distinct. For instance, pitted egg stones found as part of Ocmulgee River artifact suites are similar to forms commonly found at Florida sites especially in the Aucilla Suwannee river drainage (Blanton and Snow 1986; Neill 1971; Purdy 1986: 23, 30). I nterpretation of the Ocmulgee River drainage as a boundary area for groups throughout the region may be justified. Can the regions situated to the north and south of the Ocmulgee River boundary area then be interpreted as the ranges of distinct social grou ps, such that the Ocmulgee boundary can be considered a social boundary? While patterned variation for hafted bifaces from these geographically disparate regions is indicative of subregional technological traditions, this ultimate question cannot be defini tively answered with the data that are currently available. However, Ocmulgee River data lend support to the notion that identified subregional

PAGE 295

295 technological traditions reflect contemporary social groups, and call into question the alternative model involving group relocation over time. Based solely on geography, a minimum of three contemporary (at the coarsegrained scale of 700 calendar years) social groups might be postulate d for the Coastal Plain Southeast; one in each watershed, and one in peninsular Florida. Based on the results presented in Chapter 5, I propose that the Coastal Plain Southeast likely was home to at least three macrobands, who occupied multidrainage environments. These include the SanteeCooper/SavannahOgeechee, the Flint/Chattahooc hee, and the Aucilla Suwannee/Tampa Bay. Once again, the empirical basis for these divisions is the inequality of variances for a number of metric attributes for which I collected quantitative data and prepared percentile plots, as well as geographic trend s in basal configuration. I propose that representatives from these three macrobands met periodically at the Ocmulgee River, for marriage alliances, sharing and trade. Information exchange well outside the scope of the co resident group may have been embe dded in these social gatherings, as part of feasts, trade fairs, or planned ceremonies (after Fitzhugh et al. 2011). The notion of universal kinship likely applied to relations between and among macrobands, as hunter gatherers shared many aspects of materi al culture, if not a common language. It is likely that pidgin languages were employed during interactions within areas of aggregation (e.g., along the Ocmulgee Ri ver at the Feronia sites ), and that mate exchange resulted in bilingual and even multilingua l members of societies throughout this region. As detailed throughout Chapter 5, the artifacts in consideration are from distinct drainages throughout this region. For the purposes of the current discussion, I have aggregated these drainages, as these areas are perhaps best understood as macroband terr itories. Use of the

PAGE 296

296 term territory here is not meant to imply the exclusive use of defended areas (after Kelly 1995:163). I was able to identify numerous trends in both continuous and discrete variables for the hafted biface and uniface data I collected from these territories. These trends were illustrated in numerous percentile plots for ratio scale data; a variety of statistical tests (most notably, chisquare and Levenes test of equality of variances) provide supporting evidence for subregional traditions. It should be noted that hafted bifaces from all inferred macroband territories share similar trends with respec t to between notch width and base width data, as previously discussed. Numerous trends for the haft element of unifaces are broadly shared throughout the region. Santee Cooper/Savannah Ogeechee Separation of the SanteeCooper/SavannahOgeechee macroband territory from other portions of the research area is based primarily on trends in basal configuration. Incurvate is the most common base shape for hafted bifaces in this subregion, followed by straight. Straight is the most common type of basal configuration for unifaces here. The second most common base shape for unifaces is not a shared trait across this subregion. In general, and as de monstrated in percentile plots for ratio scale data, SanteeCooper/SavannahOgeechee hafted bifaces share similar haft element trends with tools from the Ocmulgee, Flint, and Chattahoochee River drainages. With respect to the blade portion of tools, trends are not generally shared with Chattahoochee River drainage examples. SanteeCooper/Savannah Ogeechee unifaces do not generally share trends with other macroband territories, with respect to the blade portion, as illustrated by percentile plots of ratio s cale data. The low sample size for unifaces from the SanteeCooper River drainage is noteworthy. Unifaces in this macroband territory share several similar haft element trends (with

PAGE 297

297 the exception of shoulder length) with tools from the Aucilla Suwanne e/Tam pa Bay macroband territory. Flint/Chattahoochee What most obviously distinguishes the FlintChattahoochee macroband territory from others in the region is the lack of unifaces; only one uniface was available for study. Additionally, separation of this ter ritory from others is justified by trends in hafted biface base shape. Straight is the most common base shape f or hafted bifaces in this macroband territory and incurvate is the second most common base shape. Flint/Chattahoochee hafted bifaces share simi lar haft element trends with tools from the SanteeCooper/Savannah Ogeechee macroband territory. For the blade portion of tools, Flint River drainage examples share similar trends with the SanteeCooper/Savannah Ogeechee territory; Chattahoochee River drai nage examples trend more closely with Aucilla Suwannee/Tampa Bay data. Aucilla Suwannee/Tampa Bay Separation of the Aucilla Suwannee/Tampa Bay macroband territory from other portions of the research area is based on trends in hafted biface basal configuration, as well as distinctive trends related to the haft element. The most common type of basal configuration for hafted bifaces in this macroband territory is straight, followed by excurvate. The most common and second most common base shapes for unifaces are not shared traits, although straight and excurvate bases are most prevalent throughout the subregion. In general, and as demonstrated in percentile plots for ratio scale data, AucillaSuwannee hafted bifaces share similar haft element trends that distinguish them from trends observed for

PAGE 298

298 the other study regions. With respect to the blade portion of tools, trends are generally shared with Chattahoochee River drainage examples. Aucilla Suwannee/Tampa Bay unifaces do not generally share trends with other macroband territories, with respect to the blade portion, as illustrated by percentile plots of ratio scale data. Unifaces here share several similar haft element trends (with the exception of shoulder length) with tools from the Santee Cooper/SavannahOg eechee macroband territory Early Archaic Southeast Macroband Territory Characteristics These three macroband territories each afforded different resources for hunter gatherers. The SanteeCooper/SavannahO geechee subregion provided high quality chert raw material at the Allendale chert quarries, as well as numerous creeks, rivers, and Carolina Bays for fishing, waterfowl hunting, potable water, and transportation. The Sante e Cooper and SavannahOgeechee R iver systems all drain into the Atlantic watershed. The Flint/Chattahooch ee subregion also afforded highquality Coastal Plain chert, as well as numerous creeks and rivers for fishing, potable water, and navigation. The Flint and Chattahoochee r iver drainages both are par t of the Gulf wate rshed system. Finally, the Aucilla Suwannee/Tampa Bay subregion, also part of the Gulf watershed, was home to sev eral quarry clusters where high quality cherts could be obtained. During the Early Archaic period, the Florida peninsula endured a period of dr ought known as the Bolen Drought, and settlement close to reliable water sources was a necessity, since the water table was down some seven meters compared to contemporary levels (Dunbar 2002). As a result, macroband mobility may have been more circumscribed in peninsular Florida than I would postulate for othe r portions of the Coastal Plain (cf. Sassaman 2003:9193). As indicated by a pilot study of quarry cluster sourcing of hafted bifaces from both the Aucilla Su wannee and

PAGE 299

299 Tampa Bay drainages, most side notched artifacts from these drainages can be sourced to locally available raw material. However, s ome degree of movement between the AucillaSuwannee and Tampa Bay drainages is indicated by the presence of six (out of a total of 118 analyzed) hafted bifaces apparently manufactured from nonlocal cherts. I propose that, while band mobility was largely focused along drainages, cross drainage movement was common for individual bands, for resource procurement as well as socialization (after Whallon 2006). Addi tionally, I propose that band members interacted with fellow macroband members on a regular basis in areas of habitual use, and that macrobands, comprised of individual bands from numerous adjacent drainages throughout the Coastal Plain, interacted with one another on a semi annual basis, in the Ocmulgee River drainage. Figure 61 presents the proposed geographic regions of three Coastal Plain macrobands. The Ocmulgee R iver drainage is approximately equidistant to surrounding groups; the SanteeCooper/Savannah Ogeechee macroband territory is situated to the northeast, the Aucilla Suwannee/Tampa Bay subregion is located to the south, and the Flint/Chattahoochee macroband territory is situated to the west. Also, the Ocmulgee River drainage extends across the i nterface of th e Gu lf and Atlantic watersheds ( Figure 4 12). Previous researchers have noted that this boundary likely influenced Early Archaic settlement patterns (Anderson and Hanson 1988; Blanton and Snow 1986, 1989). Its cent ral position between areas o f habitual use made the Ocmulgee R iver drainage an ideal location for social aggregation by members of macrobands from throughout the Coastal Plain Southeast. This apparently was true despite the absence of high quality lithic raw materials in the Ocmulgee River drainage. As discussed in Chapter 5, raw material curation would have been necessary for groups visiting this region. The location of the Feronia Site Complex suggests that ancient aggregation may have occurred at any zonal

PAGE 300

300 interface, whether at the Piedmont/Coastal Plain (Fall Line), or at the interface of the Atlantic and Gulf watersheds (Blanton and Snow 1989) and that the key was aggregation in places that would have been easy to find, describe, and relocate (after Miller 2011). That social aggr egation is highly indicated for Ocmulgee in spite of the absence of locally available cherts, underscores the primary importance of social group interactions during the Early Archaic (after Sassaman 2011). This aggregation may have taken multiple forms, some aspects of which were postulated in the first regional treatment of Eastern prehistory (Griffin 1952). According to that treatment, seasonal ceremonies and festivities likely occurred during congregations of large groups at foodrich locations, facilitating the exchange of information about cultural traditions as well as features of the broader regional landscape (Griffin 1952:355). As previously discussed, periodic aggregations are known to be common f eatures of hunter gatherer life, and have been well documented in the ethnographic literature (Kelly 1995:183188). Most commonly, aggregations are seen to occur in course of other activities where mobility patterns intersect, or where concentrations of specific resources draw groups together (Fitzhugh et al. 2011). Perhaps Early Archaic groups from throughout the Coast al Plain viewed the Ocmulgee R iver drainage as a key destination; a super waterhole (after Hiatt 1968:100). As other researchers have noted, there is much more to aggregation than reproductive fitness or subsistence ecology, and the social and ritual com ponents of social aggregation should not be minimized (Conkey 1980; Whallon 2006) Interactions among distinct social groups, such as those postulated for the Ocmulgee River drainage, likely reinforced social boundaries throughout the region (after Barth 1969; Parker 2006; Stark 1998). Ethnographically, we can see that hunter gatherers commonly come together for social and ritual reasons, and frequently

PAGE 301

301 aggregate for the primary purpose of extending their social networks (Damas 1969; Lee 1972, 1979). These ritual and social processes may well have occurred concomitantly with more economic activities (such as cooperative hunting and fishing) at ancient hunter gatherer aggregation locales such as the Feronia sites along the Ocmulgee River Few ancient aggregat ion sites have been identified in the Early Archaic Southeast, likely because the majority of research has not been regionally oriented in a way that is sensitive to this sort of patterned variation (after Conkey 1980). Numerous examples of equifinality ar e presented by patterned variation in the Early Side Notched Horizon. Chief among these are the widespread distribution of side notched hafted bifaces across the Southeast, and the subregional variations within that cultural horizon. First, t his variation may reflect aspects of l ocalized l earning patterns unequal skill levels, or various other processes that may or may not be a direct function of cultural choice (cf. Bamforth and Finlay 2008; Gosselain 2000) A lternatively, and a s previously noted, it is possible that subregional variations resulted from the movement of groups in response to climate change (cf. Gunn and Rovner 2003). This scenario holds that, as the Pleistocene Holocene transition led to the northward movement of isotherms, it resulted in t he northward migration of elk and bison. Groups using side notched toolkits (here, called Big Sandy variants) (Gunn and Rovner 2003:85) also moved north, resulting in an archaeological record of side notched variants throughout the Atlantic Slope. The pr imary evidence for this argument is the later dates of certain similar side notched forms in Virginia, compared to South Carolina. Currently it is not possible to test this scenario, as precise dates are lacking for the study data set, as well as for the r egion. Additionally, direct evidence of elk and bison hunting is not available for Southeastern sites east of the Appalachian Mountains; evidence connecting side notched variants to the

PAGE 302

302 northward movement of isotherms and specific game migrations remains circumstantial (Gunn and Rovner 2003:85). A growing body of evidence suggests that subregional technological traditions began post Clovis in a number of locations throughout the Southeast (Smallwood 2012; Thulman 2006). As reviewed in Chapter 3, certain (a rtifact based) Paleoindian evidence lends support to the staging area model (Anderson 1990, 1996) for initial widespread colonization and settlement, as does recent GIS based research into possible Paleoindian aggregation locations (Anderson and Sassaman 2 012:5051; Miller 2011). I would argue that the origin of subregional cultural traditions in the Early Side Notched Horizon can be traced to similar Paleoindian traditions, and that the persistence of these subregional traditions is arguably lin ked to sett lement patterns anticipated by the staging area hy pothesis. Large S cale Social Interaction in the Early Archaic Together with certain ethnographic studies, recently developed models for social networking and informationsharing have special relevance for investigating archaeological cases of social intera ction at great regional scales. Specifically, the notion of universal kinship within a large region (Barnard 1978, 2011) and the likelihood of informal regional networks (Fitzhugh et al. 2011; Whallon 2006; Whallon et al. 2011; after Wiessner 1982) may be applicable to the Early Side Notched Horizon. Members of the Aucilla Suwannee/Tampa Bay macroband may have been relatively tethered to specific Tertiary karsts for both water and highquality raw material, as their entire subregion was affected by the Bolen Drought (cf. Sassaman 2003) These groups may have been more closely tied to specific locations for raw material and for water, as well as for settlement and social interaction, than what I would propose for their contemporaries in other portions of

PAGE 303

303 the Coastal Plain. Information sharing specifically for the exchange of ecological information may have occurred much more frequently in Florida than elsewhere in the Coastal Plain (after Fitzhugh et al. 2011:9697) S ome evidence exists for the emergence of regionalization among Floridas hunter gatherers as early as the end of the Early Paleoindian period (Thulman 2006:219). Perhaps we can view Floridas Early Archaic groups as pa rt of a descendent tradition, who live d in the best possible places within their immediate social boundaries, given the clear environmental constraints they faced ( after Dunbar 1991). Alternatively, these environmental constraints may have required partici pation in a largescale information sharing network, and members of the Aucilla Suwannee/Tampa Bay macroband may have invested in a wide range of networks both locally and regionally, with many strong ties maintained at all scales of interaction. This pattern fits well with the fully integrated networks predicted b y Fitzhugh et al. (2011:9798) for cases of low environmental predictability and low interaction costs. Periodic interactions with residents of the Flint/Chattahoochee macroband territory would have created and maintained social bonds among large groups of hunter gatherers. Long distance social networks may have been key to hunter gatherer life, as they would have provided ease of movement through the landscape (after Gould 1982; Whallon 2006). A sense of universal kinship throughout the Coastal Plain may well have developed from social relationships forged out of necessity, maintained by macroband representatives, and historicized through myth. These very acts of sharing may have created the geog raphically extensive social networks reflected in Early SideNotched material culture. These cultural and historical processes arguably are visible today in the patterned variation in side notched artifacts from throughout the Coastal Plain.

PAGE 304

304 Large scale sharing networks likely always included the sharing of information, about environmental conditions as well as the social landscape. Certain commodities also were likely in trade, creating and affirming relationships across great distances. It is not possible to describe all of the commodities that were traded among Early Archaic Southeasterners. Scenarios of long distance trade of lithic raw material can be inferred for portions of the Coastal Plain. Specifically, evidence from Florida indicates the transpo rt of raw material or finished bifaces across great distances. The pilot study conducted for this research project identified two side notched hafted bifaces from the Lake Bird site (in Taylor County, Florida) that seem to have been manufactured from chert s originating over 250 km to the south, at the Hillsborough River. Ad ditionally, movement some 150 km to the south is indicated by one biface from the Jeanies Better Back site (in Lafayette County, Florida), which may have been manufactured from Brooksvil le quarry cluster chert (Austin and Mitchell 1999). These findings may provide support for the existence of long distance alliances within the Aucilla Suwannee/Tampa Bay macroband territory. Alternatively, they may reflect the range of annual mobility for peninsular Florida. The notion of landscape as kinscape stands to inform archaeological discourse about ancient hunter gatherers. While I have necessarily produced numerous maps with which to convey my findings, it is important to consider the very differe nt ways in which Early Archaic Southeasterners may have viewed the landscapes in which they lived. These groups may well have created kinscapes, or maps of social relationships, through which they negotiated and related their ties to one another and to l and (after Bender 1999:36). Many native peoples are known to have complex relationships with their landscapes, which in turn influence their social relationships (after Basso 1983: Morphy 1995; Smith 1999).

PAGE 305

305 H unter gatherers also may have had totemic ties t o land, reckoning descent to specific points of interest in the landscape, such as individual river drainages or chert quarries (after Gould 1978). They also likely had kinbased relationships to places in the landscape such as burial grounds, former home sites, and locations where rites of initiation were performed (after Hitchcock and Bartram 1998). Additionally, ritual acts such as long distance gifting and the manufacture of specially crafted bifaces can be interpreted as biographical processes. These c an be read as historical productions, since they were connected to people whose own biographies could be traced to different times and places ( after Sassaman 2011). In this way, places throughout the landscape of the Coastal Plain may well have been integral parts of social and historical memories, expressing multiple scales of social identities and relationships, and relating at once to social group organization, pr actice, and process (after Hendon 2000; Morphy 1995). No adequate term exists to denote the great regional scale of social interaction presented here for the Early Archaic Southeast. Perhaps, in the spirit of the related concept of landscape as kinscape, w e can imagine the broader landscape of the Coastal Plain Southeast as a bandscape a collectivity comprised of multiple macrobands who were engaged in broadly similar cultural traditions. Certainly, patterned variation in side notched artifacts lends itself to an appreciation of macroregional social reproduction greater than that typically described by anthropologists. Summary and Recommendations for Future Research There is some basis for the continuing use of types (including Taylor, Bolen, and Big Sandy ) as time markers for the Early SideNotched Horizon, as they were defined through trial an d error, and not by assumption, and historically have been used to distinguish discernible differences in typical forms ( Bullen 1958, 1975; Kneberg 1956; Michie 1966). I identified measurable differences in basal configuration and numerous ratio scale attributes that map onto

PAGE 306

306 geography throughout the Coastal Plain. In that sense, these types have been effectively used as generalizations to monitor cha nge through time and space (after Thomas 1986:622). However, due to the patterned variation evident in high frequencies of tools throughout the Coastal Plain, I recommend use of the term Early Side Notched to replace the artifact type names for hafted bifaces (after Dris kell 1996; Morse 1994; Randall 2002). We should seek the most parsimonious distinction between side notched artifact assemblages, as the use of type names has led to confusion in reporting investigation results throughout the region (cf. Blanton and Snow 1986, 1989; Carter and Dunbar 2006; Gunn and Rovner 2003; Michie 1996; Randall 2002). If we can view artifact types as historical, they no longer appear to be so discrete and bounded. Artifact types cannot be arbitrarily defined entities, and at the same t ime the products of historical process. The extent to which we can see artifact types as discrete depends on our spatial and temporal scales. This is wh y, looking across the many kilometers of the Southeastern Coastal Plain, we must define Early Archaic si de notched artifacts as the smallest cluster of individual artifacts that share variation. Patterned variation in material culture has significance, since we are studying change and social process, and not just difference (after Wolf 1984). This is the com plex web of history and artifact types in the Early Archaic Southeast (after Dunnell 1986:154). The best indicator for regionalization in the Coastal Plain based on data collected to date is basal configuration, trends in which can be seen to shift at the Ocmulgee River, and then again to the south and west. This regionalization, I would argue, is a product not of isolation, but of inte raction with neighboring hunter gatherers. When in teractions were routine, hunter gatherers

PAGE 307

307 likely saw themselves as distin ct from people in other groups, and differences became asserted and magnified (after Potter 2006; Sassaman 2001b). Throughout Chapter 5, I demonstrated that side notched tool base shape is directly correlated to geographic region, for both hafted bifaces and unifaces. Therefore, base shape may be an indicator of macroband affiliation. Table 6 1 shows the percentages of basal configuration for hafted bifaces and unifaces for each of the three inferred macroband territories. These data are aggregations of da ta already presented in Chapter 5, where they appeared with respect to individual drainages. Table 6 1. Percentages of basal confi guration by macroband territory Macroband territory Hafted Bifaces Edgefield Scrapers Incurvate Straight Excurvate Incurvate Straight Excurvate Santee Cooper/ Savannah Ogeechee 67 25 8 28 49 23 Flint/Chattahoochee 40 57 3 100 Aucilla Suwannee/Tampa Bay 18 38 44 8 40 52 Basal configuration cannot be seen as epiphenomenal with respect to the Early Archaic Southeast. I have suggested possible scenarios for observed differences in base shapes for hafted bifaces and unifaces throughout the region. Regardless of the many possi ble reasons behind these apparent patterns in basal configuration, this obviously is an avenue to pursue in future research. Basal configuration is the one aspect of hafted bifaces that cannot be viewed while the tool is hafted. As such, it is unlikely tha t biface base shape was related to the purposeful signaling of group identity (cf. Thulman 2006). I doubt that cultural resource management archaeologists will discontinue use of type names for side notched tools, since their utility lies mainly in their metonymic qualities. When an archaeologist recovers a sidenotched hafted biface in eastern Georgia, and calls it Taylor, the

PAGE 308

308 type name alone brings to mind an image of that artifact, which in turn brings to mind the fact that this artifact dates to the earliest portion of the Early Archaic, and that similar forms can be found throughout the Southeaster n Coastal Plain. Continued use of the Edgefield Scraper type name to describe a regionally expressed side notched uniface is recommended. This tool type is geographically restricted, from central South Carolina in the Santee Cooper River drainage, south to northcentral Florida. The type also extends west into central Georgia, in the Ocmulgee River drainage. Future research should attempt to clarify the position of Edgefield Scrapers throughout the Early Archaic. While this artifact type typically is consid ered to date to the earliest portion of the Early Archaic, and is most frequently recovered in association with side notched hafted bifaces, it also has been recovered in association with later corner notched artifacts (Sassaman et al. 2002). Additionally, microwear analyses of Edgefield Scrapers from throughout the Coastal Plain may allow for a better understanding of their use, especially in light of their frequent interpretation as seasonal technology ( after Goodyear et al. 1980). Future research should attempt to learn the raw material sources for Early Archaic artifacts. Some attempts have been made toward this end in western South Carolina and Florida, and the results of quarry cluster sourcing in Florida have been particularly rewarding, as they have a dded to our knowledge of hunter gatherer mobility (Austin and Mitchell 1999). Much more research is needed to identify the most commonly used lithic quarries in the Coastal Plain, and to establish where artifacts made from those raw materials were used an d discarded. For example, while Eocene age cherts are presumed to be the primary source for artifacts found at the Feronia locality, no microscopic analyses have been conducted on this material A deter mination of long distance movement of r aw material or bifaces to these locations from

PAGE 309

309 Florida or southwestern South Carolina would obviously enhance our interpretations of the Ocmulgee River drainage. Data for original lithic package sizes throughout the Coastal Plain also are needed. It may eventually be pos sible to explain the size differences between South Carolina and Florida side notched artifacts as a function of that variable. Admittedly, formal hafted tools such as bifaces and unifaces provide only a glimpse of the whole picture of the Early Side Notch ed Horizon. Future researchers would do well to perform debitage analysis at stratified sites, and to investigate patterns of raw material acquisition and specific lithic reduction and tool production sequences (cf. Austin and Mitchell 2010). T his analysi s is based on a select sample of Early SideNotched Horizon artifacts, in that I did not equally and arbitrarily obtain samples from all portions of the research area. Rather, I examined artifacts where known collections were accessible for study, and geog raphical groupings were created a priori as part of analysis of variation through space. Whereas I retained spatial data and used locational information as an independent variable throughout analyses, f uture researchers might consider randomly selecting data from throughout the Coastal Plain in order to conduct blind comparative analyses. Additionally, the use of multivariate statistical tools in future analyses of the Early SideNotched Horizon may allow for the simultaneous examination of relationships am ong multiple ratio scale attributes Finally, improved models are needed for explaining Early Archaic settlement and subsistence patterns. It is no longer sufficient to assume that these groups relied mostly on deer for their protein, or to assume movement on a seasonal scale (after Anderson and Hanson 1988). Aquatic resources likely formed a large part of the Early Archaic diet, and while no watercraft yet has been found for this period in the Southeast, these groups likely were proficient paddlers

PAGE 310

310 and fis hers. Future research of Early Archaic archaeological components should implement recovery techniques for data that may allow for ethnobotanical and faunal analyses, soil and other conditions permitting (after Doran and Dickel 1988; Driskell 1996; Tuross e t al. 1994). The notion of a seasonal round has aided archaeologists who may be frustrated by interpreting an incomplete record. However, while patterns of seasonal mobility are based on solid ethnographic research, they run the risk of masking the variabi lity that surely ch aracterized the lives of hunter gatherers. Differences within seasons, and from year to year, should not be underemphasized, and archaeologists mus t be wary of approaching hunter gatherer variability purely in terms of environmental vari ability, in any event (Jochim1991). It is possible that the Coastal Plain regions considered for this research were hinterlands for populations who focused primarily on coastal regions (cf. Thulman 2006:218). Early Holocene groups likely spent a great deal of time on coastlines that are now submerged along the continental shelf. Currently, paleolandscapes remain underexplored toward numerous questions of anthropological interest, and future efforts at regional modeling with respect to the Early Side Notched Horizon should be designed to correct for this where possible (Anderson and Sassaman 2012:40; Faught 2004b; Faught and Waggoner 2012). Some models have been presented which take social networks into account, but as anthropologists, we would do well to rem ember that social networks serve more than mating purposes. People define themselves and their social groups in large part according to who is different, who is not like them, for a variety of reasons. There likely were different language groups in the Southeastern Coastal Plain during the Early Archaic. There also must have been band affiliation, as well as a sense that one belonged to a group larger than his or her band. This larger group was a macroband that likely spanned numerous river drainages, and f rom which

PAGE 311

311 marriage and trading partners were sought, and with whom allies could be found during times of social or environmental instability. As archaeologists who must deal with a record that is fragmentary at the best of times, we realize we can be sever ely limited by the data at our disposal. We also realize that we will never know exactly what happened in the past, whether in the Early Archaic, or in any other period of time, in the Southeast or elsewhere. However, archaeologists can appreciate larger t emporal and spatial scales than can many of our colleagues in any other subfield of anthropology. For example, many ethnographic studies of hunter gatherers have been hampered by a limited field of vision, resulting in a worms eye view of reality (Wobst 1978:304). This largely has been a function of the need to do salvage ethnography, during which there was little concern in documenting regional and interregional social process. In contrast, archaeologists can document the dynamics of multiscalar groups in large units of space, and are well equipped to investigate spatiotemporal variability. Thus, archaeologists are uniquely positioned to make grand statements about specific groups of people, and these stand to powerfully affect the ways in which we under stand some of the earliest peoples in North America. That will serve as a contribution not only to our own field, but also to history writ large.

PAGE 312

312 Figure 6 1. Proposed Coastal Plain macroband territories

PAGE 313

313 APPENDIX A BASAL CONFIGURATION FOR HAFTED BIFACES Table A 1. Basal configuration for Santee Cooper drainage Santee Cooper drainage Elloree Museum Elloree Bank Skinner Hebert Total Incurvate 3 18 22 3 46 Straight 3 4 8 2 17 Excurvate 0 0 1 0 1 Total 6 22 31 5 64 Table A 2. Basal configuration for SavannahOgeechee drainage Savannah Ogeechee drainage Big Pine Tree Site Charles Site Topper Site Strong Owl and Cotton Creek Sites Burke County Total Incurvate 10 3 13 267 80 20 393 Straight 2 0 1 102 39 5 149 Excurvate 2 0 0 34 15 0 51 Total 14 3 14 403 134 25 593 Table A 3. Basal configuration for Ocmulgee drainage Ocmulgee drainage Feronia sites Coffee County Total Incurvate 16 13 29 Straight 17 14 31 Excurvate 2 4 6 Total 35 31 66 Table A 4. Basal configuration for Flint drainage Flint drainage Dooly County Worth County Total Incurvate 16 15 31 Straight 26 35 61 Excurvate 1 1 2 Total 43 51 94

PAGE 314

314 Table A 5. Basal configuration for Chattahoochee drainage Chattahoochee drainage Early County Miller County Barbour County Henry County Houston County Geneva County Jackson County Total Incur v ate 13 19 29 9 23 31 31 155 Straight 9 75 11 13 17 35 46 206 Excurvate 3 3 1 1 3 0 3 14 Total 25 97 41 23 43 66 80 375 Table A 6. Basal configuration for Aucilla Suwannee drainage Aucilla Suwannee drainage Applewhite Site Jefferson County Lake Bird Site Jeanie's Better Back Site Hendrix Total Incurvate 0 5 4 2 72 83 Straight 74 9 35 13 86 217 Excurvate 6 0 34 6 108 154 Total 80 14 73 21 266 454 Table A 7. Basal configuration for Tampa Bay drainage Tampa Bay drainage Harney Flats Site Harris Grove Kepler Site Pinellas County St. Petersburg Area Spring Hill Site Pasco County Total Incurvate 2 6 0 3 1 2 0 14 Straight 1 24 0 7 0 3 1 36 Excurvate 2 14 2 4 2 2 1 27 Total 5 44 2 14 3 7 2 77

PAGE 315

315 APPENDIX B METRIC ATTRIBUTES OF HAFTED BIFACES BY SITE AND COLLECTION AREA Santee Cooper Sites and Collection Areas Table B 2. Elloree Museum hafted biface metric attributes Elloree Museum sidenotched hafted bifaces Mean Min Max Frequency (out of 6) recorded s.d. c.v. Blade length (mm) 29.17 20.72 36.09 4 6.47 0.22 Shoulder width (mm) 20.85 16.98 23.72 5 2.62 0.13 Maximum thickness (mm) 6.53 4.86 8.1 6 1.29 0.2 Shoulder length (mm) 10.06 8.35 11.53 6 1.4 0.14 Between notch width (mm) 16.2 14.97 17.55 6 1.04 0.06 Notch height (mm) 5.27 4.7 6.42 6 0.64 0.12 Base width (mm) 21.73 20.18 23.8 6 1.38 0.06 Table B 3. Elloree Bank hafted biface metric attributes Elloree Bank side notched hafted bifaces Mean Min Max Frequency (out of 22) recorded s.d. c.v. Blade length (mm) 29.72 19.66 61.8 20 10.93 0.37 Shoulder width (mm) 21.61 16.04 28.98 22 3.72 0.17 Maximum thickness (mm) 6.31 4.16 8.51 22 1.18 0.19 Shoulder length (mm) 10.02 7.2 15.35 22 1.97 0.2 Between notch width (mm) 16.24 11.28 22.23 22 2.7 0.17 Notch height (mm) 4.84 3.14 6.66 22 0.98 0.2 Base width (mm) 22.71 18.79 27.06 18 2.75 0.12

PAGE 316

316 Table B 4. Skinner Collection hafted biface metric attributes Table B 5. Hebert Collection hafted biface metric attributes Skinner Collection side notched hafted bifaces Mean Min Max Frequency (out of 33) recorded s.d. c.v. Blade length (mm) 31.64 19.49 52.34 31 7.82 0.25 Shoulder width (mm) 21.81 15.84 31.31 26 3.88 0.18 Maximum thickness (mm) 6.52 4.89 8.51 33 0.76 0.12 Shoulder length (mm) 10.19 7.64 15.67 32 1.75 0.17 Between notch width (mm) 15.35 11.11 19.65 33 2.33 0.15 Notch height (mm) 4.79 3.05 8.11 32 0.89 0.18 Base width (mm) 21.8 17.25 30.13 23 3.77 0.17 Hebert Collection side notched hafted bifaces Mean Min Max Frequency (out of 5) recorded s.d. c.v. Blade length (mm) 30.82 20.99 46.96 5 10.44 0.34 Shoulder width (mm) 21.5 19.38 26.4 5 2.84 0.13 Maximum thickness (mm) 7.1 5.2 9.65 5 1.73 0.2 Shoulder length (mm) 10.49 8 13.74 5 2.11 0.2 Between notch width (mm) 15.4 14.51 17.09 5 1.07 0.07 Notch height (mm) 5.18 3.25 6.3 5 1.23 0.24 Base width (mm) 22.32 19.26 26.55 5 2.66 0.12

PAGE 317

317 Savannah Ogeechee Sites and Collection Areas Table B 6. Big Pine Tree site hafted biface metric attributes Big Pine Tree site side notched hafted bifaces Mean Min Max Frequency (out of 15) recorded s.d. c.v. Blade length (mm) 35.93 21.23 52.48 13 9.44 0.26 Shoulder width (mm) 23.75 17.46 32.6 8 4.91 0.21 Maximum thickness (mm) 6.96 5.47 8.6 15 0.82 0.12 Shoulder length (mm) 11.96 7.91 26.15 15 4.4 0.37 Between notch width (mm) 16.68 11.96 21.3 14 2.9 0.17 Notch height (mm) 6.2 3.59 10.6 15 2.04 0.33 Base width (mm) 23.03 16.66 28.49 11 3.46 0.15 Table B 7. Charles site hafted biface metric attributes Charles site side notched hafted bifaces Minimum value Maximum value Frequency (out of 3) recorded Blade length (mm) 0 Shoulder width (mm) 22.98 22.98 1 Maximum thickness (mm) 5.18 5.68 3 Shoulder length (mm) 9.11 10.86 3 Between notch width (mm) 18.75 21.49 3 Notch height (mm) 4.82 6.72 3 Base width (mm) 24.01 27.89 2

PAGE 318

318 Table B 8. Topper site hafted biface metric attributes Topper site side notched hafted bifaces Mean Min Max Frequency (out of 15) recorded s.d. c.v. Blade length (mm) 33.67 12.52 47.45 10 10.71 0.32 Shoulder width (mm) 22.06 18.92 26.47 8 2.57 0.12 Maximum thickness (mm) 6.45 4.72 8.17 13 0.94 0.15 Shoulder length (mm) 10.25 7.59 12.71 15 1.74 0.17 Between notch width (mm) 15.04 10.47 17.35 15 2.05 0.14 Notch height (mm) 5.18 3.37 7.9 13 1.34 0.26 Base width (mm) 22.66 17.04 28.6 13 3.25 0.14 Table B 9. Strong Collection hafted biface metric attributes Strong Collection side notched hafted bifaces Mean Min Max Frequency (out of 411) recorded s.d. c.v. Blade length (mm) 30.31 12.2 69.01 212 9.6 0.32 Shoulder width (mm) 21.71 12.9 35.57 261 3.25 0.15 Maximum thickness (mm) 6.43 3.89 9.35 385 0.95 0.15 Shoulder length (mm) 10.18 6.66 19.8 398 1.8 0.18 Between notch width (mm) 15.4 10.22 25.32 393 2.55 0.16 Notch height (mm) 5.35 2.7 13.67 390 1.28 0.24 Base width (mm) 21.78 10.57 39.68 368 3.64 0.17

PAGE 319

319 Table B 10. Owl and Cotton Creek sites hafted biface metric attributes Owl and Cotton Creek sites side notched hafted bifaces Mean Min Max Frequency (out of 134) recorded s.d. c.v. Blade length (mm) 36 17.56 68.06 118 9.57 0.27 Shoulder width (mm) 22.6 15.55 36.68 101 3.96 0.18 Maximum thickness (mm) 6.83 4.77 10.95 133 1.03 0.15 Shoulder length (mm) 10.7 1.53 15.77 134 1.9 0.18 Between notch width (mm) 15.25 10.35 21.97 134 2.24 0.15 Notch height (mm) 5.34 3.29 8.96 134 1.18 0.22 Base width (mm) 23.26 16.44 29.88 125 2.8 0.12 Table B 11. Burke County hafted biface metric attributes Burke County side notched hafted bifaces Mean Min Max Frequency (out of 25) recorded s.d. c.v. Blade length (mm) 39.03 23.34 62.64 25 9.67 0.25 Shoulder width (mm) 22.93 17.22 35.31 21 4.41 0.19 Maximum thickness (mm) 6.56 5.46 8.39 24 0.74 0.11 Shoulder length (mm) 9.94 7.76 13.08 25 1.3 0.13 Between notch width (mm) 15.68 12.64 21.07 25 2.46 0.16 Notch height (mm) 5.22 3.51 7.35 25 1.04 0.2 Base width (mm) 22.6 17.82 29.63 25 3.15 0.14

PAGE 320

320 Ocmulgee Sites and Collection Areas Table B 12. Feronia sites hafted biface metric attributes Feronia sites side notched hafted bifaces Mean Min Max Frequency (out of 37) recorded s.d. c.v. Blade length (mm) 28.03 12.64 48.74 33 9.72 0.35 Shoulder width (mm) 22.33 15.25 34.92 32 4.81 0.22 Maximum thickness (mm) 6.73 5.08 9.02 35 0.82 0.12 Shoulder length (mm) 10.63 8.35 14.19 35 1.53 0.14 Between notch width (mm) 15.88 10.65 24.34 37 3.77 0.24 Notch height (mm) 5.5 3.67 8.92 35 1.22 0.22 Base width (mm) 22.71 16.05 30.1 29 3.79 0.17 Table B 13. Coffee County hafted biface metric attributes Coffee County side notched hafted bifaces Mean Min Max Frequency (out of 37) recorded s.d. c.v. Blade length (mm) 25.56 13.57 48.51 24 9.23 0.36 Shoulder width (mm) 21.78 14.87 36.52 24 5.05 0.23 Maximum thickness (mm) 6.76 3.43 9.75 37 1.12 0.17 Shoulder length (mm) 11.02 6.8 15.87 35 2.19 0.2 Between notch width (mm) 15.4 11.54 23.01 32 2.71 0.18 Notch height (mm) 5.74 3.34 8.84 36 1.28 0.22 Base width (mm) 24.06 16.63 32.62 23 4.15 0.17

PAGE 321

321 Flint Sites and Collection Areas Table B 14. Dooly County hafted biface metric attributes Dooly County side notched hafted bifaces Mean Min Max Frequency (out of 45) recorded s.d. c.v. Blade length (mm) 31.37 15.65 47.42 27 7.83 0.25 Shoulder width (mm) 20.48 3.2 8.93 42 4.27 0.21 Maximum thickness (mm) 6.58 4.98 8.94 45 0.98 0.15 Shoulder length (mm) 10 6.9 17.14 44 1.97 0.2 Between notch width (mm) 14.74 10.2 18.47 45 1.49 0.1 Notch height (mm) 5.26 16.42 32.15 16 1.16 0.22 Base width (mm) 22.72 16.89 29.52 25 3.02 0.13 Table B 15. Worth County hafted biface metric attributes Worth County side notched hafted bifaces Mean Min Max Frequency (out of 52) recorded s.d. c.v. Blade length (mm) 31.21 18.05 59.14 42 8.49 0.27 Shoulder width (mm) 22 3.62 8.8 51 4.08 0.19 Maximum thickness (mm) 6.6 4.84 9.17 52 0.92 0.14 Shoulder length (mm) 10.7 6.95 15.58 52 1.68 0.16 Between notch width (mm) 14.92 11.37 19.24 52 1.9 0.13 Notch height (mm) 5.6 16.77 30.55 18 1.13 0.2 Base width (mm) 23.53 17.99 30.89 48 2.96 0.13

PAGE 322

322 Chattahoochee Sites and Collection Areas Table B 16. Early County hafted biface metric attributes Early County side notched hafted bifaces Mean Min Max Frequency (out of 26) recorded s.d. c.v. Blade length (mm) 33.91 17.54 44.19 12 6.96 0.21 Shoulder width (mm) 25.75 15.92 32.34 15 4.21 0.16 Maximum thickness (mm) 7.17 5.18 9.31 26 1.07 0.15 Shoulder length (mm) 10.54 7.72 14.37 26 1.73 0.16 Between notch width (mm) 16.71 12.04 21.56 26 2.53 0.15 Notch height (mm) 5.7 4.08 9.07 25 1.17 0.2 Base width (mm) 25.23 21.62 28.71 13 2.14 0.08 Table B 17. Miller County hafted biface metric attributes Miller County side notched hafted bifaces Mean Min Max Frequency (out of 97) recorded s.d. c.v. Blade length (mm) 33.48 18.19 49.21 54 6.26 0.19 Shoulder width (mm) 25.1 19.6 32.21 36 2.87 0.11 Maximum thickness (mm) 7.26 5.39 10.54 68 0.96 0.13 Shoulder length (mm) 10.63 7.26 14.09 68 1.52 0.14 Between notch width (mm) 14.91 11.27 21.32 68 1.95 0.13 Notch height (mm) 5.89 3.59 8.55 67 1.11 0.19 Base width (mm) 23.11 18.33 28.24 56 2.04 0.09

PAGE 323

323 Table B 18. Barbour County hafted biface metric attributes Barbour County side notched hafted bifaces Mean Min Max Frequency (out of 43) recorded s.d. c.v. Blade length (mm) 32.43 20.04 56.38 29 8.64 0.27 Shoulder width (mm) 23.96 17.42 32.33 20 4.37 0.18 Maximum thickness (mm) 6.98 4.93 9.35 41 0.9 0.13 Shoulder length (mm) 10.06 6.38 14.76 40 1.94 0.19 Between notch width (mm) 15.27 11.73 26.86 42 2.64 0.17 Notch height (mm) 5.7 2.99 8.22 42 1.17 0.21 Base width (mm) 22.74 18.77 33.04 38 2.68 0.12 Table B 19. Henry County hafted biface metric attributes Henry County side notched hafted bifaces Mean Min Max Frequency (out of 23) recorded s.d. c.v. Blade length (mm) 50.93 33.72 80.22 23 11 0.22 Shoulder width (mm) 29.27 22.12 36.62 21 4.23 0.14 Maximum thickness (mm) 7.1 5.7 9.4 23 1.02 0.14 Shoulder length (mm) 10.51 8.09 14.1 23 1.63 0.15 Between notch width (mm) 16.57 12.35 22.5 23 2.85 0.17 Notch height (mm) 5.63 3.42 8.25 23 1.38 0.24 Base width (mm) 25.13 20.01 30.33 22 2.63 0.1

PAGE 324

324 Table B 20. Houston County hafted biface metric attributes Houston County side notched hafted bifaces Mean Min Max Frequency (out of 48) recorded s.d. c.v. Blade length (mm) 45.07 18.36 81.69 40 13.7 0.3 Shoulder width (mm) 28.7 17.8 44.87 32 6.07 0.21 Maximum thickness (mm) 7.6 6.04 10.24 40 0.95 0.12 Shoulder length (mm) 10.49 6.58 13.82 40 1.68 0.16 Between notch width (mm) 15.91 13.13 24.33 40 2.65 0.17 Notch height (mm) 5.65 3.32 7.89 40 1.01 0.18 Base width (mm) 24.69 18.52 33.71 38 3.36 0.14 Table B 21. Geneva County hafted biface metric attributes Geneva County side notched hafted bifaces Mean Min Max Frequency (out of 66) recorded s.d. c.v. Blade length (mm) 36.44 19.76 68.51 39 11.6 0.32 Shoulder width (mm) 23.48 17.55 33.47 30 3.57 0.15 Maximum thickness (mm) 6.91 3.57 8.17 42 0.79 0.11 Shoulder length (mm) 10.12 7.56 15.45 42 1.51 0.15 Between notch width (mm) 14.57 11.44 21.04 42 1.99 0.14 Notch height (mm) 5.73 3.85 8.17 42 1.11 0.19 Base width (mm) 22.19 17.86 29.72 35 2.9 0.13

PAGE 325

325 Table B 22. Jackson County hafted biface metric attributes Jackson County side notched hafted bifaces Mean Min Max Frequency (out of 81) recorded s.d. c.v. Blade length (mm) 39.07 21.59 66.5 77 9.25 0.24 Shoulder width (mm) 26.36 18.17 39.74 54 4.4 0.17 Maximum thickness (mm) 7.36 5.28 9.76 79 0.94 0.13 Shoulder length (mm) 10.52 6.14 16.76 79 1.68 0.16 Between notch width (mm) 15.14 11.56 26.75 79 2.61 0.16 Notch height (mm) 5.56 3.65 8.33 78 1.02 0.18 Base width (mm) 24.25 19.3 33.88 74 2.76 0.11 Aucilla Suwannee Sites and Collection Areas Table B 23. Applewhite site hafted biface metric attributes Applewhite site side notched hafted bifaces Mean Min Max Frequency (out of 87) recorded s.d. c.v. Blade length (mm) 42.02 16.51 83.51 69 11.95 0.28 Shoulder width (mm) 28.16 22.92 34.75 45 2.7 0.1 Maximum thickness (mm) 7.05 5.21 9.09 87 0.86 0.12 Shoulder length (mm) 10.48 6.31 16.8 87 2.02 0.19 Between notch width (mm) 15.44 11.74 22.68 87 2.02 0.13 Notch height (mm) 5.87 3.6 10.29 87 1.22 0.21 Base width (mm) 24.15 17.51 29.9 81 2.62 0.11

PAGE 326

326 Table B 24. Jefferson County hafted biface metric attributes Jefferson County side notched hafted bifaces Mean Min Max Frequency (out of 14) recorded s.d. c.v. Blade length (mm) 38.31 28.39 55.07 9 9.14 0.24 Shoulder width (mm) 25.19 21.25 34.23 8 4.67 0.19 Maximum thickness (mm) 7.09 5.57 8.6 14 0.77 0.11 Shoulder length (mm) 11.54 9.19 16.61 14 2.02 0.18 Between notch width (mm) 15.37 9.87 20.78 14 3.02 0.2 Notch height (mm) 6.12 4.14 8.87 14 1.46 0.24 Base width (mm) 23.25 13.49 30.8 13 4.47 0.19 Table B 25. Lake Bird s ite hafted biface metric attributes Lake Bird site side notched hafted bifaces Mean Min Max Frequency (out of 74) recorded s.d. c.v. Blade length (mm) 38.83 22.55 69.4 50 10.77 0.28 Shoulder width (mm) 26.78 16.81 34.03 48 3.63 0.14 Maximum thickness (mm) 7.04 5.04 9.23 64 0.9 0.13 Shoulder length (mm) 11.31 7.2 15.96 67 1.78 0.16 Between notch width (mm) 14.84 11.21 20.22 74 2.03 0.14 Notch height (mm) 5.92 3.62 10.27 66 1.27 0.21 Base width (mm) 24.34 19.02 33.95 63 3.06 0.13

PAGE 327

327 Table B 26. Jeanies Better Back site hafted biface metric attributes Jeanie's Better Back site side notched hafted bifaces Mean Min Max Frequency (out of 21) recorded s.d. c.v. Blade length (mm) 33.46 20.68 61.51 18 9.84 0.29 Shoulder width (mm) 26.97 20.99 33.13 13 3.61 0.13 Maximum thickness (mm) 7.15 5.85 9 21 0.9 0.13 Shoulder length (mm) 10.65 7.42 14.53 21 1.95 0.18 Between notch width (mm) 14.97 12.38 20.63 21 2.07 0.14 Notch height (mm) 5.84 3.75 8.38 20 1.33 0.23 Base width (mm) 23.85 20.26 28.58 18 2.55 0.11 Table B 27. Hendrix Collection hafted biface metric attributes Hendrix Collection side notched hafted bifaces Mean Min Max Frequency (out of 268) recorded s.d. c.v. Blade length (mm) 41.27 18.99 113.52 238 12.72 0.31 Shoulder width (mm) 25.74 16.91 41.07 213 4.15 0.16 Maximum thickness (mm) 7.61 5.31 11.19 268 1.08 0.14 Shoulder length (mm) 13.43 7.89 22.24 268 2.64 0.2 Between notch width (mm) 16.24 10.05 25.2 268 2.73 0.17 Notch height (mm) 7.84 4.16 15.41 267 2.15 0.27 Base width (mm) 24.32 16.11 36.32 254 3.26 0.13

PAGE 328

328 Tampa Bay Sites and Collection Areas Table B 28. Harney Flats site hafted biface metric attributes Table B 29. Harris Grove Collection hafted biface metric attributes Harney Flats site side notched hafted bifaces Mean Min Max Frequency (out of 5) recorded s.d. c.v. Blade length (mm) 35.42 22.86 61.64 5 15.2 0.43 Shoulder width (mm) 24.81 20.79 28.28 5 2.77 0.11 Maximum thickness (mm) 7.72 6.34 9.12 5 1.11 0.14 Shoulder length (mm) 11.56 9.4 13.15 5 1.77 0.15 Between notch width (mm) 15.58 12.38 20.51 5 3.27 0.21 Notch height (mm) 6.8 5.42 8.54 5 1.15 0.17 Base width (mm) 22.34 18.84 24.55 4 2.61 0.12 Harris Grove Collection side notched hafted bifaces Mean Min Max Frequency (out of 44) recorded s.d. c.v. Blade length (mm) 40.82 12.94 67.27 39 12.63 0.31 Shoulder width (mm) 25.03 17.15 33.33 42 3.69 0.15 Maximum thickness (mm) 7.43 5.53 9.85 42 1.04 0.14 Shoulder length (mm) 11.72 8.21 16.05 43 2.12 0.18 Between notch width (mm) 15.34 9.72 19.97 44 2.51 0.16 Notch height (mm) 6.7 3.6 10.98 43 1.62 0.24 Base width (mm) 23.86 16.03 31.73 39 3.83 0.16

PAGE 329

329 Table B 30. Kepler site hafted biface metric attributes Kepler site side notched hafted bifaces Mean Min Max Frequency (out of 2) recorded s.d. c.v. Blade length (mm) 47.75 42.69 52.81 2 7.16 0.15 Shoulder width (mm) 32.63 31.93 33.32 2 0.98 0.03 Maximum thickness (mm) 6.59 5.79 7.38 2 1.12 0.17 Shoulder length (mm) 13.46 13.35 13.57 2 0.16 0.01 Between notch width (mm) 17.21 16.69 17.72 2 0.73 0.04 Notch height (mm) 7.09 6.29 7.89 2 1.13 0.16 Base width (mm) 27.3 25.47 29.13 2 2.59 0.09 Table B 31. Pinellas County hafted biface metric attributes Pinellas County side notched hafted bifaces Mean Min Max Frequency (out of 14) recorded s.d. c.v. Blade length (mm) 40.61 31.4 54.71 14 8.11 0.2 Shoulder width (mm) 24.67 21.51 34.89 13 3.7 0.15 Maximum thickness (mm) 7.46 6.12 8.72 13 0.73 0.1 Shoulder length (mm) 12.8 8.69 19.72 14 2.93 0.23 Between notch width (mm) 15.29 11.21 22.64 14 3.24 0.07 Notch height (mm) 6.77 4.45 8.36 14 1.32 0.19 Base width (mm) 22.99 17.6 32.21 14 3.94 0.17

PAGE 330

330 Table B 32. St. Petersburg Area hafted biface metric attributes St. Petersburg Area side notched hafted bifaces Mean Min Max Frequency (out of 3) recorded s.d. c.v. Blade length (mm) 0 Shoulder width (mm) 24.94 24.03 25.85 3 1.29 0.05 Maximum thickness (mm) 7.7 5.38 9.99 3 2.31 0.3 Shoulder length (mm) 15.31 11 22.2 3 6.03 0.39 Between notch width (mm) 17.81 12.09 25.45 3 6.88 0.39 Notch height (mm) 8.56 6.12 13.41 3 4.2 0.49 Base width (mm) 25.44 18.43 32.94 3 7.27 0.29 Table B 33. Spring Hill site hafted biface metric attributes Spring Hill site side notched hafted bifaces Mean Min Max Frequency (out of 7) recorded s.d. c.v. Blade length (mm) 34 21.84 49.26 6 10.97 0.32 Shoulder width (mm) 27.07 18.43 33.43 6 5.1 0.19 Maximum thickness (mm) 7.82 6 9.65 6 1.31 0.17 Shoulder length (mm) 11.34 8.96 14.04 6 2.02 0.18 Between notch width (mm) 17.47 12.39 20.36 7 2.9 0.17 Notch height (mm) 6.45 4.79 7.36 6 0.95 0.15 Base width (mm) 25.07 21.31 31.7 7 3.43 0.14

PAGE 331

331 Table B 34. Pasco County hafted biface metric attributes Pasco County side notched hafted bifaces Mean Min Max Frequency (out of 2) recorded s.d. c.v. Blade length (mm) 33 28.29 37.71 2 6.66 0.2 Shoulder width (mm) 25.26 23.37 27.15 2 2.67 0.11 Maximum thickness (mm) 6.7 5.96 7.43 2 1.04 0.16 Shoulder length (mm) 12.27 12.03 12.5 2 0.33 0.03 Between notch width (mm) 17.64 14.64 20.64 2 4.24 0.24 Notch height (mm) 7.41 6.55 8.26 2 1.21 0.16 Base width (mm) 24.52 22.72 26.31 2 2.54 0.1

PAGE 332

332 APPENDIX C METRIC ATTRIBUTES OF HAFTED BIFACES BY DRAINAGE SYSTEM Table C 35. Santee Cooper drainage side notched hafted bifaces Santee Cooper drainage side notched hafted bifaces Mean Min Max Frequency (out of 66) recorded s.d. c.v. s.e. Skewness Kurtosis Blade length (mm) 30.77 19.49 61.8 60 8.95 0.29 1.15 1.25 1.68 Shoulder width (mm) 21.63 15.84 31.31 58 3.58 0.17 0.47 0.8 0.1 Maximum thickness (mm) 6.5 4.16 9.65 66 1.04 0.16 0.13 0.22 0.34 Shoulder length (mm) 10.14 7.2 15.67 65 1.79 0.18 0.22 0.94 1.25 Between notch width (mm) 15.73 11.11 22.23 66 2.32 0.15 0.29 0.36 0.15 Notch height (mm) 4.88 3.05 8.11 65 0.92 0.19 0.11 0.6 1.3 Base width (mm) 22.16 17.25 30.13 52 3.1 0.14 0.43 0.44 0.35

PAGE 333

333 Table C 36. Savannah Ogeechee drainage side notched hafted bifaces Savannah Ogeechee drainage side notched hafted bifaces Mean Min Max Frequency (out of 600) recorded s.d. c.v. s.e. Skewness Kurtosis Blade length (mm) 33.03 12.2 69.01 378 10.04 0.3 0.52 0.6 0.58 Shoulder width (mm) 22.05 17.46 36.68 399 3.55 0.16 0.18 0.87 1.85 Maximum thickness (mm) 6.54 3.89 10.95 570 0.97 0.15 0.04 0.42 0.37 Shoulder length (mm) 10.33 1.53 26.15 586 1.93 0.19 0.08 1.48 9.25 Between notch width (mm) 15.4 10.22 25.32 581 2.47 0.16 0.1 0.73 0.85 Notch height (mm) 5.36 2.7 13.67 577 1.28 0.24 0.05 1.34 4.14 Base width (mm) 22.2 10.57 39.68 542 3.48 0.16 0.15 0.56 2.29

PAGE 334

334 Table C 37. Ocmulgee drainage side notched hafted bifaces Ocmulgee drainage side notched hafted bifaces Mean Min Max Frequency (out of 74) recorded s.d. c.v. s.e. Skewness Kurtosis Blade length (mm) 27.9 12.64 48.74 57 9.5 0.34 1.26 0.36 0.63 Shoulder width (mm) 22.09 14.87 36.52 56 4.87 0.22 0.65 0.83 0.43 Maximum thickness (mm) 6.75 3.43 9.75 72 0.98 0.14 0.11 0.03 2 Shoulder length (mm) 10.82 6.8 15.87 70 1.89 0.17 0.23 0.51 0.11 Between notch width (mm) 15.66 10.65 24.34 69 3.3 0.21 0.4 0.84 0.09 Notch height (mm) 5.62 3.34 8.92 71 1.25 0.22 0.15 0.59 0.08 Base width (mm) 23.3 16.05 32.62 52 3.97 0.17 0.55 0.23 0.56

PAGE 335

335 Table C 38. Flint drainage side notched hafted bifaces Flint drainage side notched hafted bifaces Mean Min Max Frequency (out of 97) recorded s.d. c.v. s.e. Skewness Kurtosis Blade length (mm) 31.27 15.65 59.14 69 8.18 0.26 0.98 0.57 0.79 Shoulder width (mm) 21.28 16.42 32.15 34 4.18 0.2 0.72 0.97 0.2 Maximum thickness (mm) 6.59 4.84 9.17 97 0.94 0.14 0.1 0.54 0.01 Shoulder length (mm) 10.38 6.9 17.14 96 1.84 0.18 0.19 0.7 1.37 Between notch width (mm) 14.84 10.2 19.24 97 1.71 0.12 0.17 0.18 0.47 Notch height (mm) 5.41 3.2 8.93 93 1.15 0.21 0.12 0.92 1.23 Base width (mm) 23.25 16.89 30.89 73 2.99 0.13 0.35 0.53 0.02

PAGE 336

336 Table C 39. Chattahoochee drainage side notched hafted bifaces Chattahoochee drainage side notched hafted bifaces Mean Min Max Frequency (out of 384) recorded s.d. c.v. s.e. Skew. Kurt. Blade length (mm) 38.54 17.54 81.69 274 11.24 0.29 0.68 0.87 1.19 Shoulder width (mm) 26.08 15.92 44.87 207 4.69 0.18 0.33 0.77 1.4 Maximum thickness (mm) 7.22 3.57 10.54 319 0.95 0.13 0.05 0.21 0.86 Shoulder length (mm) 10.43 6.14 16.76 318 1.66 0.16 0.09 0.23 0.46 Between notch width (mm) 15.62 11.27 26.86 320 2.51 0.16 0.14 1.32 2.61 Notch height (mm) 5.7 2.99 9.07 317 1.1 0.19 0.06 0.31 0.16 Base width (mm) 23.72 17.86 33.88 276 2.84 0.12 0.17 0.66 1

PAGE 337

337 Table C 40. Aucilla Suwannee drainage sidenotched hafted bifaces Aucilla Suwannee drainage side notched hafted bifaces Mean Min Max Frequency (out of 450) recorded s.d. c.v. s.e. Skewness Kurtosis Blade length (mm) 40.71 16.51 113.52 375 12.31 0.3 0.64 1.28 4.72 Shoulder width (mm) 26.29 16.81 41.07 319 3.96 0.15 0.22 0.24 0.41 Maximum thickness (mm) 7.4 5.04 11.19 440 1.04 0.14 0.05 0.58 0.63 Shoulder length (mm) 12.4 6.31 22.24 443 2.71 0.22 0.13 0.7 0.69 Between notch width (mm) 15.8 10.05 25.2 450 2.53 0.16 0.12 1.01 1.64 Notch height (mm) 7.07 3.6 15.41 440 2.08 0.29 0.1 1.05 1.31 Base width (mm) 24.27 16.11 36.32 416 3.08 0.13 0.15 0.49 0.98

PAGE 338

338 Table C 41. Tampa Bay sidenotched hafted bifaces Tampa Bay side notched hafted bifaces Mean Min Max Frequency (out of 77) recorded s.d. c.v. s.e. Skewness Kurtosis Blade length (mm) 42.84 26.35 67.27 51 10.41 0.24 1.46 0.81 0 Shoulder width (mm) 25.2 17.15 34.89 60 4.01 0.16 0.52 0.26 0.01 Maximum thickness (mm) 7.42 5.38 9.99 74 1.08 0.15 0.13 0.16 0.47 Shoulder length (mm) 12.09 8.21 22.2 75 2.49 0.21 0.29 1.29 3.26 Between notch width (mm) 15.75 9.72 25.45 77 2.97 0.19 0.34 0.52 0.43 Notch height (mm) 6.8 3.6 13.41 75 1.61 0.24 0.19 1.08 3.3 Base width (mm) 23.75 16.03 32.94 67 3.67 0.15 0.45 0.4 0.05

PAGE 339

339 APPENDIX D BASAL CONFIGURATION FOR HAFTED UNIFACES Table D 42. Basal configuration for Santee Cooper drainage Santee Cooper drainage Skinner Total Incurvate 0 0 Straight 2 2 Excurvate 2 2 Total 4 4 Table D 2. Basal configuration for SavannahOgeechee drainage Savannah Ogeechee drainage Big Pine Tree Site Strong Owl and Cotton Creek Sites Total Incurvate 1 41 8 50 Straight 2 59 24 85 Excurvate 1 36 3 40 Total 4 136 35 175 Table D 3. Basal configuration for Ocmulgee drainage Ocmulgee drainage Feronia sites Total Incurvate 1 1 Straight 0 0 Excurvate 8 8 Total 9 9 Table D 4. Basal configuration for Chattahoochee drainage Chattahoochee drainage Barbour County Total Incurvate 0 0 Straight 0 0 Excurvate 1 0 Total 1 1

PAGE 340

340 Table D 5. Basal configuration for Aucilla Suwannee drainage Aucilla Suwannee drainage Hendrix Total Incurvate 2 2 Straight 15 15 Excurvate 22 22 Total 39 39 Table D 6. Basal configuration for Tampa Bay drainage Tampa Bay drainage Harris Grove Kepler Site St. Petersburg Area Spring Hill Site Pasco County Total Incurvate 0 1 1 0 0 2 Straight 3 0 2 0 1 6 Excurvate 4 0 0 1 0 5 Total 7 1 3 1 1 13

PAGE 341

341 APPENDIX E METRIC ATTRIBUTES OF HAFTED UNIFACES BY SITE AND COLLECTION AREA Santee Cooper Sites and Collection Areas Table E 43. Skinner Collection hafted uniface metric attributes Skinner Collection side notched hafted unifaces Mean Min Max Frequency (out of 4) recorded s.d. c.v. s.e. Skewness Kurtosis Left blade length (mm) 38.43 19.74 52.58 4 13.66 0.36 6.83 0.96 2.02 Right blade length (mm) 26.98 12.96 42.23 4 11.98 0.44 5.99 0.3 1.5 Shoulder width (mm) 32.99 23.42 42.09 4 7.79 0.24 3.9 0.16 0.31 Left edge angle 52.5 45 55 4 5 0.1 2.5 2 4 Maximum thickness (mm) 10.6 7.09 12.96 4 2.66 0.25 1.33 0.87 0.82 Left shoulder length (mm) 13.78 8.69 17.66 4 7.77 0.56 4.48 0.33 0.87 Right shoulder length (mm) 16.82 11.52 24.96 3 4.78 0.28 2.76 0.59 0.67 Between notch width (mm) 24.14 15.82 28.89 4 5.76 0.24 2.88 1.56 2.66 Left notch height (mm) 9.29 6.62 12.76 4 3.11 0.33 1.55 0.25 4.56 Right notch height (mm) 8.12 3.93 13.83 4 4.19 0.52 2.1 1 1.45 Base width (mm) 31.62 20.64 38.83 4 7.81 0.25 3.91 1.28 2

PAGE 342

342 Savannah Ogeechee Sites and Collection Areas Table E 44. Big Pine Tree site hafted uniface metric attributes Big Pine Tree site side notched hafted unifaces Mean Min Max Frequency (out of 4) recorded s.d. c.v. s.e. Skewness Kurtosis Left blade length (mm) 43.26 32.22 58.37 4 13.03 0.3 6.52 0.35 4.01 Right blade length (mm) 32.38 25.14 37.86 4 6.03 0.19 3.02 0.45 3.22 Shoulder width (mm) 41.38 27.86 50.94 3 12.04 0.29 6.95 1.32 n/a Left edge angle 56.25 40 85 4 19.74 0.35 9.87 1.65 3.1 Maximum thickness (mm) 8.55 5.25 10.93 4 11.65 1.36 5.83 1.09 2.12 Left shoulder length (mm) 17.1 8.42 25 3 8.32 0.33 9.87 0.42 n/a Right shoulder length (mm) 18.9 11.5 27.8 4 6.83 0.36 3.41 0.6 0.77 Between notch width (mm) 25.35 18.98 34.28 4 6.68 0.26 3.34 0.92 0.16 Left notch height (mm) 8.87 4.52 12.1 3 3.91 0.44 2.26 1.19 n/a Right notch height (mm) 9.41 4.6 12.68 3 4.25 0.45 2.46 1.41 n/a Base width (mm) 34.18 22.44 52.59 4 13.39 0.39 6.69 1.16 0.77

PAGE 343

343 Table E 45. Strong Collection hafted uniface metric attributes Strong Collection side notched hafted unifaces Mean Min Max Frequency (out of 149) recorded s.d. c.v. s.e. Skewness Kurtosis Left blade length (mm) 36.69 16.46 63.6 106 8.27 0.23 0.8 0.67 0.84 Right blade length (mm) 26.63 9.14 60.64 103 9.76 0.37 0.96 0.92 1.73 Shoulder width (mm) 31.97 20.39 54.43 109 6.15 0.19 0.59 1.01 1.37 Left edge angle 51.55 20 80 129 14.08 0.27 1.24 0.1 0.59 Right edge angle 58.33 45 75 6 11.69 0.2 4.77 0.6 1.29 Maximum thickness (mm) 8.65 5.05 12.98 147 1.77 0.21 0.15 0.18 0.41 Left shoulder length (mm) 14.06 4.88 28.1 125 4.12 0.29 0.37 0.52 0.31 Right shoulder length (mm) 14.83 6.94 28.09 118 4.42 0.3 0.41 0.49 0.32 Between notch width (mm) 21.49 11.32 40.05 144 4.96 0.23 0.41 0.93 1.7 Left notch height (mm) 7.85 2.99 17.91 118 2.5 0.32 0.23 0.95 1.5 Right notch height (mm) 8.04 4.02 16.44 127 2.45 0.3 0.22 0.83 0.65 Base width (mm) 29.05 12.64 57.1 122 7.34 0.25 0.67 0.81 1.11

PAGE 344

344 Table E 46. Owl and Cotton Creek sites hafted uniface metric attributes Owl and Cotton Creek sites side notched hafted unifaces Mean Min Max Frequency (out of 35) recorded s.d. c.v. s.e. Skewness Kurtosis Left blade length (mm) 44.01 25.51 59.67 34 7.08 0.16 1.21 0.14 0.74 Right blade length (mm) 33.22 7.13 60.98 35 10.38 0.31 1.75 0.22 1.59 Shoulder width (mm) 37.08 28.26 50.35 32 4.59 0.12 0.81 0.54 1.01 Left edge angle 52.42 30 80 33 10.91 0.21 1.51 0.37 0.15 Right edge angle 70 1 Maximum thickness (mm) 9.34 5.41 13.2 35 1.73 0.19 0.29 0.03 0.17 Left shoulder length (mm) 17.33 9.64 45.51 33 6.65 0.38 1.16 2.65 9.83 Right shoulder length (mm) 18.13 8.84 39.76 33 5.64 0.31 0.98 1.61 5.78 Between notch width (mm) 22.77 14.2 31.03 34 3.75 0.16 0.64 0.28 0.12 Left notch height (mm) 8.8 4.47 12.76 30 2.41 0.27 0.44 0.17 1 Right notch height (mm) 9.13 4.19 15.53 33 2.66 0.29 0.46 0.56 0.16 Base width (mm) 32.34 18.27 43.79 34 5.88 0.18 1 0.8 0.34

PAGE 345

345 Ocmulgee Sites and Collection Areas Table E 47. Feronia sites hafted uniface metric attributes Feronia sites side notched hafted unifaces Mean Min Max Frequency (out of 9) recorded s.d. c.v. s.e. Skewness Kurtosis Left blade length (mm) 41.25 30.61 52.89 7 6.8 0.16 2.57 0.26 1.48 Right blade length (mm) 26.19 13.98 31.96 7 9.49 0.36 3.59 0.55 0.51 Shoulder width (mm) 34.82 26.82 40.83 8 5.03 0.14 1.78 0.88 0.46 Left edge angle 41.25 20 50 8 10.94 0.27 3.87 1.08 0.62 Maximum thickness (mm) 9.34 5.39 11.79 9 2.05 0.22 0.68 0.77 0.12 Left shoulder length (mm) 18.22 12.52 22.28 8 3.73 0.2 1.32 0.3 1.46 Right shoulder length (mm) 18.51 11.38 21.78 8 3.76 0.2 1.33 1.15 0.26 Between notch width (mm) 24.73 15.72 31.45 9 4.44 0.18 1.48 0.79 1.63 Left notch height (mm) 8.04 5.64 11.24 8 1.95 0.24 0.69 0.4 1 Right notch height (mm) 9.18 6.34 12.76 8 2.43 0.26 0.86 0.46 1.28 Base width (mm) 32.93 20.26 44.32 8 7.39 0.22 2.61 0.26 0.26

PAGE 346

346 Chattahoochee Sites and Collection Areas Table E 48. Barbour County hafted uniface metric attributes Barbour County side notched hafted unifaces Measurements (out of 1) recorded Left blade length (mm) 59.56 Right blade length (mm) 27.49 Shoulder width (mm) 53.71 Left edge angle 30 Maximum thickness (mm) 11.17 Left shoulder length (mm) 17.8 Right shoulder length (mm) 21.58 Between notch width (mm) 25.96 Left notch height (mm) 8.17 Right notch height (mm) 10.61 Base width (mm) 40.44

PAGE 347

347 Aucilla Suwannee Sites and Collection Areas Table E 49. Hendrix Collection hafted uniface metric attributes Hendrix Collection side notched hafted unifaces Mean Min Max Frequency (out of 39) recorded s.d. c.v. s.e. Skewness Kurtosis Left blade length (mm) 47.18 35.28 59.11 38 6.5 0.14 1.05 0.32 0.44 Right blade length (mm) 43 29.38 60.02 37 7.4 0.17 1.22 0.3 0.18 Shoulder width (mm) 36.51 26.55 45.01 32 4.41 0.12 0.79 0.38 0.24 Left edge angle 62.56 40 85 39 9.45 0.15 1.51 0.3 0.54 Maximum thickness (mm) 9.31 5.59 11.79 39 1.41 0.15 0.23 0.59 0.73 Left shoulder length (mm) 16.77 12.69 22.51 31 2.51 0.15 0.45 0.74 0.1 Right shoulder length (mm) 17.51 12.37 24.35 36 2.84 0.16 0.47 0.49 0.05 Between notch width (mm) 20.39 13.56 30.38 38 3.3 0.16 0.53 0.68 1.05 Left notch height (mm) 9.49 6.29 17.08 34 2.13 0.22 0.36 1.44 3.77 Right notch height (mm) 9.67 5.87 16.51 38 2.33 0.24 0.38 0.72 0.72 Base width (mm) 29.81 22.08 38.25 32 4.43 0.15 0.78 0.01 0.94

PAGE 348

348 Tampa Bay Sites and Collection Areas Table E 50. Harris Grove Collection hafted uniface metric attributes Harris Grove Collection side notched hafted unifaces Mean Min Max Frequency (out of 7) recorded s.d. c.v. s.e. Skewness Kurtosis Left blade length (mm) 47.73 39.05 54.21 7 7.71 0.16 2.91 0.05 2.1 Right blade length (mm) 40.17 25.61 56.79 7 11.26 0.28 4.25 0.08 1.16 Shoulder width (mm) 38.81 32.71 43.25 6 4.67 0.12 1.91 0.63 1.94 Left edge angle 59.29 40 75 7 14.27 0.24 5.38 0.1 1.84 Maximum thickness (mm) 9.74 7.12 13.7 7 2.21 0.23 0.83 0.87 0.62 Left shoulder length (mm) 16.66 12.21 21.42 7 2.96 0.18 1.12 0.09 0.35 Right shoulder length (mm) 17.37 14.12 22.03 7 2.84 0.16 1.07 0.39 0.48 Between notch width (mm) 22.59 14.79 28.69 7 5.26 0.23 1.98 0.46 1.59 Left notch height (mm) 8.84 8 10.59 7 1.07 0.12 0.4 1.17 0.59 Right notch height (mm) 8.84 7.55 10.74 7 1.13 0.13 0.43 0.81 0.36 Base width (mm) 32.25 23.22 37.95 7 5.99 0.19 2.26 0.26 1.45

PAGE 349

349 Table E 51. Kepler site hafted uniface metric attributes Kepler site side notched hafted unifaces Measurements (out of 1) recorded Left blade length (mm) Right blade length (mm) Shoulder width (mm) 28.84 Left edge angle 60 Maximum thickness (mm) 8.44 Left shoulder length (mm) 13.38 Right shoulder length (mm) 13.86 Bet ween notch width (mm) 20.42 Left notch height (mm) 8.36 Right notch height (mm) 7.57 Base width (mm) 29.52

PAGE 350

350 Table E 52. Pinellas County hafted uniface metric attributes Pinellas County side notched hafted unifaces Measurements (out of 1) recorded Left blade length (mm) Right blade length (mm) Shoulder width (mm) 46.95 Left edge angle 55 Maximum thickness (mm) 13.19 Left shoulder length (mm) Right shoulder length (mm) Between notch width (mm) 27.06 Left notch height (mm) Right notch height (mm) Base width (mm)

PAGE 351

351 Table E 53. St. Petersburg Area hafted uniface metric attributes St. Petersburg Area side notched hafted unifaces Minimum value Maximum value Frequency (out of 3 recorded) Left blade length (mm) 41.56 57.35 3 Right blade length (mm) 35.96 48.42 3 Shoulder width (mm) 29.06 44.31 2 Left edge angle 45 75 3 Maximum thickness (mm) 7.84 9.86 3 Left shoulder length (mm) 13.32 17.42 3 Right shoulder length (mm) 15 16.91 3 Between notch width (mm) 15 26.21 3 Left notch height (mm) 6.36 8.71 3 Right notch height (mm) 8.06 9.09 2 Base width (mm) 27.39 36.24 2

PAGE 352

352 Table E 54. Spring Hill site hafted uniface metric attributes Spring Hill site side notched hafted unifaces Measurements (out of 1) recorded Left blade length (mm) 51.64 Right blade length (mm) 36.92 Shoulder width (mm) 41.5 Left edge angle 45 Maximum thickness (mm) 10.17 Left shoulder length (mm) Right shoulder length (mm) 19.55 Between notch width (mm) 24.07 Left notch height (mm) Right notch height (mm) 8.58 Base width (mm)

PAGE 353

353 Table E 55. Pasco County hafted uniface metric attributes Pasco County side notched hafted unifaces Measurements (out of 1) recorded Left blade length (mm) 53.16 Right blade length (mm) 43.58 Shoulder width (mm) 45.83 Left edge angle 55 Maximum thickness (mm) 12.25 Left shoulder length (mm) 31.12 Right shoulder length (mm) 27.17 Between notch width (mm) 27.79 Left notch height (mm) 23.24 Right notch height (mm) 19.17 Base width (mm) 38.04

PAGE 354

354 APPENDIX F METRIC ATTRIBUTES OF HAFTED UNIFACES BY DRAINAGE SYSTEM Table F 56. SanteeCooper drainage side notched hafted unifaces Santee Cooper drainage side notched hafted unifaces Mean Min Max Frequency (out of 4 ) recorded s.d. c.v. s.e. Skewness Kurtosis Left blade length (mm) 38.43 19.74 52.58 4 13.66 0.36 6.83 0.96 2.02 Right blade length (mm) 26.98 12.96 42.23 4 11.98 0.44 5.99 0.3 1.5 Shoulder width (mm) 32.99 23.42 42.09 4 7.79 0.24 3.9 0.16 0.31 Left edge angle 52.5 45 55 4 5 0.1 2.5 2 4 Maximum thickness (mm) 10.6 7.09 12.96 4 2.66 0.25 1.33 0.87 0.82 Left shoulder length (mm) 13.78 8.69 17.66 4 7.77 0.56 4.48 0.33 0.87 Right shoulder length (mm) 16.82 11.52 24.96 3 4.78 0.28 2.76 0.59 0.67 Between notch width (mm) 24.14 15.82 28.89 4 5.76 0.24 2.88 1.56 2.66 Left notch height (mm) 9.29 6.62 12.76 4 3.11 0.33 1.55 0.25 4.56 Right notch height (mm) 8.12 3.93 13.83 4 4.19 0.52 2.1 1 1.45 Base width (mm) 31.62 20.64 38.83 4 7.81 0.25 3.91 1.28 2

PAGE 355

355 Table F 57. Savannah Ogeechee drainage side notched hafted unifaces Savannah Ogeechee drainage side notched hafted unifaces Mean Min Max Frequency (out of 188) recorded s.d. c.v. s.e. Skewness Kurtosis Left blade length (mm) 38.6 16.46 63.6 144 8.69 0.23 0.72 0.4 0.02 Right blade length (mm) 28.42 7.13 60.98 142 10.21 0.36 0.86 0.67 1.09 Shoulder width (mm) 33.13 20.39 54.43 141 6.2 0.19 0.52 0.66 0.52 Left edge angle 51.84 20 85 166 13.59 0.26 1.05 0 0.37 Right edge angle 70 45 75 7 11.55 0.16 4.71 0.11 1.92 Maximum thickness (mm) 8.78 5.05 13.2 186 1.79 0.2 0.13 0.11 0.36 Left shoulder length (mm) 14.79 4.88 45.51 161 4.97 0.34 0.39 1.91 8.93 Right shoulder length (mm) 15.64 6.94 39.76 155 4.95 0.32 0.4 0.97 2.6 Between notch width (mm) 21.81 11.32 40.05 182 4.82 0.22 0.36 0.77 1.38 Left notch height (mm) 8.05 2.99 17.91 148 2.5 0.31 0.21 0.71 0.73 Right notch height (mm) 8.29 4.02 16.44 163 2.55 0.31 0.2 0.72 0.3 Base width (mm) 29.88 14.64 57.1 160 7.36 0.25 0.58 0.6 0.73

PAGE 356

356 Table F 58. Ocmulgee drainage side notched hafted unifaces Ocmulgee drainage side notched hafted unifaces Mean Min Max Frequency (out of 9 ) recorded s.d. c.v. s.e. Skewness Kurtosis Left blade length (mm) 41.25 30.61 52.89 7 6.8 0.16 2.57 0.26 1.48 Right blade length (mm) 26.19 13.98 31.96 7 9.49 0.36 3.59 0.55 0.51 Shoulder width (mm) 34.82 26.82 40.83 8 5.03 0.14 1.78 0.88 0.46 Left edge angle 41.25 20 50 8 10.94 0.27 3.87 1.08 0.62 Maximum thickness (mm) 9.34 5.39 11.79 9 2.05 0.22 0.68 0.77 0.12 Left shoulder length (mm) 18.22 12.52 22.28 8 3.73 0.2 1.32 0.3 1.46 Right shoulder length (mm) 18.51 11.38 21.78 8 3.76 0.2 1.33 1.15 0.26 Between notch width (mm) 24.73 15.72 31.45 9 4.44 0.18 1.48 0.79 1.63 Left notch height (mm) 8.04 5.64 11.24 8 1.95 0.24 0.69 0.4 1 Right notch height (mm) 9.18 6.34 12.76 8 2.43 0.26 0.86 0.46 1.28 Base width (mm) 32.93 20.26 44.32 8 7.39 0.22 2.61 0.26 0.26

PAGE 357

357 Table F 59. Aucilla Suwannee drainage sidenotched hafted unifaces Aucilla Suwannee drainage side notched hafted unifaces Mean Min Max Frequency (out of 39) recorded s.d. c.v. s.e. Skewness Kurtosis Left blade length (mm) 47.18 35.28 59.11 38 6.5 0.14 1.05 0.32 0.44 Right blade length (mm) 43 29.38 60.02 37 7.4 0.17 1.22 0.3 0.18 Shoulder width (mm) 36.51 26.55 45.01 32 4.41 0.12 0.79 0.38 0.24 Left edge angle 62.56 40 85 39 9.45 0.15 1.51 0.3 0.54 Maximum thickness (mm) 9.31 5.59 11.79 39 1.41 0.15 0.23 0.59 0.73 Left shoulder length (mm) 16.77 12.69 22.51 31 2.51 0.15 0.45 0.74 0.1 Right shoulder length (mm) 17.51 12.37 24.35 36 2.84 0.16 0.47 0.49 0.05 Between notch width (mm) 20.39 13.56 30.38 38 3.3 0.16 0.53 0.68 1.05 Left notch height (mm) 9.49 6.29 17.08 34 2.13 0.22 0.36 1.44 3.77 Right notch height (mm) 9.67 5.87 16.51 38 2.33 0.24 0.38 0.72 0.72 Base width (mm) 29.81 22.08 38.25 32 4.43 0.15 0.78 0.01 0.94

PAGE 358

358 Table F 60. Tampa Bay drainage side notched hafted unifaces Tampa Bay drainage side notched hafted unifaces Mean Min Max Frequency (out of 14) recorded s.d. c.v. s.e. Skewness Kurtosis Left blade length (mm) 49.27 39.05 58 12 6.97 0.14 2.01 0.43 1.52 Right blade length (mm) 40.22 25.61 56.79 12 8.95 0.22 2.58 0.12 0.32 Shoulder width (mm) 39.11 28.84 46.95 12 6.43 0.16 1.86 0.6 1.17 Left edge angle 57.86 40 75 14 12.04 0.21 3.22 0.24 1.14 Maximum thickness (mm) 9.89 7.12 13.7 14 2.04 0.21 0.55 0.66 0.57 Left shoulder length (mm) 17.2 12.21 31.12 12 5.09 0.3 1.47 2.06 5.12 Right shoulder length (mm) 17.71 13.86 27.17 12 3.93 0.22 1.13 1.35 1.91 Between notch width (mm) 22.61 14.79 28.69 14 4.91 0.22 1.31 0.44 1.43 Left notch height (mm) 9.65 6.36 23.24 12 4.42 0.46 1.28 3.08 10.08 Right notch height (mm) 9.53 7.55 19.17 12 3.18 0.33 0.92 2.97 9.42 Base width (mm) 32.45 23.22 39 11 5.43 0.17 1.64 0.23 1.42

PAGE 359

359 REFERENCES CITED Adams, William Y., and Ernest W. Adams 1991 Archaeological Typology and Practical Reality: A Dialectical Approach to Artifact Classification and Sorting. Cambridge University Press, Cambridge. Adovasio, James M., and C. Andrew Hemmings 2011 Inundated Landscapes and the Colonization of the Northeastern Gulf of Mexico. Paper presented at the 76th Annual Meeting of the Society for American Archaeology, Sacramento, California. Ahler, Stanley A. 1971 Projectile Point Form and Function at Rodgers Shelter, Missouri. Missouri Archaeological Society Research Series 8. Ames, Kenne th M. 2004 Supposing Hunter Gatherer Variability. American Antiquity 69(2):364374. Anderson, David G. 1990 The Paleoindian Colonization of Eastern North America: A View from the Southeastern United States. In Early Paleoindian Economies of Eastern North America, edited by Kenneth B. Tankersley and Barry L. Isaac, pp. 163216. Research in Economic Anthropology, Supplement 5. JAI Press, Greenwich, Connecticut. 1995 Paleoindian Interaction Networks in the Eastern Woodlands. In Native American Interaction: Multiscalar Analyses and Interpretations in the Eastern Woodlands edited by Michael S. Nassaney and Kenneth E. Sassaman, pp. 1 26. University of Tennessee Press, Kn oxville. 1996 Models of Paleoindian and Early Archaic Settlement in the Lower Southeast. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 2957. University of Alabama Press, Tuscaloosa. 2003 Archaeo logy and Anthropology in the Twenty First Century: Strategies for Working Together. In Archaeology is Anthropology, edited by Susan D. Gillespie and Deborah L. Nichols pp. 111127. Archeological Papers of the American Anthropological Association 13. Ander son, David G., and J. Christopher Gillam 2001 Paleoindian Interaction and Mating Networks: Reply to Moore and Moseley. American Antiquity 66(3):530535. Anderson, David G., and Glen T. Hanson 1988 Early Archaic Settlement in the Southeastern United States : A Case from the Savannah River. American Antiquity 53(2):262286.

PAGE 360

360 Anderson, David G., and Kenneth E. Sassaman 2012 Recent Developments in Southeastern Archaeology: From Colonization to Complexity. The SAA Press, The Society for American Archaeology, Washington, D.C. Anderson, David G., and Kenneth E. Sassaman (editors) 1996 Modeling Paleoindian and Early Archaic Settlement in the Southeast: A Historical Perspective. In The Paleoindian and Early Archaic Southeast, pp. 1628. University of Alabama Press, Tuscaloosa. Anderson, David G., Charles E. Cantley, and A. Lee Novick 1982 The Mattassee Lake Sites: Archaeological Investigations along the Lower Santee River in the Coastal Plain of South Carolina. Special Publicati on, Interagency Archaeological Services, National Park Service, Atlanta. Anderson, David G., Lisa D. OSteen, and Kenneth E. Sassaman 1996 Environmental and Chronological Considerations. In The Paleoindian and Early Archaic Southeast, edited by David G. A nderson and Kenneth E. Sassaman, pp. 315. University of Alabama Press, Tuscaloosa. Anderson, David G., D. Shane Miller, Steven J. Yerka, J. Christopher Gillam, Erik N. Johanson, Derek T. Anderson, Albert C. Goodyear, and Ashley M. Smallwood 2010 PIDBA (P aleoindian Database of the Americas) 2010: Current Status and Findings. Archaeology of Eastern North America 38:128. Andrefsky, William 1998 Lithics: Macroscopic Approaches to Analysis. Cambridge University Press, Cambridge. Apicella, Coren L., Frank W. Marlowe, James H. Fowler, and Nicholas A. Cristakis. 2012 Social Networks and Cooperation in Hunter gatherers. Nature 481(7382):497501. Austin, Robert J. 1997 The Economics of Lithic Resource Use in SouthCentral Florida. Ph.D. dissertation, Department of Anthropology, University of Florida, Gainesville. Austin, Robert J., and Scott E. Mitchell 1999 Archaeological Investigations at Jeanies Better Back (8LF54), an Early Archaic Site in Lafayette County, Florida. Prepared by Southeastern Archaeological Research, Inc., for Florida Department of Transportation, District Two, Lake City, Florida. 2010 Reconstructing the Life Histories of Bolen Hafted Bifaces from a North Florida Archaeological Site. Early Georgia 38(1):327. Bamforth, Douglas B. 1986 Tech nological Efficiency and Tool Curation. American Antiquity 51(1):3850.

PAGE 361

361 Bamforth, Douglas B., and Nyree Finlay 2008 Introduction: Archaeological Applications to Lithic Production Skill and Craft Learning. Journal of Archaeological Method and Theory 15(1): 127. Barnard, Alan 1978 Universal Systems of Kin Categorization. African Studies 37:6981. 1983 Contemporary Hunter Gatherers: Current Theoretical Issues in Ecology and Social Organization. Annual Review of Anthropology 12:193214. 2011 Social Anthropology and Human Origins. Cambridge University Press, Cambridge. Barrett, John C. 1994 Fragments from Antiquity: An Archaeology of Social Life in Britain, 29001200 BC Blackwell, Oxford. Barth, Frederick (editor) 1969 Ethnic Groups and Boundaries: The Social Organization of Culture Difference. Little, Brown and Company, Boston. Bartram, Lawrence R. 1997 A Comparison of Kua (Botswana) and Hadza (Tanzania) Bow and Arrow Hunting. In Projectile Technology edited by Heidi Knecht, pp. 321343. Plenum P ress, New York. Basso, Keith H. 1983 Stalking with Stories: Names, Places, and Moral Narratives among the Western Apache. In Text, Play, and Story: The Construction and Reconstruction of Self and Society, edited by Edward M. Bruner, pp. 1955. American Ethnological Society, Washington, D.C. Beck, Charlotte, and George T. Jones 2012 Clovis and Western Stemmed Again: Reply to Fiedel and Morrow. American Antiquity 77(2):386397. Bender, Barbara 1993 Landscape: Politics and Perspectives. Berg, Oxford 1999 Subverting the Western Gaze: Mapping Alternative Worlds. In The Archaeology and Anthropology of Landscape: Shaping Your Landscape, edited by Peter J. Ucko and Robert Layton, pp. 31 45. Routledge, London. 2002 Time and Landscape. Current Anthropology 43 Supplement:S103S112. Bernard, H. Russell 2002 Research Methods in Anthropology: Qualitative and Quantitative Approaches. 3rd ed. AltaMira Press, Walnut Creek, California.

PAGE 362

362 Bettinger, Robert L., and Jelmer Eerkens 1999 Point Typologies, Cultural Transmis sions, and the Spread of Bow and Arrow Technology in the Prehistoric Great Basin. American Antiquity 64(2):231242. Binford, Lewis R. 1979 Organization and Formation Processes: Looking at Curated Technologies Journal of Anthropological Research 35:255273. 1986 An Alyawara Day: Making Mens Knives and Beyond. American Antiquity 51(3):547562. 2001 Constructing Frames of Reference: An Analytical Method for Archaeological Theory Building Using Hunter Gatherer and Environmental Data Sets. Universi ty of California Press, Berkeley. Bissett, Thaddeus G. 2003 Morphological Variation of Bolen Haftable Bifaces: Function and Style among ChippedStone Artifacts from the Early Holocene Southeast. M.A. thesis, Department of Anthropology, Florida State Unive rsity, Tallahassee. Blanton, Dennis B., and Frankie Snow 1986 Paleo Indian and Early Archaic Lithic Assemblage Composition in South Georgia: Evidence from the Feronia Locality. Paper presented at the annual Southeastern Archaeological Conference, Nashvill e, Tennessee. 1989 Paleo Indian and Early Archaic Occupations at the Feronia Locality in SouthCentral Georgia. Paper presented at the 54th Annual Meeting of the Society for American Archaeology, Atlanta, Georgia. Bleed, Peter 1986 The Optimal Design of Hunting Weapons: Maintainability or Reliability. American Antiquity 51(4): 737747. Bradley, Bruce A. 1997 Sloan Site Biface and Projectile Point Technology. In Sloan: A Paleoindian Dalton Cemetery in Arkansas, edited by Dan F. Morse, pp. 5357. Smithson ian Institution Press, Washington, D.C. Bradley, Richard 2000 An Archaeology of Natural Places. Routledge, New York. Bridgman Sweeney, Kara, Albert C. Goodyear, Tommy Charles, and Erica Shofner 2008 Achieved and Received Wisdom in the Archaeology of Edgefield Scrapers. Paper presented at the 65th Annual Southeastern Archaeological Conference, Charlotte, North Carolina.

PAGE 363

363 Brockington, Paul B. 1971 Preliminary Investigation of an Early Knapping Site in Southeastern Georgia. South Carolina Institute of Ar chaeology and Anthropology, University of South Carolina, The Notebook 3:3446. Brooks, Mark J., Barbara E. Taylor, and Andrew H. Ivester 2010 Carolina Bays: Time Capsules of Culture and Climate Change. Southeastern Archaeology 29:146163. Bullen, Ripley P. 1958 The Bolen Bluff Site on Paynes Prairie, Florida. Contributions of the Florida State Museum Social Sciences, Number 4:138 and plates. 1975 A Guide to the Identification of Florida Projectile Points. Revised Edition. Kendall Books, Gainesville, F lorida. Bush, Peter W., and Richard H. Johnston 1988 GroundWater Hydraulics, Regional Flow and GroundWater Development of the Floridan Aquifer System in Florida and in Parts of Georgia, South Carolina, and Alabama. U.S. Geological Survey Professional Paper 1403C. United States Government Printing Office, Washington, D.C. Cable, John S. 1996 Haw River Revisited: Implications for Modeling Terminal Late Glacial and Early Holocene Hunter Gatherer Settlement Systems in the Southeast. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 107148. University of Alabama Press, Tuscaloosa. Cambron, James W., and David C. Hulse 1983 Handbook of Alabama Archaeology, Part I :Point Types edited by David L. DeJarnette. Alabama Archaeological Society, Huntsville. Carter, Brinnen S. 2003 Page/Ladson (8JE591): Excavation of an Early Holocene Occupation Site in the Aucilla River, Florida. Ph.D. dissertation, Department of Anthropology, University of Florida, Gainesville. Carter, Brinnen, and James S. Dunbar 2006 Holocene Archaeology of the PageLadson Site: The Early Archaic. In First Floridians and Last Mastadons: The Page Ladson Site on the Aucilla River, edited by S.D. Webb. S pringer Press, Dordrecht, The Netherlands. Chapman, Carl H. 1948 A Preliminary Survey of Missouri Archaeology, Part IV, Ancient Cultures and Sequence. The Missouri Archaeologist 10(IV).

PAGE 364

364 Charles, Tommy 1986 The Fifth Phase of the Collectors Survey. South Carolina Institute of Archaeology and Anthropology, University of South Carolina, The Notebook 18:127. Cheshier, Joseph, and Robert L. Kelly 2006 Projectile Point Shape and Durability: The Effect of Thickness: Length. American Antiquity 71(2):353363. C hristenson, Andrew L. 1997 Side Notched and Unnotched Arrowpoints: Assessing Functional Differences. In Projectile Technology edited by Heidi Knecht, pp. 131142. Plenum Press, New York. Clausen, C.J., A.D. Cohen, C. Emeliani, J.A. Holman, and J.J. Stipp 1979 Little Salt Spring, Florida: A Unique Underwater Site. Science 203:609614. Coe, Joffre L. 1964 The Formative Cultures of the Carolina Piedmont. Transactions of the American Philosophical Society 54(5). Philadelphia. Conkey, Margaret W. 1980 The Identification of Prehistoric Hunter Gatherer Aggregation Sites: The Case of Altamira. Current Anthropology 21(5):609630. Cooney, Gabriel 1999 Social Landscapes in Irish Prehistory. In The Archaeology and Anthropology of Landscape: Shaping Your Lands cape, edited by Peter J. Ucko and Robert Layton, pp. 46 64. Routledge, London. Copeland, C.E. 1968 Geology of the Alabama Coastal Plain. A Guidebook. Geological Survey of Alabama Circular 47. Coren, Stanley, and Clare Porac 1977 Fifty Centuries of Right Handedness: The Historical Record. Science 198(4317):631632. Cowgill, George L. 1982 Clusters of Objects and Associations Between Variables: Two Approaches to Archaeological Classification. In Essays on Archaeological Typology, edited by Robert Whallon and James A. Brown, pp. 3055. Center for American Archeology Press, Evanston, Illinois. Damas, David (editor) 1969 Environment, History, and Central Eskimo Society. In Contributions to Anthropology: Ecological Essays, pp. 4069. National Museums of Canada, Ottawa.

PAGE 365

365 Damas, David 1972 The Copper Eskimo. In Hunters and Gatherers Today edited by M.G. Bicchieri, pp. 350. Holt, Rinehart, and Winston, New York. Daniel, I. Randolph, Jr. 1996 Raw Material Availability and Early Archaic Settlement in the Southeast. In The Paleoindian and Early Archaic Southeast edited by David G. Anderson and Kenneth E. Sassaman, pp. 8491. University of Alabama Press, Tuscaloosa. 1998 Hardaway Revisited: Early Archaic Settlement in the Southeast. University of Alabama Press, Tuscaloosa. 2001 Stone Raw Material Availability and Early Archaic Settlement in the Southeastern United States. American Antiquity 66(2):237265. Daniel, I. Randolph, Jr., and Mic hael Wisenbaker 1987 Harney Flats: A Florida Paleoindian Site. Baywood, Farmingdale, New York. David, Nicholas, and Carol Kramer 2001 Ethnoarchaeology in Action. Cambridge University Press, Cambridge. Dawson, John L.M. 1974 Ecology, Cultural Pressures t owards Conformity, and Left Handedness: A Biosocial Approach. In Readings in Cross Cultural Psychology edited by John L.M. Dawson and Walter J. Lonner, pp. 125149. Hong Kong University Press, Hong Kong. Deetz, James 1968 The Inference of Residence and D escent from Archaeological Data. In New Perspectives in Archaeology, edited by Sally R. Binford and Lewis R. Binford, pp. 4148. Aldine Press, Chicago. DeJarnette, David L., Edward B. Kurjack, and James W. Cambron 1962 Stanfield Worley Bluff Shelter Excav ations. Journal of Alabama Archaeology 8(1 and 2):1111. Delcourt, Hazel R., and Delcourt, Paul A. 1985 Quaternary Palynology and Vegetational History of the Southeastern United States. In Pollen Records of Late Quaternary North American Sediments edited by V.M. Bryant and R.G. Holloway, pp.1 37. American Association of Stratigraphic Palynologists Foundation. Delcourt, Paul A., and Hazel R. Delcourt 1987 Long Term Forest Dynamics of the Temperate Zone: A Case Study of Late Quaternary Forests in Eas tern North America. Springer Verlag, New York.

PAGE 366

366 Detwiler, Kandace R. 2000 Gathering in the Late Paleoindian: Botanical Remains from Dust Cave, Alabama. Paper presented at the 65th Annual Meeting of the Society for American Archaeology, Philadelphia, Pennsylvania. Dincauze, Dena F. 1996 Modeling Communities and Other Thankless Tasks. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 421424. University of Alabama Press, Tuscaloosa. Dobres, MarciaAnn e, and John E. Robb 2000 Agency in Archaeology: Paradigm or Platitude? In Agency in Archaeology edited by MarciaAnne Dobres and John E. Robb, pp. 317. Routledge, London. Doran, Glen H., and David N. Dickel 1988 Multidisciplinary Investigations at the W indover Site. In Wet Site Archaeology edited by Barbara Purdy, pp. 263289. Telford Press, Caldwell, New Jersey. Drennan, Robert D. 1996 Statistics for Archaeologists: A Commonsense Approach. Plenum Press, New York. Driskell, Boyce N. 1996 Stratified Late Pleistocene and Early Holocene Deposits at Dust Cave, Northwestern Alabama. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 315330. University of Alabama Press, Tuscaloosa. Dunbar, James S. 1991 Resource Orientation of Clovis and Suwannee Age Paleoindian Sites in Florida. In Clovis: Origins and Adaptations edited by R. Bonnichsen and K. Turnmire, pp. 185213. Center for the Study of the First Americans, Oregon State University, Corv allis. 2002 Chronostratigraphy and Paleoclimate of Late Pleistocene Florida and the Implications of Changing Paleoindian Land Use. M.A. thesis, Department of Anthropology, Florida State University, Tallahassee. Dunbar, James S., and Ben I. Waller 1983 A Distribution of the Clovis/Suwannee Paleoindian Sites in Florida: A Geographic Approach. Florida Anthropologist 36:1830. Dunbar, James S., and S. David Webb 1996 Bone and Ivory Tools from Submerged Paleoindian Sites in Florida. In The Paleoindian and Ear ly Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 331353. University of Alabama Press, Tuscaloosa.

PAGE 367

367 Dunbar, James S., S. David Webb, and Dan Cring 1989 Culturally and Naturally Modified Bones from a Paleoindian Site in the A ucilla River, North Florida. In Bone Modification, edited by Rob Bonnichsen and M. Sorg, pp. 473497. Center for the Study of the First Americans, Orono, Maine. Dunnell, Robert C. 1986 Methodological Issues in Americanist Artifact Classification. In Advan ces in Archaeological Method and Theory Volume 9, edited by Michael B. Schiffer, pp. 149207. Academic Press, Inc., Orlando. Elliott, Daniel T. 2006 The Archaic Period of Georgia, South of the Fall Line. LAMAR Institute Publication Series 80. LAMAR Institute, Inc., Savannah, Georgia. Electronic document, http://shapiro.anthro.uga.edu/Lamar/images/PDFs/publication_80.pdf accessed February 1, 2013. Elliott, Dan iel T., and Roy Doyon 1981 Archaeological and Historical Geography of the Savannah River Floodplain Near Augusta, Georgia. Laboratory of Anthropology Series, Report 22. University of Georgia, Athens. Ember, Carol R. 1975 Residential Variation among Hunter Gatherers. Behavior Science Research 3:199227. Farr, Grayal E. 2006 A Reevaluation of Bullens Typology for Preceramic Projectile Points. M.A. thesis, Department of Anthropology, Florida State University, Tallahassee. Faught, Michael K. 1996 Clovis Or igins and Underwater Prehistoric Archaeology in Northwestern Florida. Ph.D. dissertation, Department of Anthropology, University of Arizona, Tucson. 2004a Submerged Paleoindian and Archaic Sites of the Big Bend, Florida. Journal of Field Archaeology 29:273289. 2004b The Underwater Archaeology of Paleolandscapes, Appalachee Bay, Florida. American Antiquity 69(2):275289. Faught, Michael, and Brinnen Carter 1998 Early Human Occupation and Environmental Change in Northwestern Florida. Quaternary International 4950:167176. Faught, Michael K., and James C. Waggoner, Jr. 2012 The Early Archaic to Middle Archaic Transition: An Argument for Discontinuity. Florida Ant hropologist 65(3):153175.

PAGE 368

368 Faught, Michael K., Michael B. Hornum, R. Christopher Goodwin, Brinnen Carter, and S. David Webb 2003 Earliest Holocene Tool Assemblages from Northern Florida with Stratigraphically Controlled Radiocarbon Estimates (Sites 8LE210 5 and 8JE591). Current Research in the Pleistocene 20:1618. Faurie, Charlotte, Wulf Schiefenhovel, Sylvie Le Bomin, Sylvain Billiard, and Michel Raymond 2005 Variation in the Frequency of Left handedness in Traditional Societies. Current Anthropology 46(1):142147. Fiedel, Stuart J. 1999 Older Than We Thought: Implications of Corrected Dates for Paleoindians. American Antiquity 64(1):95116. Fiedel, Stuart J., and Juliet E. Morrow 2012 Comment on Clovis and Western Stemmed: Population Migration and the Meeting of Two Technologies in the Intermountain West by Charlotte Beck and George T. Jones. American Antiquity 77(2):376385. Firestone, R.B. A. West, J.P. Kennett, L. Becker, T.E. Bunch, Z.S. Revay, P.H. Schultz, T. Belgya, D.J. Kennett, J.M. Erla ndson, O.J. Dickenson, A.C. Goodyear, R.S. Harris, G.A. Howard, J.B. Kloosterman, P. Lechler, P.A. Mayewski, J. Montgomery, R. Poreda, T. Darrah, S.S. Que Hee, A.R. Smith, A. Stich, W. Topping, J.H. Wittke, and W.S. Wolbach 2007 Evidence for an Extraterrestrial Impact 12,900 Years Ago that Contributed to the Megafaunal Extinctions and the Younger Dryas Cooling. Proceedings of the National Academy of Sciences, USA 104:1601616021. Fitzhugh, Ben, S. Colby Phillips, and Erik Gjesfje ld 2011 Modeling Variability in Hunter Gatherer Information Networks: An Archaeological Case Study from the Kuril Islands, in Information and its Role in Hunter Gatherer Bands edited by Robert W. Whallon, William A. Lovis, and Robert K. Hitchcock, pp. 85115. UCLA Cotsen Institute of Archaeology, Los Angeles. Flenniken, J. Jeffrey, and Anan W. Raymond 1986 Morphological Projectile Point Typology: Replication Experimentation and Technological Analysis. American Antiquity 51(3):603614. Flenniken, J. Jeffr ey, and Philip J. Wilke 1989 Typology, Technology, and Chronology of Great Basin Dart Points. American Anthropologist 91(1): 149158. Ford, James A. 1952 Measurements of Some Prehistoric Design Developments in the Southeastern States. American Museum of Natural History, Anthropological Papers 44(3):313384.

PAGE 369

369 Fullagar, Richard, and Lesley Head 1999 Exploring the Prehistory of Hunter gatherer Attachments to Place: An Example from the Keep River Area, Northern Territory, Australia. In The Archaeology and A nthropology of Landscape: Shaping Your Landscape, edited by Peter J. Ucko and Robert Layton, pp. 322335. Routledge, London. Futato, Eugene M. 1996 A Synopsis of Paleoindian and Early Archaic Research in Alabama. In Paleoindian and Early Archaic Southeast edited by David G. Anderson and Kenneth E. Sassaman, pp. 298314. University of Alabama Press, Tuscaloosa. Gardner, William M. 1974 The Flint Run Paleoindian Complex: A Preliminary Report 1971 through 1973 Seasons. Occasional Paper No. 1. Archaeological Laboratory, Catholic University of America, Washington, D.C. Gastwirth, Joseph L., Yulia R. Gel, and Weiwen Miao 2009 The Impact of Levenes Test of Equality of Variances on Statistical Theory and Practice. Statistical Science 24(3):343360. Gerrell, Ph il R., J.F. Scarry, and J.S. Dunbar 1991 Analysis of Early Archaic Unifacial Adzes from North Florida. The Florida Anthropologist 44(1):316. Goad, Sharon I. 1979 Chert Resources in Georgia: Archaeological and Geological Perspectives. Wallace Reservoir Pr oject Contribution 3. University of Georgia Laboratory of Archaeology Series Report Number 21, Athens. Goodyear, Albert C., III 1975 An Archeological Survey of the Proposed Alternate Three Route, Southern Alternate, of the Southwestern Columbia Beltway Be tween I 26 and S.C. 48. Research Manuscript Series 77. South Carolina Institute of Archaeology and Anthropology, University of South Carolina, Columbia. 1979 A Hypothesis for the Use of Cryptocrystalline Raw Materials among Paleoindian Groups of North America. Research Manuscript Series 156. South Carolina Institute of Archaeology and Anthropology, University of South Carolina, Columbia. 1982 The Chronological Position of the Dalton Horizon in the Southeastern United States. American Antiquity 47(2):382395. 1995 [1974] The Brand Site: A TechnoFunctional Study of a Dalton Site in Northeast Arkansas. Arkansas Archaeological Survey Research Series No. 7.

PAGE 370

370 1999 The Early Holocene Occupation of the Southeastern United States: A Geoarchaeological Summary. In Ice Age Peoples of North America, Environments, Origins, and Adaptations of the First Americans edited by R. Bonnichsen and K. Turnmire, pp. 432481. Oregon State University Press for the Center for the Study of the First Americans, Corvallis. 2001 The 2001 Allendale Paleoindian Expedition and Beyond. Legacy 6:1821. 2003 Backhoes, BBQs, and B Horizons: the 2002 Allendale Paleoindian Expedition. Legacy 7:2229. 2010 Lithic Raw Material Studies in South Carolina and Their Implications for Paleoindian Mobility Patterns and Exchange. South Carolina Antiquities 42:4041. 2013 Update on 20122013 Activities of the Southeastern Paleoamerican Survey. Legacy 17(1):1012. Goodyear, Albert C., III, and Tommy Charles 1984 An Archaeological Survey of Chert Quarries in Western Allendale County, South Carolina. Research Manuscript Series 195. South Carolina Institute of Archaeology and Anthropology, University of South Carolina, Columbia. Goodyear, Albert C., and Kenn Steffy 2003 Evidence of a Clovis Occupation at the Topper Site, 38AL23, Allendale County, South Carolina. Current Research in the Pleistocene 20:2325. Goodyear, Albert C., James L. M ichie, and Barbara A. Purdy 1980 The Edgefield Scraper: A Distributional Study of an Early Archaic Stone Tool from the Southeastern United States. Paper presented at the 37th Annual Southeastern Archaeological Conference, New Orleans, Louisiana. Goodyear, Albert C., Sam B. Upchurch, Mark J. Brooks, and Nancy N. Goodyear 1983 Paleo indian Manifestations in the Tampa Bay Region, Florida. Florida Anthropologist 36(1&2):4066. Goodyear, Albert C., Sam B. Upchurch, Tommy Charles, and Alan Albright 1985 Chert S ources and Paleoindian Lithic Processing in Allendale County, South Carolina. Current Research in the Pleistocene 2:4749. Center for the Study of Early Man, University of Maine at Orono, Orono, Maine. Goodyear, Albert C., Shane Miller, and Ashley Smallwood 2007 Introducing Clovis at the Topper Site, 38AL23, Allendale County, South Carolina. Paper presented at the 72nd Annual Meeting of the Society for American Archaeology, Austin, Texas.

PAGE 371

371 Gosselain, Oliver P. 2000 Materializing Identities: An African Perspective. Journal of Archaeological Method and Theory 7(3):187217. Gould, Richard A. 1969 Yiwara: Foragers of the Australian Desert. Collins, London. 1971 The Archaeologist as Ethnographer: A Case from the Wester n Desert of Australia. World Archaeology 3(2):143177. 1978 The Anthropology of Human Residues. American Anthropologist 80(4):81535. 1982 To Have and Have Not: The Ecology of Sharing among Hunter Gatherers. In Resource Managers: North American and Austr alian Hunter Gatherers edited by Eugene S. Hunn and Nancy M. Williams, pp. 6992. Westview Press, Boulder, Colorado. Gramly, Richard Michael 1994 The Ross Bay Archaic Encampment. The Amateur Archaeologist. 1(1):4556. Grayson, Donald K. 1987 An Analysis of the Chronology of Late Pleistocene Mammalian Extinctions in North America. Quaternary Research 28:281289. Griffin, James B. (editor) 1952 Culture Periods in Eastern United States Prehistory. In Archaeology of Eastern United States pp. 325364. Unive rsity of Chicago Press, Chicago. Grinker, Roy Richard 1994 Houses in the Rainforest: Ethnicity and Inequality among Farmers and Foragers in Central Africa. University of California Press, Berkeley. Gunn, Joel 1996 A Framework for the Paleoindian/Early Archaic Transition. In The Paleoindian and Early Archaic Southeast edited by David G. Anderson and Kenneth E. Sassaman, pp. 415420. University of Alabama Press, Tuscaloosa. Gunn, Joel D., and Irwin Rovner 2003 Horse s Grazing: Point Function and Shape. North Carolina Archaeology 52:53100. Hawkes, Christopher 1954 Archaeological Theory and Method: Some S uggestions from the Old World. American Anthropologist 56(2): 155168. Hayden, Brian, Nora Franco, and Jim Spafford 1996 Evaluating Lithic Strategies and Design Criteria. In Stone Tools: Theoretical Insights into Human Prehistory, edited by George H. Odell, pp. 945. Plenum Press, New York.

PAGE 372

372 Hemmings, C. Andrew 2004 The Organic Clovis: A Single Continent Wide Cultural Adaptation. Ph.D. dissertation, Department of Anthropology, University of Florida, Gainesville. Hendon, Julia A. 2000 Having and Holding: Storage, Memory, Knowledge, and Social Relations. American Anthropologist 102(1):4253. Henrich, Joseph 2012 Hunter gatherer Cooperation. Nature 481(7382):449450. Hiatt, L.R. 1968 Ownership and Use of Land among the Australian Aborigines. In Man the Hunte r, edited by Richard B. Lee and Irven DeVore, pp. 99110. Aldine Publishing Company, Chicago. Hitchcock, Robert K., and Laurence E. Bartram 1998 Social Boundaries, Technical Systems, and the Use of Space and Technology in the Kalahari. In The Archaeology of Social Boundaries edited by Miriam T. Stark, pp. 1249. Smithsonian Institution Press, Washington, D.C Hornum, Michael B., Donald J. Maher, Clifford Brown, Julian Granberry, Frank Vento, Arlene Fradkin, and Michele Williams 1996 Phase III Data Recovery at Site 8LE2105 for the Proposed Florida Gas Transmission Phase III Expansion Project, Leon County, Florida (Volume I of IV Chapters 16). Report prepared by Goodwin and Associates for Florida Gas Transmission Company. On file, Bureau of Historic Preservation, Florida Division of Historic Resources, Tallahassee. Hussey, Tristram C. 1993 A 20,000 Year H istory of Vegetation and Climate at Clear Pond, Northeastern South Carolina. M.A thesis, University of Maine, Orono. Jochim, Michael A. 1991 Archaeology as Long Term Ethnography. American Anthropologist 93(2):308321. Jodry, Margaret A. 2005 Envisioning Water Transport Technology in Late Pleistocene America. In Paleoamerican Origins: Beyond Clovis edited by Robson Bonnichsen, Bradley T. Lepper, Dennis Stanford, and Michael R. Waters, pp.133160. Center for the Study of the First Americans, Te xas A&M University Press, College Station. Johnson, Jay K. 1979 Archaic Biface Manufacture: Production Failures; A Chronicle of the Misbegotten. Lithic Technology 10(23):26 28.

PAGE 373

373 Jones, Sin 1997 The Archaeology of Ethnicity: Constructing Identities in t he Past and Present. Routledge, London. Josselyn, Daniel W. 1964 4 New C 14 Dates from Stanfield Worley Shelter. Stones and Bones Newsletter 6(8):1 2. Keeley, Lawrence H. 1982 Hafting and Retooling: Effects on the Archaeological Record. American Antiquity 47(4):798809. Kelly, Kevin M. 1994 On the Magic Number 500: An Expostulation. Current Anthropology 35(4): 435438. Kelly, Robert L. 1988 The Three Sides of a Biface. American Antiquity 53(4):717734. 1995 The Foraging Spectrum: Diversity in Hunter Gatherer Lifeways. Smithsonian Institution Press, Washington, D.C. Kelly, Robert L., and Lawrence C. Todd 1988 Coming into the Country: Early Paleoindian Hunting and Mobility. American Antiquity 53(2):231244. K ent, Susan 1992 The Current Forager Controversy: Real versus Ideal Views of Hunter Gatherers. Man 27(1):4570. 1993 Sharing in an Egalitarian Community. Man 28(3):479514. Knapp, A. Bernard, and Wendy Ashmore (editors) 1999 Archaeological Landscapes: Con structed, Conceptualized, Ideational. In Archaeologies of Landscape: Contemporary Perspectives, pp. 130. Blackwell Publishers Inc., Malden, Massachusetts Kneberg, Madeline 1956 Some Important Projectile Types Found in the Tennessee Area. Tennessee Archaeologist 13(1): 5556. Krieger, Alex D. 1944 The Typological Concept. American Antiquity 9(3):271288. Kvamme, Kenneth L., Miriam T. Stark, and William A. Longacre 1996 Alternative Procedures for Assessing Standardization in Ceramic Assemblages. Ame rican Antiquity 61(1):116126.

PAGE 374

374 Layton, Robert 1999 The Alawa Totemic Landscape: Ecology, Religion, and Politics. In The Archaeology and Anthropology of Landscape: Shaping your Landscape, edited by Peter J. Ucko and Robert Layton, pp. 219 239. Routledge, L ondon. Lee, Richard B. 1972 The Intensification of Social Life among the !Kung Bushmen, in Population Growth: Anthropological Implications edited by Brian Spooner, pp. 343350. MIT Press, Cambridge. 1979 The !Kung San: Men, Women, and Work in a Foraging Society. Cambridge University Press, Cambridge. 1992 Art, Science, or Politics? The Crisis in Hunter Gatherer Studies. American Anthropologist 94:3154. Lee, Richard B., and Irven DeVore (editors) 1968 Man the Hunter Aldine Publishing Company, Chicago. Lemonnier, Pierre 1992 Elements for an Anthropology of Technology. University of Michigan, Ann Arbor. Levene, Howard 1960 Robust Tests for Equality of Variances. In Contributions to Probability and Statistics: Essays in Honor of Harold Hotelling, edited by Ingram Olkin, S. G. Ghurye, W. Hoeffding, W. G. Madow, and H. B. Mann, pp. 278292. Stanford University Press, California. Lewis, Thomas M.N., and Madeline Kneberg 1960 Aaron B. Clement Collection. Tennessee Archaeologist XVI:1 Tennessee Archaeological Society, Knoxville. Lewis, Thomas M.N., and Madeline Kneberg Lewis 1961 Eva: An Archaic Site. A University of Tennessee Study in Anthropology, The University of Tennessee Press, Knoxville. Lovis, William A., Robert Whallon, and Randolph E. Donahue 2006 Social and Spatial Dimensions of Mesolithic Mobility. Journal of Anthropological Archaeology 25(2):271274. Luchterhand, Kubet 1970 Early Archaic Projectile Points and Hunting Patterns in the Lower Illinois Valley. Report of Investigations, No. 19. Illinois State Museum, Springfield, Illinois. Lyman, R. Lee, Michael J. OBrien, and Robert C. Dunnell 1997 The Rise and Fall of Culture H istory. Plenum Press, New York.

PAGE 375

375 McGahey, Samuel O. 1996 Paleoindian and Early Archaic Data from Mississippi. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 354384. University of Alabama Press, Tuscaloosa. Marrion, Leslie V. 1986 Writing Hand Differences in Kwakiutls and Caucasians. Perceptual and Motor Skills 62:760762. Marshall, Lorna 1976 The !Kung of Nyae Nyae. Harvard University Press, Cambridge. Martin, Paul S. 1973 The Discovery of Americ a. Science 179:969974. 1984 Prehistoric Overkill: The Global Model. In Quaternary Extinctions: A Prehistoric Revolution, edited by Paul S. Martin and Richard G. Klein, pp. 354403. University of Arizona Press, Tucson. Meltzer, David J., and James I. Mea d 1983 The Timing of Late Pleistocene Mammalian Extinctions in North America. Quaternary Research 19:130135. Michie, James L. 1966 The Taylor Point. The Chesopiean 4(5 6):123 124. 1968 The Edgefield Scraper. The Chesopiean 6(2):3031. 1969 Excavations at Toms (Thoms) Creek. South Carolina Institute of Archaeology and Anthropology, University of South Carolina, The Notebook 1(10):216. 1996 The Taylor Site: An Early Occupation in Central South Carolina. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 238269. University of Alabama Press, Tuscaloosa. Midgette, Gordon M. 2005 Discovery and History of the Theriault/Waring Site on Brier Creek, Burke Count y, Georgia. Paper presented at the 62nd Annual Southeastern Archaeological Conference, Columbia, South Carolina. Milanich, Jerald T. 1994 Archaeology of Precolumbian Florida. University of Florida Press, Gainesville.

PAGE 376

376 Miller, D. Shane 2011 Rivers, Rocks and EcoTones: Modeling Clovis Landscape Use in the Southeastern United States. Paper presented at the 76th Annual Meeting of the Society for American Archaeology, Sacramento, California. Morphy, Howard 1995 Landscape and the Reproduction of the Anc estral Past. In The Anthropology of Landscape: Perspectives on Place and Space, edited by Eric Hirsch and Michael OHanlon, pp. 184209. Clarendon Press, Oxford. Morse, Dan F. 1994 Comments on the Dust Cave Investigation. Journal of Alabama Archaeology 40:1 and 2. 232236. 1996 An Arkansas View. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 425429. University of Alabama Press, Tuscaloosa. Morse, Dan F. (editor) 1997a Description of the Artifacts. In Sloan: A Paleoindian Dalton Cemetery in Arkansas, pp. 1452. Smithsonian Institution Press, Washington, D.C. 1997b Sloan: A Paleoindian Dalton Cemetery in Arkansas. Smithsonian Institution Press, Washington, D.C. Morse, Dan F., and Albert C Goodyear, III 1973 The Significance of the Dalton Adze in Northeast Arkansas. Plains Anthropologist 18:316322. Morse, Dan F., David G. Anderson, and Albert C. Goodyear 1996 The PleistoceneHolocene Transition in the Eastern United States. In Humans at the End of the Ice Age: The Archaeology of the Pleistocene Holocene Transition, edited by Lawrence Guy Straus, Berit Valentin Eriksen, Jon M. Erlandson, and David R. Yesner, pp. 319338. Plenum Press, New York. Nassaney, Michael S., and Kenneth E. Sassam an (editors) 1995 Introduction: Understanding Native American Interactions, in Native American Interactions: Multiscalar Analyses and Interpretations in the Eastern Woodlands pp. xix xxxviii. University of Tennessee Press, Knoxville. Neill, Wilfred T. 1971 A Florida Paleoindian Implement of Ground Stone. Florida Anthropologist 24:6170. OSteen, Lisa D. 1983 Early Archaic Settlement Patterns in the Wallace Reservoir: An Inner Piedmont Perspective. M.A. thesis, Department of Anthropology, University of Ge orgia, Athens.

PAGE 377

377 1996 Paleoindian and Early Archaic Settlement along the Oconee Drainage. In The Paleoindian and Early Archaic Southeast edited by David G. Anderson and Kenneth E. Sassaman, pp. 92106. University of Alabama Press, Tuscaloosa. Odell, George H. (editor) 1996 Economizing Behavior and the Concept of Curation. In Stone Tools: Theoretical Insights into Human Prehistory pp. 5180. Plenum Press, New York. Owen, Roger C. 1965 The Patrilocal Band: A Linguistically and Culturally Hybrid Social Unit. American Anthropologist 67(3):675:690. Parker, Bradley J. 2006 Toward an Understanding of Borderland Processes. American Antiquity 71(1):77100. Potter, James M. 2004 The Creation of Person, The Creation of Place: Hunting Landscapes in the American Southwest. American Antiquity 69(2):322338. Powell, John 1990 Points and Blades of the Coastal Plains: A Guide to the Classification of Native American Hafted Implements in the Southeastern Coastal Plain Region. American Systems of the Southeast, Inc., West Columbia, South Carolina. Purdy, Barbara A. 1975 Fractures for the Archaeologist. In Lithic Technology: Making and Using Stone Tools, edited by Earle Swanson, pp. 133141. Mouton, The Hague. 1986 Floridas Prehistoric Stone Technology University Presses of Florida, Gainesville. RadcliffeBrown, A. 1930 Former Numbers and Distribution of the Australian Aborigines. Official Yearbook of the Commonwealth of Australia 23:671696. 1931 Social Organization of Australian Tribes. Oceania I:34 63, 206 246, 322341, 426456. 1956 On Australian Local Organization. American Anthropologist 58:363367. Randall, Asa R. 2001 Untangling Late Paleoindian and Early Side Notched Stone Tool Assemblages at Dust Cave, Alabama. Paper presented at the 58th Annual Southeastern Archaeological Conference, Chattanooga, Tennessee.

PAGE 378

378 2002 Technofunctional Variation in Early Side Notched Hafted Bifaces: A View from the Middle Tennessee River Valley in Northwest Alabama. M.A. thesis, Department of Anthropology, University of Florida, Gainesville. 2010 Remapping Histories: Archaic Period Comm unity Construction along the Middle St. Johns River, Florida. Ph.D. dissertation, Department of Anthropology, University of Florida, Gainesville. Randall, Asa R., and Kandace Detwiler 2002 Stanfield Worley Revisited: The 2001 and 2002 Field Seasons. Paper presented at the 59th Annual Southeastern Archaeological Conference, Biloxi, Mississippi. Sassaman, Kenneth E. 1996 Early Archaic Settlement in the South Carolina Coastal Plain. In The Paleoindian and Early Archaic Southeast edited by David G. Anderson and Kenneth E. Sassaman, pp. 5883. University of Alabama Press, Tuscaloosa. 1998 Crafting Cultural Identity in Hunter Gatherer Economies. In Craft and Social Identity, edited by Cathy Lynne Costin and Rita P. Wright, pp. 93107. Archeological Papers of the American Anthropological Association No. 8. 2001a Articulating Hidden Histories of the Mid Holocene in the Southern Appalachians. In Archaeology of the Appalachian Highlands edited by Lynne P. Sullivan and Susan C. Prezanno, pp. 102120. University o f Tennessee Press, Knoxville. 2001b Hunter Gatherers and Traditions of Resistance. In The Archaeology of Traditions: Agency and History Before and After Columbus edited by Timothy K. Pauketat, pp. 218236. University Press of Florida, Gaine sville. 2003 Crescent Lake Archaeological Survey 2002: Putnam and Flagler Counties, Florida. Technical Report Number 5, Laboratory of Southeastern Archaeology, Department of Anthropology, University of Florida, Gainesville. Electronic document, http://www.anthro.ufl.edu/LSA/publications/LSATechReport5.pdf accessed March 13, 2013. 2010 The Eastern Archaic, Historicized. AltaMira Press, Lanham, Maryland. 2011 History and Alterity in the Eastern Archaic. In Hunter Gatherer Archaeology as Historical Process, edited by Kenneth E. Sassaman and Donald H. Holly Jr., pp. 187208. The University of Arizona Press, Tucson. Sassaman, Kenneth E., and Donald H. Holly Jr. (editors) 2011 Transformative Hunter Gatherer Archaeology in North America. In Hunter Gatherer Archaeology as Historical Process, pp. 113. The University of Arizona Press, Tucson.

PAGE 379

379 Sassaman, Kenneth E., and Wictoria Rudolphi 2001 Communities of Practice in the Early Po ttery Traditions of the American Southeast. Journal of Anthropological Research 57:407425. Sassaman, Kenneth E., Glen T. Hanson, and Tommy Charles 1988 Raw Material Procurement and the Reduction of Hunter Gatherer Range in the Savannah River Valley. Southeastern Archaeology 7:7994. Sassaman, Kenneth E., I. Randolph Daniel, Jr., and Christopher R. Moore 2002 G.S. Lewis East: Early and Late Archaic Occupations Along the Savannah River, Aiken County, South Carolina. Savannah River Archaeological Research Papers 12. Occasional Papers of the Savannah River Archaeological Research Program, South Carolina Institute of Archaeology and Anthropology, University of South Carolina, Columbia. Savannah Upper Ogeechee Water Planning Council 2011 SavannahUpper Ogeechee Regional Water Plan Prepared for the Environmental Protection Division of the Georgia Department of Natural Resources, Atlanta. Electronic document, http://www.savannahupperogeechee.org/documents/SUO_Adopted_RWP.pdf accessed March 22, 2013. Schroder, Lloyd E. 2002 The Anthropology of Florida Points and Blades. American Systems of the Southeast, Inc., West Columbia, South Carolina. Service, Elman 1962 Primitive Social Organization. Random House, New York. 1966 The Hunters. PrenticeHall, Englewood Cliffs, New Jersey. Sherwood, Sarah C., Boyce N. Driskell, Asa R. Randall, and Scott C. Meeks 2004 Chronology and Stratigraphy at Dust Cave, Alabama. Am erican Antiquity 69(3):533554. Smallwood, Ashley M. 2010 Clovis Biface Technology at the Topper site, South Carolina: Evidence for Variation and Technological Flexibility. Journal of Archaeological Science 37:24132425. 2012 Clovis Technology and Settlement in the American Southeast: Using Biface Analysis to Evaluate Dispersal Models. American Antiquity 77(4):689713. Smith, Claire 1999 Ancestors, Place and People: Social Landscapes in Aboriginal Australia. In The Archaeology and Anthropology of L andscape: Shaping your Landscape, edited by Peter J. Ucko and Robert Layton, pp. 189205. Routledge, London.

PAGE 380

380 Snow, Frankie 1977a An Archaeological Survey of the Ocmulgee Big Bend Region. Occasional Papers from South Georgia 3, South Georgia College, Dougl as. 1977b A Survey of the Ocmulgee Big Bend Region. Early Georgia 5(1 and 2):3760. Spaulding, Albert C. 1953a Review of Measurements of Some Prehistoric Design Developments in the Southeastern States by J.A. Ford. American Anthropologist 55:588591. 1953b Statistical Techniques for the Discovery of Artifact Types. American Antiquity 18(4):305313. 1954a Reply (to Ford). American Anthropologist 56:112114. 1954b Reply (to Ford). American Antiquity 19(4):391393. 1982 Structure in Archaeological Data: Nominal Variables. In Essays on Archaeological Typology, edited by Robert Whallon and James A. Brown, pp. 1 20. Center for American Archeology Press, Evanston, Illinois. Stanford, Dennis J., and Bruce A. Bradley 2002 Ocean Trails and Prairie Paths? Thoughts about Clovis Origins. In The First Americans: The Pleistocene Colonization of the New World, edited by Nina G. Jablonski, pp. 255271. Memoir of the California Academy of Sciences, Number 27, San Francisco. Stark, Miriam T. (editor) 1998 Technical Choices and Social Boundaries in Material Culture Patterning: An Introduction. In The Archaeology of Social Boundaries pp. 111. Smithsonian Institution Press, Washington, D.C. Steponaitis, Vincas P. 1986 Prehistoric Archaeology in the Southeastern Uni ted States, 19701985. Annual Review of Anthropology 15:363404. Steward, Julian H. 1936 The Economic and Social Basis of Primitive Bands. In Essays in Anthropology Presented to Alfred Louis Kroeber, edited by Robert H. Lowie, pp. 331 350. University of C alifornia Press, Berkeley. 1969 Postscript to Bands: On Taxonomy, Processes, and Causes. In Contributions to Anthropology: Band Societies edited by David Damas. National Museums of Canada, Bulletin 228:288295.

PAGE 381

381 Stuiver, Minz, Paula J. Reimer, Edouard Bard, J. Warren Beck, G.S. Burr, Konrad A Hughen, Bernd Kromer, Gerry McCormac, Johannes van der Plicht, and Marco Spurk 1998 INTCAL98 Radiocarbon Age Calibration, 240000 cal B.P. Radiocarbon 40(3):10411083. Swar tz, Marian 1997 A History of Dogs in the Early Americas. Yale University Press, New Haven. Thomas, David Hurst 1986 Points on Points: A Reply to Flenniken and Raymond. American Antiquity 51(3):619627. Thomas, Julian 1993 The Hermeneutics of Megalithic Space. In Interpretative Archaeology edited by Christopher Tilley, pp. 7397. Berg, Oxford. Thulman, David K. 2006 A Reconstruction of Paleoindian Social Organization in North Central Florida. Ph.D. dissertation, Departmen t of Anthropology, Florida State University, Tallahassee. 2007 A Typology of Fluted Points from Florida. The Florida Anthropologist 60(4):6375. Tilley, Christopher 1994 A Phenomenology of Landscape: Places, Paths, and Monuments. Berg, Oxford. Titmus, Gene 1985 Some Aspects of Stone Tool Notching. In Stone Tool Analysis: Essays in Honor of Don E. Crabtree, edited by Marc G. Plew, J.C. Woods, and Max G. Pavesic, pp. 243264. University of New Mexico Press, Albuquerque. Tomczak, Paula D., and Joseph F. P owell 2003 Postmarital Residence Practices in the Windover Population: Sex Based Dental Variation as an Indicator of Patrilocality. American Antiquity 68(1):93108. Towner, Ronald H., and Miranda Warburton 1990 Projectile Point Rejuvenation: A Technologic al Analysis Journal of Field Archaeology 17:317321. Tuross, Noreen, Marilyn L. Fogel, Lee Newsom, and Glen H. Doran 1994 Subsistence in the Florida Archaic: Stable Isotope and Archaeobotanical Evidence from the Windover Site. American Antiquity 59(2):288303. University of South Alabama 2004 Ocala Chert Distribution Electronic document, http://www.geoarchaeology.southalabama.edu/ocala_distribution.html accessed October 10, 2012.

PAGE 382

382 Upchurch, Sam B., Richard N. Strohm, and Mark G. Nuckels 1982 Methods of Provenance Determination of Florida Cherts. Report prepared for the Florida Division of Archives, History, and Records Management by the Department of Geology, University of South Florida, Tampa. 1984 Petrology of Selected Lithic Materials from the South Carolina Coastal Plain, Appendix A. In An Archaeological Survey of Chert Quarries in Western Allendale County, South Carolina, edited by Albert C. Goodyear and Tommy Charles pp. 125160. Research Manuscript Series 195. Institute of Archaeology and Anthropology, University of South Carolina, Columbia. Van Buren, G.E. 1974 Arrowheads and Projectile Points with a Classification Guide for Lithic Artifacts. Arrowhead Publishing Company, Garden Grove, California. Walker, Renee B. 2007 Hunting in the Late Paleoindian Period: Faunal Remains from Dust Cave. In Foragers of the Terminal Pleistocene in North America, edited by Renee B. Walker and Boyce N. Driskell pp. 99115. University of Nebraska Press, Lincoln. Waller, Ben I. 1971 Hafted Flake Knives. Florida Anthropologist 24:173174. Ward, H. Trawick 1983 A Review of Archaeology in the North Carolina Piedmont: A Study of Change. In Prehistory of North Carol ina: An Archaeological Symposium edited by M.A. Mathis and J.A. Crow, pp. 5381. North Carolina Division of Archives and History, Department of Cultural Resources, Chapel Hill. Waters, Michael R., and Thomas W. Stafford Jr. 2007 Redefining the Age of Clovis: Implications for the Peopling of the Americas. Science 315:11221126. Watts, W.A. 1980a Late Quaternary Vegetation History at White Pond on the Inner Coastal Plain of South Carolina. Quaternary Research 13(2):187199. 1980b Late Wisconsin Climate of Northern Florida and Origin of Species Rich Deciduous Forest. Science 210:325327. Webb, S. David, Jerald T. Milanich, Roger Alexon, and James S. Dunbar 1984 A Bison antiquus Kill Site, Wacissa River, Jefferson County, Florida. A merican Antiquity 49(2):384392.

PAGE 383

383 Whallon, Robert 2006 Social Networks and Information: Non- utilitarian Mobility among Hunter gatherers. Journal of Anthropological Archaeology 25(2):259270. Whallon, Robert W., William A. Lovis, and Robert K. Hitchcock (editors) 2011 Information and its Role in Hunter Gatherer Bands UCLA Cotsen Institute of Archaeology, Los Angeles. Whatley, John S. 2002 An Overview of Georgia Projectile Points and Selected Cutting Tools. Early Georgia 30(1). Wiessner, Polly 1977 Hxaro: A Regional System of Reciprocity for Reducing Risk among the !Kung San. Ph.D. dissertation, Department of Anthropology, University of Michigan, Ann Arbor. 1982 Risk, Reciprocity, and Social Influences on !Kung San Economics. In Politics and H istory in Band Societies, edited by Eleanor Leacock and Richard B. Lee, pp. 6184. Cambridge University Press, Cambridge. 1983 Style and Social Information in Kalahari San Projectile Points. American Antiquity 48(2): 253276. 1984 Reconsidering the Behavioral Basis for Style: A Case Study among the Kalahari San. Journal of Anthropological Archaeology 3:190234. Willey, Gordon R., and Philip Phillips 2001 [1958] Method and Theory in American Archaeology, edited by R. Lee Lyman and M ichael J. OBrien. University of Alabama Press, Tuscaloosa. Williams, Nancy 1982 A Boundary is to Cross: Observations on Yolngu Boundaries and Permission. In Resource Managers: North American and Australian Hunter Gatherers, edited by Nancy Williams and E ugene S. Hunn, pp. 131153. Westview Press, Boulder, Colorado. Wobst, H. Martin 1978 The Archaeo Ethnology of Hunter Gatherers or the Tyranny of the Ethnographic Record in Archaeology. American Antiquity 43(2):303309. Wolf, Eric R. 1982 Europe and the People Without History The University of California Press, Berkeley. 1984 Culture: Panacea or Problem? American Antiquity 49(2)393400.

PAGE 384

384 Woodburn, James 1968 Stability and Flexibility in Hadza Residential Groupings. In Man the Hunter edited by Richard B. Lee and Irven L. DeVore, pp. 4955. Aldine Publishing Company, Chicago. 1995 African Hunter Gatherer Social Organization: Is It Best Understood as a Product of Encapsulation? In Hunters and Gatherers, Vol. 1: History, Evolution, and Social Change, edited by Tim Ingold, David Riches, and James Woodburn, pp. 3164. Berg, New York. Wright, Henry T. 1996 Comments. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 430433. University of Alabama Press, Tuscaloosa. Yellen, John E. 1977 Archaeological Approaches to the Present: Models for Reconstructing the Past. Academic Press, New York. Yellen, John, and Henry Harpending 1972 Hunter Gatherer Populations and Archaeological Inference. World Archae ology 4:244253. Yerkes, Richard, and Linda M. Gaertner 1997 Microwear Analysis of Dalton Artifacts. In Sloan: A Paleoindian Dalton Cemetery in Arkansas, edited by Dan F. Morse, pp. 5871. Smithsonian Institution Press, Washington, D.C.

PAGE 385

385 BIOGRAPHICAL SKETCH Kara Bridgman Sweeney graduated from the South Carolina Honors College at the University of South Carolina in 1995 with a B.A. in anthropology, and received a M.A. in archaeology from National University of Irelands University College Cork in 1999. After three years of employment as a cultural resource management archaeologist in Charleston, South Carolina, she began her coursework at the University of Florida. Kara served as an instructor of anthropology at Kennesaw State University (in the Georgia University sys tem) fo r five semesters, culminating in 2006. In 2006 and 2007, she served as Regional Director of the Southwest Region of the Florida Public Archaeology Network. Currently, Kara is an archaeologist with Brockington and Associates, Inc. in S avannah, Georgia.