A Household Approach to Reconstructing the Townsend Sites in East Tennessee, U.S.A.: Foodways and Daily Practice Within a Mississippian Settlement

University of Tennessee, Knoxville TRACE: Tennessee Research and Creative Exchange

Doctoral Dissertations Graduate School

8-2017

A Household Approach to Reconstructing the Townsend Sites in East Tennessee, U.S.A.: Foodways and Daily Practice within a Mississippian Settlement

Jessie Luella Johanson University of Tennessee, Knoxville, [email protected]

Follow this and additional works at: https://trace.tennessee.edu/utk_graddiss

Part of the Archaeological Anthropology Commons

Recommended Citation Johanson, Jessie Luella, "A Household Approach to Reconstructing the Townsend Sites in East Tennessee, U.S.A.: Foodways and Daily Practice within a Mississippian Settlement. " PhD diss., University of Tennessee, 2017. https://trace.tennessee.edu/utk_graddiss/4630

This Dissertation is brought to you for free and open access by the Graduate School at TRACE: Tennessee Research and Creative Exchange. It has been accepted for inclusion in Doctoral Dissertations by an authorized administrator of TRACE: Tennessee Research and Creative Exchange. For more information, please contact [email protected]. To the Graduate Council:

I am submitting herewith a dissertation written by Jessie Luella Johanson entitled "A Household Approach to Reconstructing the Townsend Sites in East Tennessee, U.S.A.: Foodways and Daily Practice within a Mississippian Settlement." I have examined the final electronic copy of this dissertation for form and content and recommend that it be accepted in partial fulfillment of the requirements for the degree of Doctor of Philosophy, with a major in Anthropology.

Kandace D. Hollenbach, Major Professor

We have read this dissertation and recommend its acceptance:

David G. Anderson, Barbara J. Heath, Sally P. Horn

Accepted for the Council:

Dixie L. Thompson

Vice Provost and Dean of the Graduate School

(Original signatures are on file with official studentecor r ds.) A Household Approach to Reconstructing the Townsend Sites in East Tennessee, U.S.A.: Foodways and Daily Practice within a Mississippian Settlement

A Dissertation Presented for the Doctor of Philosophy Degree The University of Tennessee, Knoxville

Jessie Luella Johanson August 2017

DEDICATION

To my husband, Erik and our daughter Lilia

iii

ACKNOWLEDGEMENTS

I would first like to thank my advisor, Dr. Kandace Hollenbach, for her steady support and mentoring throughout the completion of my MA and PhD at the University of Tennessee. Dr. Hollenbach’s genuine passion for the disciplines of archaeology and paleoethnobotany inspired me to pursue this path. She spent countless hours across the table at the microscope with me to teach and discuss the intricacies of macrobotanical identification. She also gave me many opportunities to become involved in research projects, both in the laboratory and field, that financially supported my academic career. I thank her for her encouragement to move beyond a specialization and to become a balanced archaeologist. I am also grateful for her willingness to work with me to meet deadlines that seemed impossible at times.

The other members of my doctoral committee provided guidance throughout my academic career at the University of Tennessee. Dr. David Anderson, in addition to providing immeasurable insight into Mississippian period archaeology, was always a friendly conversation away. Drs. Barbara Heath and Sally Horn provided a refreshing perspective on my research, providing comments and suggestions that greatly improved this dissertation.

There are many people at the Department of Anthropology and Archaeological

Research Laboratory (ARL) with the University of Tennessee who helped me along the way. The Townsend Archaeological Project was an undertaking that involved years of field and laboratory analysis. This dissertation research owes much to the previous

work that it was built upon, especially that of the Mississippian period team, Shannon

Koerner and Dr. Lynn Sullivan. I would also like to thank Mike Angst, director of the

ARL, for frequently answering questions about the Townsend project and selflessly assisting me with accessing curation materials that were always on the highest shelf.

I couldn’t have completed this dissertation without the dedication and infinite support of my family. My parents, Marilyn and Bob, and my mother-in-law Sandra provided countless hours of loving babysitting that allowed me to complete my work knowing that Lilia was in the best hands. My sister Sarah was an especially supportive friend and person to talk with throughout the process. I would also like to thank Jubal,

Curtis, Karna, Derrick, Ella, and Emmett. Most importantly, the final years of my dissertation research I gratefully got to share with my husband, Erik, and our baby girl,

Lilia, both of whom made a challenging time the happiest point of my life. Lilia, thank you for the giggles and smiles. Erik, thank you for your encouragement and advice and for being there for “all” of the moments as a partner and friend.

ABSTRACT

This study examines how foodways differences between the multiple

Mississippian settlements that were occupied circa 900 to 1300 CE at the Townsend sites

(40BT89, 40BT90, and 40BT91) in East Tennessee, U.S.A., reflect the distinct choices people made in response to variation in the social conditions they faced in a boundary location. Located in a narrow valley cove at the foothills of the Smoky Mountains, these sites lie between two physiographic provinces, the Ridge and Valley Province to the west and the Blue Ridge Mountains Province to the east, as well as between two cultural traditions, the Hiwassee Island to the west and the Pisgah to the east. The location of the sites can be viewed as an opportunity to escape regional social pressure, or alternatively as a strategy to place a community as a point of contact between diverse natural resources and cultural groups.

I contribute to the growing body of work in the Southeast that challenges the construction of Mississippian lifeways as driven by elites and large Mississippian polities. I use foodways as a lens to understand how the actions of everyday life, the performance of daily activities, reflect group decisions made in response to social and/or environmental circumstances. Using the multiple settlements, both farmsteads and villages, across the sites, I perform a household analysis to identify differences in the use and placement of features, as well as the distribution and spatial patterning of plant taxa and ceramic vessels with different temper and surface treatments.

Variation in foodways traditions between the different settlements indicate that people responded to their social circumstances in various ways, particularly growing and processing corn in different places. At Village 1, I suggest that residents used infield instead of outfield cultivation, which is an indication of increasing social pressure in the region. Multiple lines of evidence at Village 1 also point to the occupation of this single settlement by both Hiwassee Island and Pisgah related people, which I argue is evidence of a unique instance of cultural solidarity in this highland location.

vii

TABLE OF CONTENTS

CHAPTER 1 INTRODUCTION ...... 1

CHAPTER 2 ARCHAEOLOGICAL PERSPECTIVES ON MISSISSIPPIAN IDENTITY: WHAT DOES IT MEAN TO BE MISSISSIPPIAN? ...... 9 2.1 Introduction ...... 10 2.2 A Construction of Mississippian Identity in the Southeast: From the Traditional to New Approaches ...... 12 2.2.1 The Mississippian Emergence ...... 12 2.2.2 Where Do Sites Fit Within the Chiefdom Model: How to Define Complexity? 15 2.2.3 What is the source of power and inequality and how is it expressed? ...... 19 2.3 Situating Mississippian Foodways at the Townsend Sites ...... 21 2.3.1 Agent-Oriented Archaeology and Foodways ...... 22 2.3.2 The Benefits of the Household as a Comparative Lens at the Townsend Sites 24 2.4 Summary ...... 28

CHAPTER 3 THE NATURAL AND CULTURAL LANDSCAPE OF TUCKALEECHEE COVE: THE TOWNSEND ARCHAEOLOGICAL SITES IN CONTEXT ...... 30 3.1 Introduction ...... 31 3.2 The Natural Landscape ...... 32 3.2.1 Physiography ...... 32 3.2.2 Hydrology of the Little River ...... 34 3.2.3 Geology and Soil Morphology ...... 36 3.2.4 Climate ...... 39 3.2.5 Surrounding Plant and Animal Communities ...... 42 3.2.5.1 Fauna ...... 43 3.2.5.2 Flora ...... 44 3.2.6 A Changing Environment ...... 49 3.3 Cultural History of the Region ...... 50 3.3.1 Paleoindian Era Settlement and Land Use (>13,250–11,700 cal year BP) ...... 50 3.3.2 The Archaic Period: Establishing Patterns (11,500–3200 cal yr BP) ...... 53 3.3.3 The Woodland Period: Trends Writ Larger (3000–1000 cal yr BP) ...... 58 3.3.4 The Mississippian Period (1020–500 cal yr BP) ...... 63

CHAPTER 4 ARCHAEOLOGICAL EXCAVATIONS AT THE TOWNSEND SITES ...... 73 4.1 Introduction ...... 74 4.2 Artifact and Data Management ...... 78 viii

4.3 The Townsend Mississippian Occupation: Sites 40BT89, 40BT90, and 40BT91 ...... 79 4.3.1 Identification of Mississippian Period Deposits ...... 80 4.3.2 Material remains from the Kinzel Springs Site (40BT89) ...... 86 4.3.3 Material Remains from the Apple Barn Site (40BT90) ...... 87 4.3.4 Cultural remains from the Pony Ride Site (40BT91) ...... 100 4.3.4 Summary of Previous Interpretations ...... 112 4.4 Classifications Used in this Study ...... 114

CHAPTER 5 ANALYSIS AND INTERPRETATION OF FEATURES ...... 122 5.1 Introduction ...... 123 5.2 Feature Classification ...... 128 5.3 A Comparison of Storage Practices at the Townsend Sites ...... 132 5.4 Differences in the Distribution of Cooking and Hearth Features ...... 143 5.5 A Comparison of the Domestic Yard between Farmsteads and Villages ...... 147 5.6 A Comparison of Feature Content ...... 154 5.7 A Comparison of the Storage Capacity of Features ...... 155 5.8 Summary and Conclusions ...... 162

CHAPTER 6 PLANT REMAINS AND THE MISSISSIPPIAN PEOPLE OF TUCKALEECHEE COVE ...... 165 6.1 Introduction ...... 166 6.2 Methods ...... 167 6.2.1 Field Recovery and Sampling Procedures ...... 167 6.2.2 Identifications of Plant Remains ...... 168 6.2.3 A Consideration of Biases within the Dataset ...... 170 6.2.4 Statistical Analyses Employed for Interpretation ...... 172 6.3 Paleoethnobotanical Analysis Results ...... 173 6.4 Interpretations of Mississippian Plant Use in Tuckaleechee Cove ...... 187 6.4.1 Overview of Plant Categories Commonly Found at the Sites ...... 188 6.4.1.1 Mast Resources ...... 189 6.4.1.2 Crops and Farming ...... 190 6.4.1.3 Fruits ...... 194 6.4.1.4 Indigenous Seed Crops ...... 195 6.4.1.5 Miscellaneous Taxa ...... 196 6.5 Seasonality of the Sites Based on the Plant Remains ...... 197 6.6 Temporal Patterns of Plant Use ...... 198 6.7 A Comparison of the Various Contexts at the Sites ...... 199 6.7.1 Unique Features from across the Sites ...... 205 ix

6.7.2 A Comparison of Plant remains by Feature Function ...... 208 6.7.3 The Importance of Place: Patterns of Plant Consumption Based on Settlement Type ...... 212 6.8 Summary and Conclusions ...... 224

CHAPTER 7 HOUSEHOLD ANALYSIS OF THE TOWNSEND SITES ...... 227 7.1 Introduction ...... 228 7.2 Correspondence Analysis: Use in Paleoethnobotanical Assemblages ...... 229 7.3 Correspondence Analysis: The Townsend Sites ...... 231 7.3.1 Methods ...... 231 7.3.2 Results ...... 233 7.4 Discussion ...... 250 7.5 Conclusions ...... 253

CHAPTER 8 SUMMARY AND CONCLUSIONS ...... 254

REFERENCES ...... 264

APPENDIX ...... 294

VITA ...... 346

LIST OF TABLES

Table 3.1. Dominant Soil Series in the Townsend Archaeological Site Area ...... 40 Table 4.1. Mississippian Cultural Phases in East Tennessee ...... 84 Table 4.2. Mississippian Cultural Phases in Western North Carolina ...... 85 Table 4.3. Radiocarbon Dates from Site 40BT89 ...... 89 Table 4.4. Radiocarbon Dates from Site 40BT90 ...... 98 Table 4.5. Radiocarbon Dates from Site 40BT91 ...... 102 Table 4.6. All Households with Associated Subterranean Storage Features at the Townsend Sites ...... 118 Table 4.7. Dates of the Settlements at Sites 40BT90 and 40BT91 ...... 120 Table 5.1. Mississippian Period Feature Classifications ...... 124 Table 5.2. Comparison of Distance of Cooking Features from Associated Households Across Farmsteads and Villages at the Townsend Sites ...... 150 Table 5.3. Comparison of Storage Feature Distance from Associated Households Across Farmsteads and Villages at the Townsend Sites ...... 151 Table 5.4. List of Storage Capacity of All Households at the Townsend Sites...... 160 Table 6.1. Analyzed Mississippian Flotation Samples from the Townsend Sites...... 175 Table 6.2. Plant Remains from Mississippian Contexts at the Townsend Sites ...... 180 Table 6.3. Ubiquity of Plant Remains Recovered from 88 Features Across the Townsend Sites. . 184 Table 6.4. Comparisons of Nutshell to Maize from Regional Mississippian Sites ...... 192 Table 7.1. Diagnostic Statistics for Correspondence Analysis of Plant Taxa from Townsend Mississippian Households ...... 236 Table 7.2. Loading of Variables on Factors 1 and 2 of Plant Taxa from Mississippian Townsend Sites ...... 237 Table 7.3. Diagnostic Statistics for Correspondence Analysis of Ceramic Temper and Exterior Treatment from Townsend Mississippian Households ...... 242 Table 7.4. Loading of Variables on Factors 1 and 2 of Ceramic Temper and Exterior Surface Treatment from Mississippian Townsend Households ...... 243 Table 7.5. Diagnostic Statistics for Correspondence Analysis of Combined Plant Taxa and Ceramic Temper and Exterior Surface Treatment from Townsend Mississippian Households .. 247 Table 7.6. Loading of Variables on Factors 1 and 2 of Combined Plant Taxa and Ceramic Temper and Exterior Surface Treatment from Mississippian Townsend Households ...... 248 Table A.1. Mississippian Pit Features (n=115) from the Townsend Sites……………………...…...295 Table A.2. Domestic Architecture from the Townsend sites……………………………...…………298 Table A.3. Community Buildings from the Townsend Site of 40BT91…………...………………...300 Table A.4. Corncrib Structures from the Townsend Sites…………………….……………………...301 Table A.5. Primary Enclosures at the Townsend Sites………………………………...……………..302 Table A.6. Mississippian Plant Remains Recovered by Context from the Townsend Sites……...303

LIST OF FIGURES

Figure 3.1. Location of the Townsend archaeological sites along the Little River ...... 35 Figure 3.2. Landforms of the Townsend sites ...... 38 Figure 3.3. Appalachian Broadleaf Ecoregion (Level II) setting with Townsend sites ...... 46 Figure 3.4. Blue Ridge Ecoregion (Level III) setting with Townsend sites...... 47 Figure 4.1. Townsend site overview for 40BT89, 40BT90, and 40BT91 ...... 76 Figure 4.2. Architectural remains from sites 40BT90 (A) and 40BT91 (B) ...... 82 Figure 4.3. Two community buildings, Structures 1 (designated as M here) and 13 (designated as N here) located in Village 1 at site 40BT91 ...... 83 Figure 4.4. Four areas of investigation at site 40BT89 designated as Area A, B, C, and C-Short Creek ...... 88 Figure 4.5. The distribution of Early, Middle, and General Mississippian features from site 40BT89 showing isolated feature in northwest corner of the site ...... 90 Figure 4.6. The distribution of Early, Middle, and General Mississippian features from site 40BT89 moving east (1) ...... 91 Figure 4.7. The distribution of Early, Middle, and General Mississippian features from site 40BT89 moving east (2) ...... 92 Figure 4.8. Three areas of investigation, A, B, and C, at site 40BT90 ...... 93 Figure 4.9. The distribution of Early, Middle, and General Mississippian features and postholes representative of structures and palisade remnants from site 40BT90 moving east ...... 94 Figure 4.10. The distribution of Early, Middle, and General Mississippian features and postholes representative of structures and palisade remnants from site 40BT90 moving east (2) ...... 95 Figure 4.11. The distribution of Early, Middle, and General Mississippian features and postholes representative of structures and palisade remnants from site 40BT90 moving east (3) ...... 96 Figure 4.12. The distribution of Early, Middle, and General Mississippian features and postholes representative of structures and palisade remnants from site 40BT90 moving east (4) ...... 97 Figure 4.13. Three areas of investigation, Area A, B, and C, at site 40BT91 ...... 101 Figure 4.14. Village 1 located at the western edge of site 40BT91 ...... 103 Figure 4.15. Pisgah affiliated Structure 22 in the northern portion of Village 1 at site 40BT91. .... 106 Figure 4.16. Isolated Pisgah affiliated farmstead in the southeastern portion of site 40BT91 ...... 107 Figure 4.17. One of two isolated farmsteads/domestic debris at site 40BT91 including a single feature, and postholes representative of a single domestic structure (Structure 5), corncrib structures (Structures 14 and 15), and palisade remnants ...... 108 Figure 4.18. Second of two isolated farmsteads at site 40BT91 including features and potential domestic debris (Structure A) ...... 109 Figure 4.19. Domestic compound on eastern boundary of site 40BT91 (designated as Village 2 in this dissertation) including features, a floor midden (Structure 9), and postholes representative of

xii

domestic structures (28, 29, 30, 32, and 23/24), corncribs (Structure 25 and 7), and palisade remnants...... 110 Figure 4.20. Location of farmsteads, villages, and households from Townsend site 40BT90...... 116 Figure 4.21. Location of farmsteads, villages, and households from Townsend site 40BT91...... 117 Figure 5.1. Example of the morphology of a cooking basin in profile (top: Feature 10 from 40BT90) compared to a cooking pit in profile (bottom: Feature 590 from 40BT90)...... 133 Figure 5.2. Example of a storage basin in profile (top: Feature 580 from 40BT90) compared to a storage pit (bottom: Feature 2002 from 40BT90) in profile...... 134 Figure 5.3. Example of a hearth in profile (Feature 530 from 40BT91)...... 135 Figure 5.4. Histogram of volume of Mississippian basins...... 136 Figure 5.5. Histogram of volume of Mississippian pits...... 137 Figure 5.6. Histogram of FCR density (L) of Mississippian basins ...... 138 Figure 5.7. Histogram of FCR density (L) of Mississippian pits ...... 139 Figure 5.8. Boxplots comparing the density of FCR (g/L) by Mississippian ...... 140 Figure 5.9. Histogram depicting the distribution of distance of features from the closest household structure in "farmstead" settlement type...... 148 Figure 5.10. Histogram depicting the distribution of distance of features from the closest household structure in "Village" settlement type...... 149 Figure 5.11. Boxplots comparing the density of lithic debitage and ceramic sherds by Mississippian morphological category ...... 156 Figure 5.12. Boxplots comparing the density of ceramic sherds and lithic debitage by Mississippian functional category ...... 157 Figure 6.1. Comparison of plant and wood density by temporal-cultural affiliation ...... 200 Figure 6.2. Comparison of the relative densities of nuts (hickory, black walnut, and acorn) by temporal cultural affiliation ...... 201 Figure 6.3. Comparisons of the relative density of corn kernels and corn cupules by temporal cultural affiliation ...... 202 Figure 6.4. Comparison of kernel:cupule ratio by temporal cultural affiliation ...... 203 Figure 6.5. Comparison of relative density of fruits and edible seeds by temporal cultural affiliation ...... 204 Figure 6.6. Comparison of plant and wood density by feature function ...... 209 Figure 6.7. Comparison of the relative density of wood and nuts by feature function ...... 210 Figure 6.8. Comparison of corn cupules and kernels by feature function ...... 211 Figure 6.9. Comparison of plant and wood density by community type ...... 213 Figure 6.10. Comparison of relative density of acorn, black walnut, corn cupules and kernels, and fruit by community ...... 214 Figure 6.11. Comparison of the relative density of hickory by community ...... 215 Figure 6.12. Comparison of corn kernel:cupule ratio by community ...... 216 Figure 6.13. Comparison of relative density of hickory by community type ...... 219 Figure 6.14. Comparison of relative density of acorn and black walnut by community type ...... 220 xiii

Figure 6.15. Comparison of the relative density of fruit by community type ...... 221 Figure 6.16. Comparison of relative density of corn remains by community type ...... 222 Figure 7.1. Biplot depicting the results of the correspondence analysis of plant remains across Mississippian households at the Townsend sites...... 234 Figure 7.2. Biplot of plant taxa identified in Mississippian household features from the Townsend sites...... 235 Figure 7.3. Biplot depicting the results of the correspondence analysis of ceramic temper and exterior treatment across Mississippian households at the Townsend sites...... 240 Figure 7.4. Biplot of ceramic temper and exterior treatments identified in Mississippian household features from the Townsend sites...... 241 Figure 7.5. Combined biplot depicting the results of the correspondence analysis of plant taxa and ceramic temper and exterior treatment identified in Mississippian household features from the Townsend sites ...... 245 Figure 7.6. Combined biplot of plant taxa and ceramic temper and exterior treatments identified in Mississippian household features from the Townsend sites...... 246 Figure 7.7. Biplot depicting the results of the correspondence analysis suggesting a clustering of Mississippian into four general groups...... 251 Figure A.1. Household A from the Townsend site of 40BT90………………………………………332 Figure A.2. Household B from the Townsend site of 40BT90……………………………………….333 Figure A.3. Household C from the Townsend site of 40BT90….……………………………………334 Figure A. 4 Household D from the Townsend site of 40BT90.………………………………………335 Figure A.5. Households E and F from the Townsend site of 40BT90……………………………….336 Figure A.6. Household G from the Townsend site of 40BT90……………………………………….337 Figure A.7. Households H and J from the Townsend site of 40BT91……………………………….338 Figure A.8. Community Buildings M and N from the Townsend site of 40BT91…………………339 Figure A.9. Household O from the Townsend site of 40BT91……………………………………….340 Figure A.10. Household P from the Townsend site of 40BT91……………………………………...341 Figure A.11. Household Q from the Townsend site of 40BT91……………………………………...342 Figure A.12. Household R from the Townsend site of 40BT91……………………………………...343 Figure A.13. Household S from the Townsend site of 40BT91………………………………………344 Figure A.14. Households T and U from the Townsend site of 40BT91……………………………..345

xiv

CHAPTER 1

INTRODUCTION

Archaeological work on a greater breadth of geographic areas in the Southeast, upland and rural locations, and on smaller sites are refocusing attention on the people and the communities that have been overshadowed archaeologically by the more conspicuous cultural material of larger site complexes such as mounds, status goods, and iconographic resources and that are often unaccounted for in traditional social models. This more recent body of work is not a rejection of previous research that has driven our understanding of the influence of large Mississippian centers, but an appreciation of the decisions of individuals to participate in, challenge, or adapt to the often-imposing set of cultural traditions known as Mississippian.

Large mound complexes such as Cahokia in the American Bottom and

Moundville and Etowah in the Southeast typify the kind of political centers that would have exerted significant social, economic, and political control over populations both locally and regionally. However, Cahokia’s Mississippian rise and expansion was a result of complex intergroup interactions that involved multiple immigrant groups and sustained relationships with upland and hinterland peoples (Alt 2002; Pauketat 2003).

Increasing archaeological coverage of a wider range of sites is revealing the myriad of ways that people in the Mississippian period responded to a changing cultural landscape, and a result is a growing appreciation for greater diversity in the social, political, and economic fabric of the everyday Mississippian experience (e.g. Alt 2001;

Boudreaux 2013; Clay 2006; Emerson 1997; Maxham 2000; Mehrer 1995; Meyers 2002;

Scarry and Scarry 2005; Thomas 2001).

Daily life in smaller Mississippian communities in the uplands of the Southern

Appalachian remain poorly known. Extensive archaeological research of Mississippian period sites in the greater Tennessee and Little Tennessee River Valleys (e.g. Chapman

1979; Hatch 1976; Honerkamp et al. 1989; Honerkamp 1990a, 1990b; Koerner et al. 2011;

Lewis et al. 1995; Lengyel et al. 1999; Lewis and Kneberg 1946; Polhemus 1987; Salo

1969a; Schroedl 1975; Schroedl et al. 1985; Sullivan 1987, 1989, 2007; Sullivan and Humpf

2001), was prompted both by the Tennessee Valley Authority’s reservoir projects in the

1930’s and the beginning of the Cultural Resource Management era in the 1960’s.

Although these East Tennessee excavations (e.g. Norris, Chickamagua, and Watts Bar,

Douglas Reservoir, Hiwassee Island, Tellico Reservoir) revolutionized our understanding of Tennessee prehistory, they were primarily located along the large river valleys where flooding was imminent, at the expense of greater knowledge of smaller tributaries and upland areas (Faulkner 2002:179).

In this study, I contribute to the developing body of work in the Southeast that challenges the construction of Mississippian identity as driven by elites and large

Mississippian polities. I take advantage of one of the few excavations that have been conducted in the Southern Appalachian region of East Tennessee, of sites 40BT89,

40BT90, and 40BT91, collectively known as the Townsend sites. The geographic location of the Townsend sites lies between two physiographic provinces, with the Ridge and

Valley Province to the west and the Blue Ridge Mountains Province to the east, as well as between two cultural groups, the Hiwassee Island to the west and the Pisgah to the

East. Thus, the frontier location of the site can be viewed as an opportunity to escape regional social pressure, or alternatively as a strategy to place a community as a point of contact between diverse environmental zones and cultural groups.

I use the material correlates of the actions of everyday life—plant foodways and ceramic technology—from feature contexts to reconstruct how people in Tuckaleechee

Cove made different choices that were a result of particular social and cultural settings

(Wills 2009:286). Foodways are the collective daily practices involved in production, procurement, preparation and processing, storage, cooking, presentation, and consumption (Atalay and Hastorf 2006; Johannessen 1993). I also use foodways to address the different ways that groups of people at this location may have negotiated

Mississippian cultural change, whether by conforming to, adapting to, tolerating, or disputing social pressures. Considering a level of warfare and violence present across much of the Southeast that altered daily subsistence strategies (VanDerwarker and

Wilson 2016), I argue for the need to continue to investigate how foodways in the upland areas of the Southern Appalachians can reflect the social and political dynamics between this and nearby regions.

An additional way to contextualize the differences in foodways at Townsend is as a marker of group identity. This approach has been successfully applied archaeologically from early prehistory to colonial times to examine group interrelations particularly during times of culture contact, both novel and sustained (e.g. Bardolph

2014; Clark 2001; Wilson et al. 2017; Voss 2005). Because foodways are a conservative

aspect of culture, in situations of cultural contact they are less likely to change as rapidly as other forms of more ostentatious material culture (Clark 2001). The material record of foodways, in plants and ceramics, are created by the recursive acts of individuals, and thus by the choices made as individuals perform daily routines. These daily practices often conform to the norms of one’s respective society, regardless of whether one belongs to the dominant or subordinate political or economic group (Scott 2008).

Because foodways are linked to identity, they have interpretive potential and serve as a reflection of who was or wanted to be considered as a member or outsider. For these reasons, foodways can be used as a marker of ethnic and cultural identity (Scott 2008).

Because of the high potential for culture contact at the Townsend sites between two cultural groups, it is constructive to consider differences in foodways as a reflection of group identity and cultural affiliation.

The occupation of the Townsend sites during the Mississippian period is represented by 18 multiple households and multiple settlement types consisting of eight farmsteads and two villages. The diverse settlements at the Townsend sites allows me to examine variability between the different residents that occupied the Cove for over

400 years. I employ household archaeology because the household is the place where foodways and aspects of identity are practiced daily, both conformed to and challenged.

Therefore, a comparison of the differences in foodways between households and settlements at Townsend can inform us about the individual properties of households

such as social status or domestic composition, along with the larger environmental and socio-political contexts.

I address several aspects of the Townsend sites, including the ways plants used in foodways and feature data contribute to an understanding of site use; the degree to which the residents were invested in Mississippian foodways traditions; how the type of settlement lived in, farmstead or village, influenced the Mississippian experience; how differences in foodways reflect distinctive cultural traditions; and whether status was expressed through plant foods and the activities associated with them, or people used material culture to minimize differences. While access to particular animal resources or cuts of meat can also inform us about Mississippian cultural traditions and expressions of status (Scott 2008; VanDerwarker 2009), I did not include them in my analysis because of their poor preservation at the Townsend sites. My work places a high priority on acknowledging and highlighting the contributions of women through decisions surrounding foodways to the creation of a Mississippian lifeways and economy.

I present my research in eight chapters. Chapter 2 is an examination of the traditional construction of Mississippian culture from archaeological research in the

Southeast, followed by an introduction to the new approaches that are being used to challenge a monolithic view of Mississippian peoples and cultures. I also detail the theoretical framework I use to interpret the material remains at the study sites. Chapters

3 and 4 provide an overview of the sites, both the natural and cultural landscape of the study area, Tuckaleechee Cove, and a background to the site excavation, laboratory

analysis, and previous interpretations of the site complex. In Chapter 4, I also introduce my organization of space at the sites, describing my use of terminology for settlement types and household constructions that I employ in the interpretations presented in

Chapters 5, 6, and 7.

I present my analysis of feature contexts from the Townsend sites in Chapter 5. I categorize the many features across the sites to interpret their use. I also use feature quantity, distance to household, and storage capacity across households to evaluate status and social differences, as well as to reveal larger shifts in foodways practices.

Chapters 6 and 7 present the results of the plant foodways analysis. I emphasize how plants were integral to the daily and seasonal patterns of life in Tuckaleechee Cove.

Using a comparative approach I explore how variation in plant use between the different households and settlements at the sites contribute to an understanding of the distinct choices people made to address unique circumstances. In Chapter 7 I integrate the plant and ceramic remains using correspondence analysis to try to determine differences in the assemblages of households and settlements that are otherwise unclear in independent data analysis.

In Chapter 8, I summarize my interpretations of the Townsend sites. I point to how plant foodways and the activities surrounding them can signal group ideology, both through shared traditions and potential differences. I emphasize the need to continue to investigate the social dynamics between the settlements in the upland areas of the Southern Appalachians and larger site complexes to the west and east. I also

identify how this single complex of Mississippian sites demonstrate the large amount of variation that was present in Mississippian societies at large.

Included within the appendices are the original classifications of pit features, domestic architecture, community buildings, corncrib structures, and primary enclosures. A detailed table of plant remains identified from individual feature contexts analyzed in this dissertation is also presented in this section, along with individual maps of each of the households and associated features that I employed for comparative purposes throughout this study. Original field excavation site forms and drawings are located at the Archaeological Research Laboratory (ARL) of the University of

Tennessee’s Anthropology Department. The data obtained in this project will be archived digitally at the University of Tennessee and all paleoethnobotanical samples are curated at the ARL’s curation facility at the University of Tennessee.

CHAPTER 2

ARCHAEOLOGICAL PERSPECTIVES ON MISSISSIPPIAN IDENTITY: WHAT

DOES IT MEAN TO BE MISSISSIPPIAN?

2.1 Introduction

Archaeologists have traditionally defined Mississippian period cultures, dating from approximately 900–1500 CE in the Southeast, by a set of common characteristics: chiefdom-level political organization, maize-based subsistence, shell-tempered pottery, dense population centers, and monumental architecture. The Mississippian period, with highly visible and aesthetically appealing trademarks such as platform mounds, elaborate grave goods, and rich iconography, has also been a centerpiece of Southeastern archaeology (e.g. Anderson and Sassaman 2012; Pauketat 2007, 2013). The collective body of work has traditionally defined this culture based on this material expression in terms of complexity, social hierarchy, and a shared identity founded on an elaborate ceremonial complex. A large amount of social change in the Mississippian period was likely the result of the instability caused by shifting hierarchies and chiefly cycling

(Anderson 1994, 1996a; Blitz 2009; Gavrilets et al. 2010).

However, recent work on the Mississippian period in the Southeast has challenged the sustainability of the traditional Mississippian characteristics across time and space and instead has successfully used archaeological approaches at smaller scales of analysis, that of the household and of specific sites, to understand how differences within and across communities would have impacted lifeways (Alt 2001; Boudreaux

2013; Clay 2006; Emerson 1997; Maxham 2000; Mehrer 1995; Scarry and Scarry 2005;

Thomas 2001; Wilson 2008, 2010). A growing body of research challenges both the singular idea of a Mississippian identity and the validity of models that depend on the 10

political control of the elite to produce change (Pauketat 2007:36). Sites across the

Southeast, and particularly in peripheral regions such as the Southern Appalachians, do not fit comfortably into previous theoretical models (King and Meyers 2002; Meyers

2002; Sullivan and Rodning 2011; Jefferies 1996; Clay 2006). Consequentially, models of the Mississippian period have become increasingly complex to account for the large amount of diversity across communities.

This diversity is particularly true of Mississippian period sites in the Southern

Appalachian region. This area of the Southeast is often considered a Mississippian frontier, a geographic area where Mississippian societies were in the process of forming by incorporating and/or changing Mississippian ideology (King and Meyers 2002). As a result of geographic location positioned between dominant cultural traditions, frontier communities often have some mainstream Mississippian characteristics, but their expression may be altered, likely shaped to meet different needs and obligations

(Meyers 2002).

I begin this chapter with a discussion of how traditional archaeological approaches have constructed Mississippian identity in the Southeast. I follow with a report on the current state of Mississippian period archaeology, focusing on the contribution of research that has built an alternative narrative of Mississippian identity, especially regarding geographic areas that are rural, in upland locations, or on the frontier. I end the chapter with a discussion of the theoretical approaches I use to frame the material remains from the Townsend sites during the Mississippian period.

2.2 A Construction of Mississippian Identity in the Southeast: From the

Traditional to New Approaches

I guide my discussion by exploring three pervasive questions that have traditionally shaped the construction of Mississippian identity, the way that a community defines who is and who is not a member, and that are being explored in new ways to challenge that construction: 1) What is the origin of Mississippian identity? 2)

Where do sites fit within the chiefdom model: how do we define complexity? 3) What is the source of power and inequality and how is it expressed?

2.2.1 The Mississippian Emergence

Research into the origin of Mississippian culture is an important avenue to answer fundamental questions about the processes of cultural change. In the period beginning around 900 CE, societies in the Southeast underwent fundamental transformations that resulted in what most scholars agree was a chiefdom-level society.

However, more recently the concept of chiefdom has been called into question for its neoevolutionary terminology and its overgeneralization of social organization (Pauketat

2007). Thus, archaeologists have been interested in several facets of this transition, the reasons for this change, the ways in which culture was disseminated, and the products of that change, particularly related to the development of hereditary inequality. Two main theories have been proposed to explain the emergence of Mississippian culture from the Late Woodland period: local adaptation and external influence. 12

The local adaptation model was championed by Smith (1984, 1990), who proposed that Late Woodland cultures developed into Mississippian chiefdoms as a response to local environmental conditions. Smith suggested that Woodland communities across the Southeast repeatedly faced similar challenges: their growing populations were increasingly confined in resource-rich zones with high agricultural productivity that escalated competition and the responsibility to protect boundaries

(Smith 2007). The Mississippian cultural endeavor in various regions were similar adaptive responses to analogous circumstances.

The more recent, and more widely accepted model of Mississippian emergence places the prehistoric city of Cahokia as the birthplace of Mississippian culture (Pauketat

2004a, 2005, 2007, 2008). Pauketat (2004a 2007) argued that Mississippian identity spread materially through characteristics such as wall-trench architecture, pottery styles, and chunkey stones among others, and ideologically as societies accepted a new, different way of life. The “Cahokia-centric” (Blitz 2010:12) model of Mississippian origins points to a variety of historical circumstances in particular areas that led to

Mississippian acculturation, communities coming together, coming apart, people interacting and migrating in what is being promoted as an Oikumene, or web of mutually known polities, of pan-North America (Peregrine and Lekson 2012).

Whereas Smith’s adaptive response model can be critiqued for reducing local communities to a series of passive responses to increasing environmental and cultural pressure, Pauketat (2003) emphasized active culture creation that is the product of both

local and external interactions and decisions. While the ingenuity and uniqueness of

Cahokia is evident from its monumental architecture, settlement hierarchy, population size, craft production, and organizational diversity (Blitz 2010:11), most research does not point to complete replacement of Late Woodland culture from outward migrating groups, but instead to negotiations that involved local communities choosing, rejecting, and reshaping Mississippian ideology into practicable and comfortable traditions.

In eastern Tennessee, Mississippian origins have been explained by three proposed models that fall generally in line with that proposed for the greater Southeast:

1) ethnic replacement (Lewis and Kneberg 1946); 2) cultural diffusion (Faulkner 1972,

1975); and 3) in situ cultural development (Boyd et al. 1983; Helmkamp 1985; Schroedl

1978b; Schroedl and Boyd 1985; Schroedl et al. 1985). Schroedl et al. (1990) proposed that internal change is widely supported by the continuity of Late Woodland practices into the Mississippian period, such as continued use of triangular points and limestone- tempered plain and cordmarked ceramics, the Hamilton burial mound complex, and single-post wall and wall-trench houses. Smith (1984) also suggested that without a source of cultural influence or evidence of chronological radiation, in situ development was a more reasonable explanation of change. However, refined chronologies that have helped to track the spread of Mississippian traits indicate both an origin in the site of

Cahokia, as well as radiation through emulation or immigration (Anderson and

Sassaman 2012).

2.2.2 Where Do Sites Fit Within the Chiefdom Model: How to Define Complexity?

A major focus of Mississippian research has been to define the overall complexity of Mississippian period societies, and to situating individual sites into a settlement hierarchy. Cobb (2003) suggested that Southeastern archaeologists have used both the fluorescence of material culture and sixteenth-century ethnohistoric information from the region to build a model of Mississippian identity that leans primarily on the chiefdom concept and systems of political complexity. He argued that while there has been no clear definition of complexity, Mississippian chiefdoms are usually ordered from simple to complex or paramount based on their scale, as defined by physical size, geographic reach, and population size; their position within a settlement hierarchy as defined by mound complexes, degree of craft specialization and labor investment per large-scale projects like mound or palisade construction; and elaboration of burial treatment and iconographic representation. Simple chiefdoms are further defined as only consisting of one tier of loosely applied political leadership, while complex chiefdoms can contain two-to-three-tiered hierarchies exhibiting a tightly controlled class structure and redistributive power (Steponaitis 1978:420–421).

The simple-complex settlement chiefdom model (Anderson 1994) has been very influential in explanations of relationships between primary and secondary mound centers in the Southeast. But Blitz (1999) argued that Hally’s (1993, 1996) review of platform-mound sites across the South Appalachian area proves that the simple- complex model cannot account for the spatial variation in Mississippian settlements and 15

thus cannot explain the complexity of each or relationships between them. Blitz alternatively suggested that mound quantity, placement, and polity size cannot always be associated with a chain of hierarchical administrative centers, but instead may be related to processes of dispersal and concentration within a geographic area.

Despite the attractiveness of a corresponding site-size-to-status relationship, much unexplained variation is present in Mississippian settlement organization and relationships. Cobb (2003) suggested that due to a lack of large-scale surveys, we are only getting a small glimpse of the variation that was likely present temporally and spatially in this period. He pointed to two areas of research that impact our appreciation for the diversity of the period: work on rural areas that are lumped together categorically as small village and hamlet sites, and the study of peripherally located sites in the hinterland. Sites located along frontiers, on the boundary between two cultural traditions, in upland or rural areas, or that are smaller have been at the forefront of research that challenges traditional constructions of Mississippian identity.

While they display some undisputable Mississippian traits, they often do not fit well into the tiered settlement hierarchy and display a lot of social and economic variability.

For example, Maxham (2000) used the ceramic vessel assemblage of a site in the rural Black Warrior Valley to counter the idea that complex activities such as feasting and ritualistic consumption only occurred at political mound centers. Even in rural areas, commoners participated in ritualistic ceremonialism that was separate from elite interactions. A similar picture of diversity and social complexity emerges at upland

settlements. People in Cahokia’s uplands of the Richland Complex seem to have adopted cultural aspects at different rates and in different ways that allowed them to retain traditional aspects of their community (Alt 2002). In addition to actively participating in processes of cultural transmission, upland communities that were positioned in resource-rich zones were also potential economic forces that were intertwined with the rise and perpetuation of floodplain communities.

Similarly, rather than forcing hinterland settlements into the settlement hierarchy for the purpose of solidifying their Mississippian identities, Southeastern archaeologists are increasingly focusing on historical trajectories to explain local diversity (Clay 2006).

A good example of the work that is emphasizing Mississippian cultural heterogeneity has been conducted in parts of the Southern Appalachians including sites in southwest

Virginia (King and Meyers 2002; Meyers 2002), western North Carolina (Beck and Moore

2002), and southeastern Kentucky (Jefferies et al. 1996). For instance, Meyers (2008) argued for the presence of chiefdoms in southwestern Virginia during the fourteenth century despite the lack of burial stratification or intensive corn agriculture. Instead of these traditional markers of economic control, she suggested that control over salt production and trade drove economic relationships and social distinctions in this frontier location. Likewise, Jefferies et al. (1996) concluded that variation in foodways, particularly a reduced emphasis on corn in favor of nut in resources, was present at the peripheral site of Croley-Evans in the Cumberland Valley of southeastern Kentucky.

Although residents of this simple chiefdom made shell-tempered pottery, constructed

platform mounds, were involved in exotic trade, and grew some corn, they continued to use nuts as a primary food source and did not invest in many pottery forms or ceramic decorations. For these reasons, Jefferies et al. (1996) suggested that occupants of this site modified their subsistence strategy to meet local economic and social demands.

Differences in the production and expression of power along frontier regions can also be archaeologically variable. In the Malpaso Valley of Mexico, the expression of social status in the prehispanic Mesoamerican frontier does not appear to have been based exclusively on the control of resources as it was defined in the core or mainstream settlements (Turkon 2004). In the absence of prestige items, elaborate burials, and monumental architecture, Turkon (2004) used the differential distribution of food- related correlates to determine variation in status, and suggested that domestic activities such as the preparation and serving of food can be important alternatively used material markers to signify or display social difference.

Whether a community can truly be defined as a simple or complex chiefdom, a primary or secondary center, or as politically centralized or decentralized remains unresolved. The lack of a platform mound or the lack of detectable differences in status does not necessarily equate with reduced social complexity. Furthermore, the lack of these features may not reduce a site to a status subordinate or subsidiary to sites that have these characteristics. While archaeologists have generally used settlement size, the number of mounds, plaza size, and the number of high status burials as an indicator of

the level of social organization of a site, it is clearly difficult to determine the relationship of multiple sites to one another (Clay 2006; Wood 2014).

2.2.3 What is the source of power and inequality and how is it expressed?

Archaeologists often rely upon the presence of exotic goods, rare resources, and other prestige items to identify markers of status, social inequality, and power in

Mississippian communities. Because Mississippian societies have traditionally been classified as centralized chiefdoms, power has routinely been associated with elite control over the political and economic sphere (Blitz 2009:4). These models generally emphasize control over material goods, production, and labor; this control is archaeologically visible in the management and redistribution of valuable artifacts and food, in surplus production and tribute payments, and in construction efforts such as palisade and mound building.

Indeed, while status goods can be useful indicators of social complexity in the

Mississippian period (Anderson 1994:196; Welch 1991; Welch and Scarry 1995), especially in mound contexts at very centralized sites such as Cahokia, Moundville, and

Etowah, more unrecognizable forms of ranking such as by gender, age, and/or faction may require a reconceptualization of social order (Brumfiel 1995). A significant body of research employs the concept of heterarchy to examine social relationships, power, and inequality (Brumfiel 1995; Crumley 1995; Maxham 2000; Sullivan and Rodning 2001).

One advantage of an exploration of horizontal relationships is that often-overlooked

channels of power—frequently those associated with the household and women—are recognized. Sullivan and Rodning (2001, 2011) used a heterarchical approach to examine how gender relationships were expressed in burial practices in the Southern

Appalachians. They argued that the separate burial spaces of men and women from the thirteenth to the seventeenth centuries, with burials of mostly women near the domestic space and burials of men in the public domain, is a spatial reflection of gender duality.

Rather than assign a marginalized role to women because they are not buried publically,

Sullivan and Rodning suggested that residential burial of women is an indication of their kin-based authority over the household.

Heterarchical models have also shifted exclusive power and control away from political elites and acknowledged the power and influence of communities and kinship.

In the absence of strong material evidence to suggest extensive elite control over resources from many sites across the Southeast, power is being defined in terms of ideological control (Blitz 2009). For example, Levy (1995) demonstrated that ritual ceremony and esoteric knowledge can drive complexity in the absence of elite economic control. At the Town Creek site in western North Carolina, the eventual residential abandonment of the community gave way to the location of a kin-centered ceremonial/burial precinct (Boudreaux 2013). The occupational history of this site suggests that social relationships were complex, in that social groups within a ritual context governed power relations.

2.3 Situating Mississippian Foodways at the Townsend Sites

Despite the growing body of work that has challenged previous notions of homogeneity in Mississippian identity, there remains a tendency to view communities that do not have mound construction, do not appear to have a large degree of social hierarchy, or that are geographically remote as problematic. Because of several important aspects of the Townsend sites—their smaller size and lack of a mound, frontier location between two environmental and cultural areas, multiple settlement types, multiple households, and occupation over a period of roughly 400 years—they have much to contribute to the narrative and construction of Mississippian lifeways in

Southeastern archaeology.

I approach the foodways at the Townsend sites by using an agent-oriented perspective at the scale of the household. The Townsend sites’ proximity to Hiwassee

Island sites to the west in East Tennessee and Pisgah sites to the east in western North

Carolina undoubtedly influenced and contributed to their own culture making. The larger cultural landscape of the Southeast was complex and the interrelationships between the large Mississippian centers such as Cahokia and rural areas cannot be disregarded (Clay 2006). I view the variability within the plant assemblage, in the routines of everyday life, as products of dynamic interactions of people on the local and regional scale.

2.3.1 Agent-Oriented Archaeology and Foodways

The advantage of using an agent-oriented perspective is that it interprets the archaeological record as the result of individual decisions (Barfield 1997). Gender, class, and faction are all recognized as important aspects of an individual’s status impacting social negotiation, and thus social change (Brumfiel 1992; Rodning and Eastman 2001:3).

This approach has been suggested as an alternative to the ecosystems approach that views cultures as systems and cultural change as an adaptive response. The ecosystems or processualist approach is limiting because it tends to hide the activities of certain sets of social actors such as women, children, and less dominant communities; reduces the role that social categories play in producing social change; and compels us to search for external rather than internal causes of social change (Brumfiel 1992; Twiss 2007).

A variety of factors contribute to the agency of individuals within a household or community, and an individual’s social role is capable of changing both by situation and through time. In the Southern Appalachians, the social role of women likely shifted depending on whether they were in a household setting, which was a seat of kin-based women’s power, or in the public realm, where they took a more passive role (Sullivan and Rodning 2011). People take on various roles throughout their lives; social roles change over time, for example with age or marital status.

Using an agent-oriented perspective, I approach the feature, plant, and ceramic data in this dissertation with the understanding that people do not make choices or organize their space haphazardly. For example, when we discuss food in the 22

archaeological record we must not only consider dietary requirements (e.g. Lupton

1996), but we must also discuss how people, depending on their gender, class, and faction, negotiated their social positions around subsistence activities and how this negotiation is then reflected in the material record.

Plant foods are integral components of daily life. People select food for consumption, decide whom to share a meal with, and choose how to consume that meal

(Hastorf 2012). Many of these choices and activities are subconscious practices that have been learned and enforced from childhood and thus, they are folded into our sense of self (Atalay and Hastorf 2006; Hastorf 2012:68). Because people participate in cultural traditions associated with food and habitually perform these practices, food is a social fact that creates and reflects cultural identity (Atalay and Hastorf 2006:284; Douglas

1972:61; Twiss 2007). Prehistoric people had to negotiate both ecological and social constraints associated with the production, preparation, consumption, and storage of plant resources. Viewed in this way, the archaeological record—the plant remains, the organization of activities around them (e.g. gathering, processing, consumption, and storage), the ceramic technology used in association, and the decisions surrounding the selection of material and the stylistic production of those ceramics—is a proxy for understanding how individuals experienced the past.

Employing an agent-oriented approach allows me to recognize the contributions of the hidden actors, particularly women, in the past by assigning rather than ignoring the specific activities associated with their gender (Brumfiel 1992). It is probable that

community members other than women (men, elders, and children) participated in field preparation and maintenance activities. Nonetheless, I proceed with the assumption that women were responsible for the organization and labor associated with farming activities, the processing of plant resources, and the methods and practices of consumption, as is abundantly evident from the ethnohistoric data in the eastern U.S.

(Bridges 1989; Emerson 1989:52–56; Hann 1986; Hudson 1976; Prentice 1986; Scarry and

Scarry 2005; Swanton 1946:306, 710–717; Thomas 2001: 32; Wenhold 1936:13). By recognizing women’s contributions to the subsistence base, I address how women at the

Townsend sites contributed to the political economy of their Mississippian society and how potential shifts in foodways practices may have impacted economic and social organization.

2.3.2 The Benefits of the Household as a Comparative Lens at the Townsend Sites

The social dynamics within and between households have been of much interest to Southeastern archaeologists. The reasons for this are twofold: the household follows an architectural grammar that has been viewed as a mirror of the larger Mississippian community organization (Muller 1997; Wilson 2008), but it is also an entrance into the everyday lives of people in the past (Pluckhahn 2010:332). The household is a constructive scale of analysis for several reasons. First, archaeologists have been able to identify households as distinct groupings of structures and associated features within sites. Second, the household, while fluid and dynamic, is a bounded social, economic,

and biological unit (Rogers 1985). Third, because the household unit is the largest contributor to agrarian economic production, political change begins in this arena

(Brumfiel 1992:555). For these reasons, by comparing distinct households, archaeologists can identify differences across households in composition, status, and occupation, as well as ideology (Pluckhahn 2010). Households reflect the accepted norms of society, thus providing us with one avenue to discover ethnic associations or social divisions/factions within a community (Rogers 1985:19). Additionally, domestic architecture size, extravagance, and associated special function structures, along with artifacts of greater quality, quantity, and diversity, can be good indications of a higher- status household (Hirth 1993:24–125). However, this is not always the case, and in some instances, such as the Moundville I phase in the Black Warrior Valley, status and wealth between households were minimized in daily life, even though they were elevated in ceremonialism (Wilson 2008:129–130).

Exactly what constitutes a household has been thoroughly debated in archaeology. Many archaeological approaches, especially in the prehistoric Southeast, have elected to use a narrow definition that links the household specifically to the collective activities that are bound to a domestic dwelling (e.g. Nash 2009; Wesson 2008).

Alternatively, some scholars (e.g. Ashmore and Wilk 1988; Wilk and Netting 1984;

Pluckhahn 2010) define the household based on the collective activities performed by a group of people such as production, consumption, reproduction, co-residence, and transmission. This definition is productive because of its flexibility—while co-residence

may be practiced, it is not the only recognized living arrangement (Pluckhahn 2010). For example, these activities may be performed by multiple family households. The household is defined here as the minimal, co-residential social unit in society (Allison

1999; Ashmore and Wilk 1988; Franklin 2004; Hirth 1993; Nash 2009; Robin 2003; Wilson

2008:999). I find it useful to extend this definition to include the household as a unit where cooperative social and economic production is based from a domestic context

(Mrozowski et al. 2008). However, defining the discrete households in the Mississippian period at the Townsend sites is conditional on determining the components that constitute a Mississippian household. A household is more than a domestic structure; that is, a household is a social spatial unit where a variety of domestic activities are performed. Mississippian households and communities across the Southeast followed an architectural grammar that was likely a mirror of the organization of social or kin groups (Wilson 2008:12).

Wilson (2008) archaeologically defined the Mississippian households at

Moundville as consisting of domestic structure/structures (e.g. a summer and winter house), and a combination of storage facilities, trash deposits, and work and ceremonial spaces that are separated from another household cluster by roughly 20 to 40 meters. I find a similar approach useful in eastern Tennessee as the excavation of substantial

Mississippian communities indicates that household units were organized in a similar way, as single or multiple household clusters that utilized a winter and summer house, an open space, and a variety of storage facilities (Polhemus 1987; Schroedl 1998; Sullivan

1995). There does appear to be pronounced differences over time in household organization in East Tennessee, from the Dallas, to Mouse Creek, to Overhill Cherokee phase in house size and storage patterns that are tied to shifting social relationships in the region. There is also much variability in household patterns between sites that is related to the spatial setting, size of the population, occupational history, and environmental setting of individual sites (Rogers 1995). For example, while the 20- to -

40 meter distance threshold between households has been a useful marker of space at many sites across the Southeast and in Central America (Wilson 2008), in the American

Bottom alone, there is a lot of variation in this pattern. Smaller farmsteads usually consisted of a single household cluster with several structures (Finney 1985; Jackson

1980; Mehrer 1995; Milner 1983) and at Cahokia there was a continuous distribution of houses indicative of multihousehold residence (Pauketat 1998).

At the Townsend sites of 40BT90 and 40BT91 there are 18 distinct household compounds believed to have been occupied by separate households, which appear as single or paired structures and their associated features as defined based on distance to structure. I extend the theoretical lens of household archaeology at the Townsend sites by also comparing the multiple settlements at the sites under this analytical umbrella.

Across 40BT90 and 40BT91, I define and compare eight farmsteads and two village settlements. I argue that it is useful to view the multiple settlements at the sites in a similar way as households, because they are distinct from one another and were likely composed of a group of people with presumably similar ideologies that were living with

similar opportunities, challenges, and constraints (Nash 2009:218). By comparing the plant remains of the distinct household units at Townsend, I will attempt to identify the different strategies that households employed that were potentially related to their composition, status, and occupation duration, as well as ideology.

2.4 Summary

Our understandings of culture are often constructed based on the beliefs, technology, and other social expressions of the more visible and the dominant in society.

These reconstructions are thus fragmented and subject to bias and misrepresentation.

The complexity of the Mississippian period requires a holistic approach that recognizes the contributions of diverse types of social expressions and social actors.

In the following chapters, I use this research at Townsend to create a more holistic view of the Mississippian period by including the often-overlooked perspective into the canon—that of smaller, upland communities. I use the Townsend sites, taking advantage of one of the few extensive excavations that has been conducted in the

Southern Appalachian region of East Tennessee, to address how residents at this location actively produced their own culture, by negotiating, conforming to, tolerating, or disputing social pressure. I examine the variability in foodways within the sites and between the different households and settlements and situate it relative to other small peripherally located sites, such as 40BT47, Croley-Evans, and Carter-Evans, and larger, more centralized settlements nearby in eastern Tennessee and western North Carolina

The growing challenge to a monolithic Mississippian identity within

Southeastern archaeology calls for research that contributes to a more nuanced understanding of social organization and political economy in this period. An agent- centered theoretical approach implemented at the scale of the household puts everyday decisions at the nucleus Mississippian social life. By using material correlates of the actions of everyday life—food remains—I reconstruct how geographic place contributed to the formation of a community and the households within it. Interpretations from the

Townsend sites lead to a better understanding of how people in frontier, upland locations in the Mississippian Southeast responded to outside pressures, while creating vibrant and distinct cultural traditions. In the following chapter I describe how the natural and cultural landscape contributed to community formation in Tuckaleechee

Cove.

CHAPTER 3

THE NATURAL AND CULTURAL LANDSCAPE OF TUCKALEECHEE

COVE: THE TOWNSEND ARCHAEOLOGICAL SITES IN CONTEXT

3.1 Introduction

The Townsend archaeological sites exist in a unique geographic position: along the physical boundary of two physiographic provinces, the Ridge and Valley and Blue

Ridge, and between two cultural units in the Mississippian period—the Tennessee

Valley and Pisgah. An in-depth consideration of the archaeological landscape can give us a context to understand how place affected people’s daily decisions, whether they were driven by economic, political, or social motivations (Knapp and Ashmore 1999).

The geographic location of the Townsend site complex provided many advantages for people living in the area. Access to multiple environmental zones granted opportunities to use many different resources including raw materials and diverse plant and animal communities; and easy access to a waterway, the Little River, provided transportation as well as fertile bottomlands for farming. What is less clear is whether the seemingly isolated location of the sites, the occupation of a space between two cultures, was culturally limiting or whether it created special opportunities for residents to act as an interface between multiple groups, potentially absorbing and transforming useful information and ideas to create a distinct cultural tradition.

In this chapter, I describe the natural and cultural environment of the Townsend sites to explore the impact of place on the Mississippian people that settled there.

Attitudes have shifted over time about the degree to which the environment shapes societies, from passive to deterministic to holistic, but the environment continues to serve as an explanatory framework for cultural development and change across the 31

social and environmental sciences (Kowalewski 1995:147; Moran and Brandizio 2013;

Willey and Sabloff 1993). Viewed holistically, human-environment interactions are the product of reciprocal interactions within the ecosystem between human populations in a historical context, and all parts of the natural environment, the geology, climate, flora, and fauna (Butzer 1982; Waters 1992). This chapter places the Townsend sites in their natural context, followed by a review of the cultural history of the area. I emphasize both how people responded to shifts in the natural environment and how people actively created and shaped their surrounding natural and cultural landscapes.

3.2 The Natural Landscape

3.2.1 Physiography

The same landscape features that attract modern-day people to visit to

Townsend in Blount County, Tennessee, are likely the aspects of the natural environment that made this location a special place for native peoples. The Townsend sites are located along the Little River in Tuckaleechee Cove, which is situated at the foothills of the Smoky Mountains in eastern Tennessee. Tuckaleechee Cove is one of several coves in the Blue Ridge province, including Cades Cove, Miller Cove, and Wear

Cove; areas of gentler, rolling topography and deep, fertile alluvial soils that are set against the backdrop of the high peaks of the southern section of the Appalachian

Mountains (Neuman and Nelson 1965:D3). Tuckaleechee Cove lies within the Level IV

Subecoregion of the Limestone Valleys and Coves, which is defined as a small but distinct lowland area created by erosional forces that have worn down overlaying deposits, resulting in surface deposits of Ordovician-age limestones (Griffith et al. 1997).

These coves might have been places of refuge for Mississippian peoples, hospitable pockets of livable and arable land that undoubtedly were viewed as a valued resource, especially in the Southeast where farmable land may have been in great demand for communities relying on maize agriculture.

The site complex is located entirely within the southern section of the Blue Ridge province of eastern Tennessee (Fenneman 1938, 1946). This area is characterized by a topography dominated by ridges and highlands that separate the rolling plateau of the

Piedmont to the east and the long, narrow crests and valleys of the Ridge and Valley

Province to the west (Fenneman 1938:174). The Great Smoky Mountains, a subrange of the Appalachians, dominate the portion of the Blue Ridge province surrounding

Tuckaleechee Cove. Based on topographic relief and underlying geology and drainage patterns, the Great Smoky Mountains can be divided topographically into the foothills, the main range or uplands, and the coves (Neuman and Nelson 1965:D3; Southworth et al. 2003; Southworth et al. 2005). The foothills have a smoother topographic profile than the uplands, descending from an elevation of 1524 to 1829 m at maximum heights along the crest in the main range to an elevation of approximately 93 m near the Little

Tennessee River (Neuman and Nelson 1965:D3).

3.2.2 Hydrology of the Little River

The archaeological site area lies along a 5-km stretch of the Little River, a 97-km northwest-flowing tributary within the Little River Drainage District. The Little River begins at high elevation near Clingmans Dome in the Great Smoky Mountains National

Park at an elevation of approximately 2025 m and acts as a drainage system for much of the rugged terrain of the mountains (Mast and Turk 1999) (Figure 3.1). The river drops significantly in elevation to 250 m above sea level as it descends and empties into the

Tennessee River.

The river is perennial, but the flow depends on several factors including precipitation patterns, topography, and underlying geology. With abundant rainfall, mean monthly discharge ranges from 3.5 m3/s in the fall season to 14.8 m3/s during the peak flow season in March (Mast and Turk 1999). Stream gradients for the Little River span from 16 m/km in downstream locations to 110 m/km in the steeper headwater tributaries (Mast and Turk 1999). In effect, navigation of the river by watercraft ranges from not navigable in low-water areas, challenging due to high discharge in areas where streams merge such as the Three Forks area upstream from Townsend, and easily managed in areas such as Tuckaleechee Cove that are dominated by moderate flow.

Land use along the Little River Watershed is predicated on location within three distinct geographic subdivisions: upper, middle, and lower (Ezzell et al. 2005). The location of the Townsend sites is on the boundary between the upper portion of the

Figure 3.1. Location of the Townsend archaeological sites along the Little River.

watershed, which is currently dominated by mixed forest, and the middle portion, which is predominantly used for agricultural purposes (USDA 2011).

3.2.3 Geology and Soil Morphology

The geology of Tuckaleechee Cove is unique to the surrounding Blue Ridge

Mountains. The cove’s diverse bedrock and surficial deposits were caused by the erosion of surface deposits that resulted in the exposure of Ordovician-age rocks

(Neuman and Nelson 1965). Thus, the geology of the Cove is more comparable to the

Ridge and Valley deposits to the west than to the Blue Ridge Mountains to the east.

The underlying geology of Tuckaleechee Cove and the surrounding area directly affects the availability of raw lithic materials, as well as the quality of the soil. The location of the Townsend sites in a geologically diverse landscape gave the occupants access to a wide range of raw stone material. A comparison of the Townsend sites’ lithic assemblage and a survey of stone sources in the Little River drainage indicated that most of the raw material for everyday stone implements was acquired from local sources—most commonly, the Little River (Sweat 2010). For example, Sweat (2010:145) determined that local materials made up 98 to 100 percent of the Late

Woodland/Mississippian and Late Mississippian/Historic categories of projectile points, with Knox chert being the most prominently used material. Furthermore, 91 percent of chipped and groundstone artifacts at Townsend derived from a local source. The geologic deposits in Tuckaleechee Cove, along the Little River Drainage, and in many of

the tributaries of the Little Tennessee River, provided access to a wide range of raw materials including sandstone, slate, quartzite, quartz, and siltstone that were used to make utilitarian/expedient tools such as percussion, grinding, and pulverizing instruments, and decorative and ritualistic items such as gorgets, celts, and pipes (Sweat

2010). There was no immediate access in Tuckaleechee Cove to high-quality Knox chert or steatite, but sources were located to the northwest and east and materials could have been easily transported along the Little River (Sweat 2010:193).

The soil and geomorphology of Tuckaleechee Cove are critical to understanding a variety of factors at the Townsend sites. A reconstruction of soils is a primary way to understand the environmental and depositional histories of prehistoric landforms, which can then be used, first, to evaluate how soil formation and erosion contributed to the preservation and visibility of the archaeological record, and second, to determine the human occupation or cultural history of the landscape (Windingstad et al. 2011). The

Townsend archaeological sites are composed of three landforms: the current floodplain

(T0), the lower terrace (T1), and the upper terrace (T2) (Figure 3.2). The visibility of the archaeological deposits at the sites is a general result of the northward migration of the

Little River, which resulted in a natural process of erosion in areas north of the channel and deep sedimentation and burial in southerly areas (Windingstad et al. 2011:145). In effect, contemporary Late Archaic deposits had variable depths of burial, with deposits occurring superficially up to approximately 1 m in depth. Further

Figure 3.2. Landforms of the Townsend sites (Windingstad et al. 2011:Figure 8).

complicating stratigraphic integrity, historic sedimentation and plowing activity in the site area has leveled topographic relief and disrupted contexts.

There is great diversity of soil types in Tuckaleechee Cove. Alluvial soils can exhibit wide variation due to the complexity of deposition and sedimentation that is a product of fluvial activity (Ferring 1992; Windingstad et al. 2011). However, all the soil types that are predominant across the Townsend site area are suitable for farming. The

Townsend sites cross several soil series (Table 3.1) with the largest portion of the sites’ area (79.8 percent) existing within the Sequatchie loam (Sc), Sequatchie silt loam (Sd),

Staser fine sandy loam (Sh), and Staser silt loam (Sl) soil series, all classified as “prime farmland” (USDA NRCS 2017). The Sh and Sl series are also very deep and well- drained, a result of alluvial, floodplain formation processes (United States Department of Agriculture National Resources Conservation Service 2017).

3.2.4 Climate

The detailed studies of the climate records of the Great Smoky Mountains

(Fridley 2009; Shanks 1953; Stephens 1969; Whittaker 1956) have resulted in a good understanding of the factors that control weather patterns, and have also produced long-term reconstructions of climate histories in the region. These records are invaluable for understanding how weather patterns and changes in climate

Table 3.1. Dominant Soil Series in the Townsend Archaeological Site Area.

Internal Frequency Farmland Soils A Material Landform Drainage of Flooding Relief Slope Acid Classification A Suitability

Short rotations Sequatchie Mixed Stream Well None Level 1–3 % Medium Prime Farmland or intensive loam Alluvium terraces drained use

Short rotations Sequatchie Mixed Stream Well Nearly Rare 1–3 % Medium Prime Farmland or intensive silt loam Alluvium terraces drained level use

Sandstone/ Level/ Staser fine Flood Well Mixed Occasional Nearly 0–2 % Medium Prime Farmland Intensive use sandy loam plains drained Alluvium level

Slate and Staser silt Flood Well Nearly Shale/ Mixed Occasional 0–2 % Slightly Prime Farmland Intensive use loam plains drained level Alluvium

A All soils present have moderate organic material with low erosion risk and favorable to adequate moisture relations. B Prime Farmland classification is defined by the NRCS as having the best combination of physical and chemical characteristics for producing food. This includes soil properties, growing season, and moisture supply. Table compiled from the USDA NRCS 2017 and Elder et al. 1959

impacted plant and animal populations, as well as peoples in Tuckaleechee Cove.

The climate of the region is classified as temperate within the mesothermal humid subtropical zone (Köppen 1931). Mesothermal climates are located at mid- latitudes, with mild winters and hot, humid summers. The precipitation and temperature gradients of the Great Smoky Mountains are controlled mainly by elevation and they produce a series of microclimates, each with distinct ecosystems (Busing et al.

2005; Shanks 1954). For example, although the Great Smoky Mountains National Park

(GSMNP) gets an average annual precipitation of approximately 3480 mm/year, precipitation increases sharply from 1473 mm/year in the valley floors to approximately

2286 mm/year at an elevation of 1372 to 1524 m. Likewise, temperatures fall along this elevation gradient and decrease approximately 2° F/305 m gained in altitude, resulting in a 10 to 15-degree difference in temperatures between the valley floors and high elevation peaks (Shanks 1954; Whittaker 1956). Thus, microclimates in the park can vary from humid mesothermal in lower elevation locations to perhumid microthermal in the upper elevations, all within an 18-km air radius (Busing et al. 2005).

Besides elevation, a variety of factors contribute to the microclimates of the

GSMNP that are closely related to the ruggedness of the landscape; a more variable landscape will create more opportunity for greater contrasts between areas (Hicks 1979).

Factors such as slope direction, degree of exposure, and canopy cover can contribute to ground conditions. The potential effect in landscape features such as coves is a tendency

for overall wetter and cooler conditions than are experienced on exposed peaks and ridges (Hicks 1979).

Precipitation patterns in the park also vary according to their east-west location: the western portion is generally drier than the eastern (Underwood 2013). At an elevation of approximately 305 to 335 m and with a westward geographic location, average annual precipitation is on the lower end of the gradient in Townsend, at approximately 1422 mm per year. Temperatures are seasonal and range from an average annual winter low of 43.9 to an annual high of 66.3 degrees F, although temperatures can drop to an overnight low of 25 degrees in January and a daytime high of 82 degrees in July (US Climate Data 2017). Largely determined by temperature in temperate climate zones, the length of the growing season determines the types of plants and the productivity or crop yields of plants in an area. Tuckaleechee Cove falls within the USDA plant hardiness zone 7, with 189 frost free days per year and an average frost cycle beginning in late October and ending in mid-April. Blount County has approximately 191 frost free days per year, and is recognized as favorable for the growth of all of the common crops, especially in areas of well-drained soils (Elder 1959).

3.2.5 Surrounding Plant and Animal Communities

The availability of plant and animal resources on the landscape is instrumental in determining patterns of human movement and the location of settlements. Since the middle to late Paleoindian period in the southeastern United States, the broad

distribution of very diverse communitites of plants and animals contributed to increased familiarization with and adaptive use of distinct local environments, resulting in sub- regional culture areas defined by diversified economic systems and more generalized, locally produced toolkits (Anderson 1995:32; Anderson and Sassaman 1996:32; Bense

2009; Meltzer 1988). Throughout the subsequent cultural periods, people in this region continued to depend on the diverse wild flora and fauna available to them, despite eventually cultivating and domesticating several indigenous weedy seed species before transitioning to maize agriculture.

3.2.5.1 Fauna

The site area and nearby Smoky Mountains contain a wide diversity of amphibians, reptiles, and fish, as well as an abundance of small-and medium-size mammals, several species of large mammals, rodents, and birds. The park currently contains 65 species of mammals, over 200 species of birds, 67 native fish species, and over 80 species of reptiles and amphibians, and these numbers were undoubtedly higher prior to human impacts on the ecosystem (National Park Service 2017). Because of human activities, many important native species have been eradicated in this area, including the bison, elk, gray and red wolf, fisher, river otter, Peregrine falcon, and multiple fish species (National Park Service 2017).

The diversity of fauna in the region is directly related to the abundant and varied vegetation, the numerous tributaries and mountain streams, and the elevation gradients

that create unique habitats. With a wide variety of nuts, fruits, and seeds available, many economically important species of animals for humans, such as white-tailed deer, wild turkey, black bear, squirrels, red and gray foxes, and elk, were drawn to similar patches of resources and thus were regarded as a source of food and competition for plant resources (Hinkle et al. 1993:225). The 1127 km of streams in the park display a wide variation in temperatures and flow, supporting many different native species of fish. The diversity and number of birds in the park varies seasonally; however, the diverse vegetation and elevation of the park supports a wide variety of northerly and southerly migrating species (National Park Service 2017).

3.2.5.2 Flora

A comparison of the availability of economically important wild plant foods used by indigenous peoples in different sub-regions of the Eastern Woodlands highlights the exceptional diversity of the Interior Southeast, an area that includes eastern Tennessee and western North Carolina. For example, this sub-region is the most abundant in nut-producing trees, with 27 species available for consumption (Scarry

2003). The plant diversity in the immediate vicinity of the site area, and nearby in the

GSMNP, is unmatched in the eastern United States. It contains over 1570 species of vascular plants including 130 species of trees, 1300 species of flowering plants, and a rich understory that includes at least 3500 species of macrofungi, various mosses, and ferns (Houk 1993; King and Stupka 1950; Sharkey 2001; Whittaker 1956). The wide

range of plants available to occupants of the site area is a result of the relatively temperate climate and moderate rainfall patterns in Tuckaleechee Cove, combined with the complex topography and deep geologic age of the surrounding Great Smoky

Mountains (Wofford and Chester 2002:1).

The site area lies within the Central Appalachian Broadleaf ecoregion (Figure

3.3), defined by a mixed oak-pine forest composition. Common trees of this ecoregion include the yellow poplar (Liriodendron tulipifera), white and black oak (Quercus alba and

Quercus velutina), hemlock (Tsuga spp.), red maple (Acer rubrum), black locust (Robinia pseudoacacia), and sweet birch (Betula lenta) (Arends 1981). However, as with weather and climate, vegetation in the Blue Ridge is heavily controlled by elevation, and the tall, closed deciduous canopy of elevations under 1250 m transition to coniferous trees and shrubs in higher elevations above 1450 m (McNab et al. 2005).

The Blue Ridge Mountains section (Figure 3.4), which includes the site area, is dominated by oak-hickory, white pine, and oak-pine forests at elevations below 1500 m, but above this limit is characterized by coniferous spruce-fir forests (McNab et al. 2005).

Just to the west of the site area, the Northern Ridge and Valley Section is dominated by oak-hickory cover. Oak-hickory forests can vary from open to closed canopy and can have a variable understory (McNab et al. 2005). Based on soil moisture, the Central

Appalachian Broadleaf ecoregion is also classified as Mixed Mesophytic or

Figure 3.3. Appalachian Broadleaf Ecoregion (Level II) setting with Townsend sites.

Figure 3.4. Blue Ridge Ecoregion (Level III) setting with Townsend sites.

Appalachian Oak (Hinkle et al. 1993; Shelford 1963; Stephenson et al. 1993; Whittaker

1956). The site area, below an elevation of 1250 m and receiving a moderate amount of rainfall, can be classified as Mixed Mesophytic. Classified as a Cove-Hardwood Forest by Whittaker (1956:45–46), this type of forest is a center of biodiversity; this uniqueness is derived from high plant species richness, uneven-aged stands, and a complex stand composition (Braun 1950; Hinkle et al. 1993; Martin 1992). With sheltered valleys and deep alluvial soils, Tuckaleechee Cove can be considered among the grouping that is the most botanically diverse of the mesophytic forests. Eight species including eastern hemlock (Tsuga canadensis), mountain silverbell (Halesia monticola), yellow buckeye

(Aesculus flava), white basswood (Tilia heterophylla), sugar maple (Acer saccharum), yellow birch (Betula alleghaniensis), yellow poplar (Liriodendron tulipifera), and American beech

(Fagus grandifolia) make up 80 to 90 percent of the canopy, while seasonal ferns and herbs complete the understory with little competition from small trees and woody growth (Whittaker 1956:45).

The Appalachian-Oak forests are more common in higher elevation (>450 m

AMSL) and more xeric habitats, typifying the ridges that surround Tuckaleechee Cove

(Whittaker 1956). Dominant trees in the Mississippian period would have included chestnut (Castanea dentata) and several species of oak, along with pignut hickory (Carya glabra), black gum (Nyssa sylvatica), sweet gum (Liquidambar styraciflua), red maple (Acer rubrum), black locust, sassafras (Sassafras albidum) and sourwood (Oxydendrum arboreum), while a variety of lower trees, woody shrubs including blueberries (Vaccinium spp.), and

herbs comprised the understory (Shelford 1963:39; Stephenson et al. 1993:268; Whittaker

1956).

3.2.6 A Changing Environment

The Southern Appalachians appear to have been a place of refuge for plant and animal species over time, offering a wide variety of habitats and remaining unglaciated throughout the Pleistocene (Whittaker 1956:3). Long-term forest records in the region are based on paleoenvironmental reconstructions from pollen and charcoal in sediments and soils at archaeological sites. Generally, these records indicate a great degree of stability in forest composition for the past 5,000 years in the region (Crites and

Hollenbach 2011; Delcourt and Delcourt 1987). However, changes in pollen and charcoal do indicate that human activities such as burning and clearing impacted forests species composition for as long as 10,000 years in some areas (Crites and Hollenbach

2011; Delcourt et al. 1986; Underwood 2013).

In the nearby Little Tennessee River, Delcourt et al. (1986) identified changes in pollen composition associated with increases in forest clearance and cultivation that correlate to the intensification of indigenous horticulture and maize agriculture.

Underwood (2013:131–132) also found spikes in radiocarbon dates on soil charcoal at approximately 3500 cal yr BP and 1000 cal yr BP across study areas located approximately 13 to 29 km away from the Townsend sites. These spikes correspond chronologically to the transition to the Eastern Agricultural Complex and to the

widespread farming of maize. Underwood (2013) attributed these occurrences in multiple records to intensifications in human occupation and use of the landscape that were likely also taking place in Tuckaleechee Cove during the Late Archaic and the

Mississippian periods.

The stable landscape of Tuckaleechee Cove and the surrounding area provided a familiar and well-known resource base for its inhabitants. Access to multiple environments, all within a relatively short travel distance, presented people with many opportunities to exploit a variety of plants and animals, which would have been an additional advantage in times of climate instability. A diverse resource base undoubtedly proved important to reduce food shortages that sedentary farming communities faced, like risks associated with short-term weather events such as droughts and floods or slightly cooler or warmer temperatures, as well as problems equated with longer periods of climate stress (Anderson et al. 1995; Anderson and

Sassaman 2012; Kidder 2006).

3.3 Cultural History of the Region

3.3.1 Paleoindian Era Settlement and Land Use (>13,250–11,700 cal year BP)

The Paleoindian period in the Southeast can be divided into three subperiods:

Pre-Clovis (ca. >13,250 cal yr BP), Clovis (ca. 13,250–12,850 cal year BP) and post-Clovis fluted and unfluted (ca. 12,850–11,700 cal year BP), each defined by changes in climate

and culture (Anderson 2001, 2004, 2005; Anderson and Sassaman 2012; Anderson et al.

2015). Although there are a number of sites in the region that claim Pre-Clovis occupation (e.g. Cactus Hill, Virginia; Saltville, Virginia; Coats-Hines, Tennessee; Little

Salt Spring, Florida; Page-Ladson, Florida; and Topper, South Carolina), there is still ample scholarly debate over the artifact assemblages and contexts from these sites

(Anderson and Sassaman 2012). Early Paleoindian incursions into the region were also likely few and far between. This situation changed drastically in the Middle and Late

Paleoindian, where we see a marked increase in the number of sites and artifact styles across the Southeast (Anderson et al. 2015).

It was once widely thought that Paleoindian peoples were highly mobile, technologically dependent, and exclusively targeting large megafauna. However, increasing evidence suggests that at least in some instances, Paleoindian peoples likely took advantage of resource-rich environments for longer periods of time by establishing staging or habitual-use areas (Anderson 1990, 1996b; Dincauze 1993). For instance, based on the abundance of Clovis points and early dates of sites, it appears that Clovis peoples quickly mapped on to the resource-rich river valleys in the Southeast. The

Cumberland and Tennessee rivers in Tennessee, for example, have large numbers of fluted points and in some cases, large, diverse site assemblages (Anderson et al. 1996:5;

Broster and Norton 1995).

While there are growing numbers of megafauna kill sites in the Southeast

(Breitburg et al. 1996; Brush and Smith 1994; Graham et al. 1981; Webb et al. 1984), the

archaeological record, specifically growing faunal and botanical assemblages from deeply stratified and well-preserved Late Paleoindian sites (Dunbar and Vojnovski 2007;

Hollenbach 2007, 2009; Walker 2007), increasingly points to a generalized subsistence strategy with a heavy focus on high-return plants and small game that developed from the Early to Late Paleoindian periods (Anderson and Sassaman 2012:65; Cannon and

Meltzer 2003). There is also increasing evidence of much more variation in toolkit assemblages than was previously assumed (Anderson et al. 2015). These differences are likely a reflection of the environmental and social conditions of individual sites

(Anderson and Sassaman 2012). It is increasingly clear that there is much more diversity within the Paleoindian umbrella than was previously understood or expected.

Late Paleoindian peoples responded to a changing environment, global warming, glacial melting, and megafauna extinctions by creating local cultures that are defined archaeologically by successive projectile point styles from fully fluted forms to unfluted forms that were present across the Southeast (Anderson and Sassaman 2012).

Among the most well-known in the region, the Dalton culture exemplifies the increasing technological specialization and ceremonialism that characterizes the Late Paleoindian.

Although there is little direct evidence of ritualistic symbolism through portable or rock art, ritual activity is signified by the production of ceremonial points such as the Sloan, the creation of caches, and the elaboration of burial practices (Anderson and Sassaman

2012:62).

It is likely that the human occupation of Tuckaleechee Cove in the Paleoindian period was limited. Currently, there is little evidence to suggest sustained use of the

Appalachian Summit region in the Paleoindian period (Anderson and Sassaman

2012:50). At the Townsend sites the Paleoindian occupation is documented by the presence of one Quad projectile point that was found at site 40BT90 (Driskell 2011:34).

However, based on the increasing number of isolated Paleoindian sites, as well as a growing awareness of the more intensified use of some sites in Tennessee and the

Southeast in general, it is likely that the Paleoindian presence in east Tennessee will continue to increase in visibility and diversity.

3.3.2 The Archaic Period: Establishing Patterns (11,500–3200 cal yr BP)

The end of the Younger Dryas and the onset of the Holocene mark the boundary between the Late Paleoindian and the Archaic period. Although Archaic peoples amplified, sometimes at a swift rate, patterns that began in the preceding Late

Paleoindian period including growing populations, increasing interactions and trade networks, and intensified and more long-term use of certain places on the landscape, there is a substantial continuity in settlement and subsistence patterns. Based primarily on climate trends, and changes in technology in the Southeast, the Archaic is divided into three subperiods, the Early (11,500–8900 cal yr BP), Middle (8900–5800 cal yr BP), and Late (5800–3200 cal yr BP) (Anderson and Sassaman 2012:66).

Populations in the Early Archaic period continued patterns of reorganization and adjustment that began in the Late Paleoindian in response to rapid climate and environmental change. The beginning of the Holocene, and thus the Early Archaic, is defined by an increase in global temperatures, intense hydrologic activity, and an expansion of hardwood forests. Use of Late Paleoindian forms of projectile points continued into the Archaic, but side- and corner-notched points, along with bifurcate- based hafted bifaces primarily made from local sources, hint at continuing shifts in the way that people were using the landscape. Overall subsistence strategies moved to generalized foraging centered on locally available plant and animal resources and place- based settlement tied to landscape features (Anderson and Sassaman 2012:72).

Despite decreased archaeological visibility of Early Archaic sites due to erosion and heavy sedimentation caused by fluvial processes associated with glacial run-off from large stream systems (Chapman 1985a; Schuldenrein 1996:3; Sherwood et al. 2004), the density of Early Archaic artifacts and sites greatly increases from the preceding

Paleoindian period (Anderson and Sassaman 2012:72). Multiple deeply buried sites, including those excavated in the Tellico Reservoir (Chapman 1975, 1977, 1979, 1985b;

Kimball 1996), exemplify this demographic change in eastern Tennessee. Coupled with

Bass’s (1977) survey of sites in the Great Smoky Mountains, these data give us a good understanding of the settlement strategies of Early Archaic peoples in this area, suggesting intensive seasonal utilization of riverine locales combined with temporary exploitation of upland environments. Based on the presence of a modern assemblage of

plants and animals, a variety of plant and animal processing tools, and scant, but revealing botanical and faunal assemblages from archaeological sites, it appears that a diversified subsistence pattern was in place by this time.

Dramatic changes in social life occurred in the Middle Archaic, likely in response to changing environments that were a product of the Hypsithermal or Holocene Climate

Optimum. While there was a large degree of variation in the conditions experienced across the Southeast during this climate event (see Anderson and Sassaman 2012:73 and

Anderson et al. 2007 for regional overview; see Delcourt 1979, Delcourt and Delcourt

1985 for local example from Tennessee), the period is generally defined by increased warming and drying, as well as seasonal extremes. Additionally, over the course of the middle Holocene in the Southeast, geologic processes stabilized, resulting in reduced sedimentation and stream channel normalization (Schuldenrein 1996:3).

Although it is unclear to what degree there were unified cultural traditions based on economics and trade or religious ideology (Saunders 2010:71–72), in at least some parts of the Southeast during the Middle Archaic, there was a profusion of new social traditions (Anderson and Sassaman 2012:74). This heightening of social complexity and interaction is most visible in the creation of large shell accumulations and the construction of earthen mounds and mound complexes. The appearance of stemmed biface technology, along with apparently dramatic social changes, has raised questions about the possibility of migrations of people into the area (Anderson and Sassaman

2012; Coe 1964; Sassaman 2010), but it is still unclear whether these innovations in

culture and technology were of local or external origins (Anderson and Sassaman

2012:74).

In East Tennessee during the Middle Archaic, perhaps in response to climate fluctuations (Chapman 1985a:149; Ward and Davis 1999:70), people appear to have occupied dispersed settlements in both the uplands and valleys and coves (Ward and

Davis 1999:70). An influx of new technology such as stemmed projectile points, netsinkers, and atlatl weights (Chapman 1977, 1985a:148; Davis 1990) changed the ways that people were acquiring wild game, but there was a large amount of stability in the use of wild plant foods (Chapman 1977:125; Chapman and Shea 1981; Lewis and Lewis

1961:40–43).

At the Townsend sites, based on the minimal number of Early and Middle

Archaic diagnostic projectile points, as well as the lack of period contexts, the use of

Tuckaleechee Cove during these periods appears to be ephemeral and strictly focused on resource extraction (Hollenbach and Yerka 2011:394). It was not until the subsequent period, the Late Archaic, that there was a greater investment in and more permanent occupation of the area.

For thousands of years, people in the Southeast were settling in, becoming well versed in the intricacies of their local environments. In addition to this cultural familiarity, climate conditions in the Late Archaic stabilized, adding a new level of predictability and security to daily life. If people in the Middle Archaic experienced a

revolution in social life, in the Late Archaic people began a revolution in the way that they interacted with their environment by growing their own crops.

Beginning in the Late Archaic period, in addition to hunting and gathering, people domesticated a suite of indigenous seed crops, including sumpweed (Iva annua), sunflower (Helianthus annuus), and chenopod (Chenopodium berlandieri), and likely tended and cultivated several more including maygrass (Phalaris caroliniana), little barley

(Hordeum pusillum), erect knotweed (Polygonum erectum), and giant ragweed (Ambrosia trifida). The presence of larger and more permanently occupied settlements in many areas of the Southeast, including in East Tennessee and western North Carolina

(Chapman 1981, 1985a:150; Davis 1990; Ward and Davis 1999:70–71), testifies to the increasingly complex ties that people must have developed to specific places on the landscape. Although common attributes in projectile points were shared regionally, sub-regional variations occur. These regional differences are demonstrated by the

Savannah River Stemmed and the later, smaller Iddins Undifferentiated Stemmed in

East Tennessee as compared to the Otarre Stemmed points of western North Carolina

(Keel 1976:231; Ward and Davis 1999:71).

People established stronger ties to individual places on the landscape, and they also began to participate more in extra-local interactions. The events and processes that unfolded at the large and impressive mound complex of Poverty Point in eastern

Louisiana unquestionably influenced regional and local culture. In eastern Tennessee, and from the Townsend sites, soapstone vessels with Poverty Point style designs (Wells

2006) point to the regional movement of goods, but also beliefs and social relationships

(Anderson and Sassaman 2012:75–76).

The investment in Tuckaleechee Cove is much greater in the Late Archaic, evident in the 82 features represented by two hearths, nine basins, 63 pits, three multi- pits, and three rock concentrations (Hollenbach and Yerka 2011:63). Based on the presence of a wide range of feature types, including several large and extra-large category storage pits, along with the spatial positioning of features, it appears that people visited the sites repeatedly, or more likely lived at them for extended periods of time (Hollenbach and Yerka 2011:71–84).

3.3.3 The Woodland Period: Trends Writ Larger (3000–1000 cal yr BP)

Whereas the Archaic period is subdivided into periods that align with climate events, the Woodland period’s subdivision into Early (3000–2150 cal yr BP), Middle

(2150 BP–1550 cal yr BP), and Late (1550 BP–950 cal yr BP) is linked to interwoven processes of technological, economic, and social change, many of which were put into motion in the preceding Late Archaic. The Woodland period is archaeologically synonymous with ceramic vessel technology and smaller, stemmed and triangular projectile points (Anderson and Mainfort 2002:1–2). Although the period experienced ebbs and flows, over the course of the Woodland, small, semi-sedentary, mostly egalitarian horticultural communities transformed into large, permanent, and socially stratified agricultural villages.

The Early Woodland is difficult to distinguish culturally from the Late Archaic; however, archaeologically, the widespread adoption of pottery that coalesced into distinctive ceramic cultures, including that of the South Appalachian area, defines the period (Anderson and Mainfort 2002:5). Based on temper and surface decoration treatments, the Watts Bar, Long Branch, and Swannanoa traditions characterize East

Tennessee, while Swannanoa vessels represent the dominant tradition in western North

Carolina. Speaking to the chronological utility of ceramic assemblages, cultural phases in eastern Tennessee are defined by percentages of the aforementioned ceramic types

(Lafferty 1981:500; McCollough and Faulkner 1973:93-94; Schroedl 1978a).

Limited data suggest that settlement size was small, populations were relatively mobile, and exchange was reduced in many areas of the Southeast, but there is evidence that Adena influence from the north, especially mortuary practices, filtered into some areas of Kentucky, south-central Tennessee, Mississippi, Arkansas, and Alabama

(Anderson and Mainfort 2002:7). In eastern Tennessee, as in preceding periods, settlements continued to focus on the floodplains, with ephemeral use of the upland environments as well (Davis 1990:229–230; Lafferty 1981:516–524).

Evidence from East Tennessee suggests that horticultural practices intensified in the Early Woodland, with routine recovery of native cultigens alongside gathered staple resources such as nuts (hickory, acorn, and black walnut) and various wild fruits

(Brewer 1973; Chapman and Shea 1981; Crites 1996; Knott 1981). This pattern is consistent with the plant remains recovered from Late Archaic and Early Woodland

contexts at the Townsend sites, in which mast resources remain the primary component of the gathered plant diet, but we begin to see more evidence of cultigens, specifically maygrass and little barley (Hollenbach and Yerka 2011:362). Also in East Tennessee, at the Birdwell and Nease sites, people focused on the floodplain setting of the landscape, burning and preparing the area for the cultivation of these weedy seed species

(Johanson 2010).

The Middle Woodland is most often marked by the extent of a community’s participation in the burial mound tradition of the Hopewell culture. What was once viewed as a panregional Hopewellian culture has been tempered across most areas of the Southeast (Anderson and Mainfort 2002:9). The Southern Appalachian region is a good example of what Anderson and Mainfort (2002) considered to be a “thin veneer” of

Hopewellian influence throughout the Southeast in the Middle Woodland. Hopewellian artifacts are clearly found at sites throughout the region, including the Townsend sites

(Chapman and Keel 1979; Cridlebaugh 1981:180-181; Hollenbach and Yerka 2011; Keel

1976:157; Keel and Egloff 1984:34; Wetmore 1996:227), but the occupants were engaged in a growing interaction sphere which included several areas, including eastern

Tennessee, western North Carolina, and northern Georgia. Furthermore, the interactions taking place were not unilateral; ideas and cultural material from communities in the Southeast moved northward into Hopewell as well (Walthall

1985:252; Wetmore 2002:263), which makes it difficult to separate cultural influences.

This is especially true in light of the longstanding tradition of mound construction in the

Southeast, which may have made Hopewell traditions less intrusive (Anderson and

Sassaman 2012:132).

A large amount of variation in interregional relationships appears to have been present geographically and temporally in the Southeast. For instance, in the

Appalachian Summit at the site of Garden Creek, there is evidence of both direct face-to- face interaction with Hopewell in Enclosure No. 1 and less formal interaction at Mound

1 (Wright 2014). Based on very different manifestations of Hopewellian culture at this single site, Wright (2014:292) argued that interactions with Hopewell cannot be explained by “unitary visions”, but rather need to be evaluated within dynamic historical frameworks.

In the Little Tennessee River Valley and the Southern Appalachians, river valleys and bottomlands were still important loci for residential sites, but there was continued use of upland environments (Davis 1990:234; Purrington 1982:58, 1983:139; Ward and

Davis 1999:154; Wetmore et al. 2000:142). Based on the increasing number and variety of features at sites across the region, sites were growing larger and becoming occupied seasonally or semi-permanently (Hollenbach and Yerka 2011:47). This is the case for the

Townsend sites, where the Middle Woodland component is extensive, represented by

449 features and more than 527 postholes, and is the first time we see construction of burial mounds (Hollenbach and Yerka 2011).

Likely tied to the exchange network that placed East Tennessee into contact with

Hopewell and nearby regions, during this period maize was introduced in meager

amounts and used alongside native cultigens and wild resources at approximately 200

CE (Chapman and Crites 1987:353; Hollenbach and Yerka 2011:396). Maize agriculture did not contribute greatly to the diet until much later, but its early presence and association with Hopewell seems to indicate that it played a social, religious, and/or economic role in Middle Woodland society.

The Late Woodland in the Southeast is variable; some areas experienced what appears to be abandonment, while other areas saw unabated social and economic growth, resulting by the later stage of the period in very Mississippian-like characteristics such as large civic-ceremonial centers, social stratification, platform mound construction, shell-tempered pottery, wall-trench construction, and intensive maize agriculture (Anderson and Mainfort 2002:18).

In East Tennessee, due to few diagnostic Late Woodland technologies, this period is often hard to identify; however, evidence suggests that social construction was not completely halted in this region—some mound building continued and trade networks with north Georgia remained in place (Hollenbach and Yerka 2011:50). While the use of riverine resources is highly visible in many assemblages (Bogan 1982:41;

Faulkner and Graham 1966:131- 132; McCollough and Faulkner 1973:124-125), there was consistency in traditional subsistence patterns and people continued to grow indigenous crops, collect wild plant foods, and use corn in small amounts.

Interestingly, the Townsend sites seem to display the opposite of this regional pattern and conform more to traditional ideas of Late Woodland cultural decline. Based

on ceramic assemblages, only nine features have Late Woodland contexts and within those, none have yielded maize.

3.3.4 The Mississippian Period (1020–500 cal yr BP)

Archaeologists have traditionally defined Mississippian cultures by a set of common characteristics: chiefdom-level organization, maize-based subsistence, shell tempered pottery, dense population centers, and monumental architecture. While some of these trends may have in fact started in the Late Woodland/Emergent Mississippian period, they began to take a much larger role and influenced everyday life to a much greater extent in the Mississippian.

The Mississippian cultural period spans from approximately 1000 to 1550 CE and includes many large mound complexes such as Cahokia in the American Bottom and

Moundville in central Alabama ,and Etowah in northern Georgia. These sites likely exerted significant social, economic, and political control over populations both locally and regionally. Pauketat (2007) argued that Cahokia may have even been the first North

American city-state that served as the ideological foundation, a “pax Cahokiana,” of

Mississippian society. Based on the size and cultural dynamism of Cahokia, it is not difficult to suggest that in the Mississippian there was, as Peregrine and Lekson (2012) argued, a North American “oikoumene”—an interconnected web of cultures, present across the Southeast, and that Cahokia was at the center. However, large sites are just one of the common settlement layouts in this period and many people resided in smaller

villages and single-family farmsteads in upland and rural locations. There is variability in the degree of interconnectedness between these settlements and larger centers.

If Cahokia was a vehicle for ideas and cultural change, then maize agriculture was the economic tool used to underwrite the process. In most areas of the Southeast, farmers had fully adapted maize agriculture by 1000 CE. Maize surpluses fueled growing populations, increasing social complexity, and stratification. Consequently,

Mississippian foodways studies have traditionally assessed the level of participation of communities in maize agriculture, the scale of production (e.g. Baden and Beekman

2001; Scarry 2008; Schroeder 1999, 2001), and the social manipulation of surplus stores namely through feasting (Blitz 1993; Maxham; VanDerwarker et al. 2007) and status- building enterprises (Welch and Scarry 1995).

Paleoethnobotanical analyses have been critical in providing insight into the social enterprise of Mississippian foodways. The scale of production varied geographically; women in Mississippian societies grew maize in both household and communal lots and surplus stores provided nutrition for families, communities, and elite enterprises (Scarry 2008). The abundance of maize macrobotanical remains from sites has been used as an obvious indication of the crops’ importance. However, comparisons of maize macrobotanical remains to other categories of plant taxa, such as ratios of maize refuse to nut refuse or indigenous crops, have provided more accurate estimations of the importance of maize across sites (Jefferies et al. 1996; Scarry 2003:84).

Feasting activity is often identified based on the presence of special combinations of

plants that contradict the normal domestic pattern (Briggs 2016; Welch and Scarry 1995).

Lastly, difference in status-related foodways has been defined in terms of the kinds and portions of food, along with differences in serving and preparation that are visible in ceramic vessel assemblages (Briggs 2016; VanDerwarker 1999; Welch and Scarry 2005).

Welch and Scarry (2005) employed a comparison of corn kernel to corn refuse ratios to identify potential instances of tribute payments (Welch and Scarry 2005). In instances where there is a deficit of corn kernels at one site and a surplus of kernels at another, it could be inferred that tribute in the form of consumable material was transferred between locations.

Paleoethnobotanical analyses also contribute to an understanding of how foodways can both contribute to and reflect Mississippian social circumstances. Briggs

(2016) argued that Mississippian identity and daily practice was impacted markedly by a transition to hominy cooking. However, VanDerwarker and Wilson (2016) also pointed to the substantial impact that social interactions at large have on foodways.

Comparing the changing abundance of nuts, fruits, and wild greens/seeds, all of which are foraged foods, they suggested that residents in the Central Illinois River Valley likely reduced foraging activity at the expense of their health due to the sustained threat of violence and warfare.

In a similar vein, VanDerwarker and Wilson (2016) raised the possibility that residents in times of stress may have transitioned from outfield to infield cultivation.

Because kernels represent consumption and cupules and glumes represent processing

and discard activity, a low kernel to cupule ratio is a good indicator of processing activity (Scarry and Steponaitis 1997:117). Therefore, a higher cupule ratio within a palisaded area, is potentially highly related to closely positioned fields and nearby processing activity. Rather than risk endangerment away from the protection of palisade walls, they suggested that cultivation in fields just outside the palisade walls would have been an effective strategy to minimize exposure to danger. Extending

Killion’s Infield/Outfield model (1987), which suggested that processing occurs near the source of cultivation, VanDerwarker (2010) used the decrease in the kernel to cupule ratio at the site of La Joya in Mexico to suggest a transition towards infield cultivation.

Food production strategies were likely greatly impacted by negative inter-group interactions throughout the Mississippian Southeast and novel paleoethnobotanical approaches that identify daily practice have much to contribute to interpreting the structure of daily lives.

A chronology of Mississippian culture in East Tennessee was first developed by

Lewis and Kneberg (1946) based on excavations in the Norris Reservoir and

Chickamauga Basin. The period was initially divided into Early

Mississippian/Hiwassee Island Culture (1000–1200 CE) and Late

Mississippian/Dallas/Mouse Creek Culture (1200–1600 CE). This chronology was later revised based on further excavations of radiocarbon-dated Mississippian sites in the area, including those conducted during the Tellico Archaeological Project in the Lower

Tennessee River Valley (Schroedl et al. 1990). The modified chronological sequence

divides the Mississippian period into four phases: Mississippian I/Martin Farm (900–

1000 CE), Mississippian II/Hiwassee Island (1000–1300 CE), Mississippian III/Dallas,

Mouse Creek (1300–1600 CE), and Mississippian IV/Overhill Cherokee (1600–1838 CE).

These phases have been almost exclusively derived from the varieties and quantities of ceramic types found within their respective archaeological assemblages, but have also been associated with changes in technologies, settlement layouts, mortuary patterns, and social complexity. The Early Mississippian occupation in East

Tennessee belong to the Martin Farm phase and Hiwassee Island phases. Martin Farm is viewed as a transitional period where new cultural traditions emerged from Late

Woodland populations. The transitional nature of this phase is accentuated by the use of limestone and shell-tempered ceramics, along with construction of Hamilton conical and platform mounds (Koerner and Sullivan 2011:25). During the Hiwassee Island phase, of which the Townsend sites are most closely affiliated with temporally, shell- tempering became the dominant ceramic type. In addition, well-adorned burials and a new use of Mississippian symbols hint at expanding relationships with regional centers such as Etowah to the south (King 2002; Lewis and Kneberg 1946; Lewis et al. 1995;

Sullivan 2009). Settlements ranged from dispersed farmsteads to small villages that were comprised of wall-trench and single-post rectangular architecture that were occasionally associated with a mound, plaza, and palisade enclosure (Shroedl 1998).

The Late Mississippian in East Tennessee is represented by the Dallas and Mouse

Creek phases. Dallas phase settlements differ significantly from their Hiwassee Island

predecessors and movement into nucleated defensible villages became more common in the Tennessee Valley (Koerner and Sullivan 2011:29). Differences between the two phases are based primarily on more pronounced social and political complexity with

Dallas phase, which is inferred from additional categories of public structures, mortuary patterns suggestive of ascribed status, and a larger range of house size (Sullivan 1995;

Schroedl 1998). Even with these differences, there is continuity in household assemblages including in the architectural grammar of the household, which is represented by a winter and summer house, a cleared activity space, and storage facilities (Sullivan 1995), and in the use of above-ground storage (Schroedl 1998:82).

There is a wide range of diversity in the type of Mississippian-period sites in East

Tennessee, from villages with and without associated mound complexes, to hamlets and farmsteads. Most of the sites excavated were located along the Tennessee and Little

Tennessee rivers. The nearest large centers to the Townsend sites and perhaps most relevant are Citico (40MR7) and Toqua (40MR6), both sites along the Little Tennessee

River that represent larger mound and habitation complexes. The size of the mound at

Citico, along with mortuary data, is suggestive of its potential influence in the Hiwassee

Island phase. Similarly, Toqua also has platform mounds and habitation areas, with the addition of a bastioned palisade. Occupied from the Hiwassee Island to Dallas phase, it is likely that this center exerted regional influence, but its relationship with upland sites is unknown.

Other excavated sites in East Tennessee outside the major river valleys include the Martin Farm site (40MR20) and the Davis Farm site (40BT47), both of which are more comparable to the Townsend sites in settlement size and cultural material. Martin Farm, located on the Little Tennessee River, was an early, small settlement typified by subterranean storage pits and a collection of mostly shell-tempered, plain pottery; however, it did have one platform mound (Koerner and Sullivan 2011:42; Schroedl et al.

1985). Davis Farm is one of the only other excavations that has been conducted on the

Little River. Similar to the Townsend sites, it was not a mound complex, but rather appeared to be the remains of a small village site. Despite its short-term occupation, from 1200 to 1290 CE, it also was palisaded and contained mostly Hiwassee Island-style cultural material mixed with small amounts of Pisgah style pottery (Greene 1991;

Koerner and Sullivan 2011:39). Several chiefdoms dating to the Hiwassee Island and the

Dallas phase were present in the Norris Basin clustered in the Clinch and Powell rivers

(Meyers 2002:31). Some of these are characterized by multiple mounds, while others were large without mounds.

Further complicating cultural chronologies along the eastern border of

Tennessee, the ceramics in this region display a strong affinity to ceramic assemblages found in western North Carolina. The culture chronology for western North Carolina and the Southern Appalachian region is separated into the Pisgah Phase (1000–1450 CE) and the subsequent Qualla Phase, which is associated with the Cherokees (after 1350

CE). Pisgah culture is commonly identified by pottery types, which are dominated by

rectilinear complicated stamped designs in the Early subphase (1000–1250 CE) and rectilinear complicated stamped and curvilinear designs in the Late subphase (1250–

1450 CE) (Ward and Davis 1999). Both of these phases have strong ties to traditions in north Georgia (Ward and Davis 1999:169). Located primarily in floodplains suitable for corn agriculture, Pisgah settlements often included paired earthen lodges, platform mounds, and palisades. The Warren Wilson site and the Garden Creek site are both mound centers that have helped to define the Mississippian period in this area. While

Pisgah communities display many Mississippian characteristics such as construction of platform mounds, palisades, and a subsistence strategy based on maize agriculture, they are generally viewed as more egalitarian than the communities in the Tennessee River

Valley, with a wide range of variability in settlement size and political influence (Ward and Davis 1999).

Cahokia is not only closely linked to the rise of Mississippian identity, but can also be considered emblematic of the widespread changes that occurred at the end of the period across the Southeast. These changes are represented by site abandonment, population decline and movement, and increasing warfare that were intimately tied to the challenges presented to large sedentary agricultural communities by a changing climate (Anderson and Sassaman 2012:165). Warfare in Mississippian societies is evidenced by heightened levels of palisade building, skeletal trauma, settlement nucleation, and militaristic iconography (Anderson and Sassaman 2012; Cobb and Butler

2002). These data indicate that endemic violence was pervasive throughout the region

(Cobb and Butler 2002). By the thirteenth century in the Central Illinois River Valley, intensification of warfare led to the disruption of everyday subsistence strategies, causing food shortages and nutritional insufficiency (VanDerwarker and Wilson 2016).

In East Tennessee, the extent of intergroup violence is unclear. Fortifications have been documented at many of the floodplain sites. Because Larson (1972:384) linked fortification to agriculturally productive bottomlands, he suggested that fortified sites were not frequent in the piedmont or mountainous sections of the Southeast. However, excavations of upland sites in East Tennessee (e.g. the Townsend sites and Davis Farm) contradict this notion. Likewise, palisades are common in western North Carolina’s

South Appalachian Mississippian tradition, which is suggestive of the need for agricultural communities to protect their surplus stores (Ward and Davis 1999). Both the Warren Wilson and Garden Creek sites have villages encompassed within palisaded walls. Based on low incidences of skeletal trauma in the Chickamauga Reservoir, Smith

(2003) argued that there may have been less intergroup violence in this area. However, she acknowledged the potential for undocumented violence in interpolity buffer zones.

There is a continued need for more work on sites in East Tennessee and the

Southern Appalachian Mississippian tradition in western North Carolina to reveal the interrelationships between these areas and the large chiefdoms in the Tennessee Valley, northern Georgia, and even Cahokia. I suspect that similar to the variation present in the uplands around Cahokia, there are many different reasons for variation in site

settlement patterns and artifact assemblages in the Southern Appalachian region of the

Southeast.

CHAPTER 4

ARCHAEOLOGICAL EXCAVATIONS AT THE TOWNSEND SITES

4.1 Introduction

The Townsend archaeological sites were excavated as a large-scale mitigation effort conducted to mitigate the negative effects of the widening of U.S. Highway 321 in

Blount County, Tennessee. Fieldwork was completed from 1999 to 2001 by the

University of Tennessee Transportation Center’s Archaeological Studies Group with funding from the Tennessee Department of Transportation (TDOT). Continued research of materials excavated from the sites was directed by Boyce Driskell (Principal

Investigator) at the Archaeological Research Laboratory (ARL) with the University of

Tennessee’s Anthropology Department. To date the largest archaeological excavation conducted by TDOT in the state of Tennessee, the Phase III work completed at this site complex resulted in cultural remains that date from the Archaic to the Historic Period.

In this chapter I summarize the three-year excavation and over ten-year laboratory analysis and report preparation of the Townsend archaeological sites, emphasizing the Mississippian-period analysis and interpretation as it currently stands.

In this dissertation, I build upon this extensive work, both in the field and laboratory.

Following a review of the identifications of Mississippian architectural remains and features from each respective site as determined by Koerner and Sullivan (2011), I detail how I constructed settlements and households to understand the plant foodways in this study.

The nature of the Townsend Archaeological Project as a mitigation effort provided advantages and disadvantages for archaeological research that were 74

considered in the interpretations for this dissertation. As a full-scale data recovery project, the scope of the excavation and quantity of cultural material that was recovered is largely unmatched outside of the cultural resource management industry. The total site area at Townsend is large; it is approximately 4.9 miles long and covers approximately 175 acres (Figure 4.1). Altered and re-used for thousands of years, the

Townsend sites are a testimony to the repeated human occupation of Tuckaleechee

Cove. Untangling the often-mixed deposits of individual cultural periods was a significant challenge for interpretation of the sites. An additional challenge was the recovery and analysis of 1,200 boxes of artifacts and samples, containing over 1.4 million artifacts (Driskell 2011:184).

In addition to the magnitude of the project itself, the planning and implementation of the work benefited from it being a joint undertaking, with significant input from multiple agencies and with the consultation of 10 tribal authorities including the Eastern Band of the Cherokee Indians (Driskell 2011:320). The Memorandum of

Agreement (MOA) mandated that any burials exposed during the excavation were left in place when possible. Therefore, the 104 features that were identified as burials, and their associated artifacts, were given a temporal cultural affiliation when possible based strictly on their in-field evaluation of artifact associations, feature morphology, or association with affiliated period deposits (Driskell 2011:411).

Figure 4.1. Townsend site overview for 40BT89, 40BT90, and 40BT91.

Data recovery methods included intensive flotation sampling that resulted in multiple, comprehensive datasets, including the botanical dataset, that contribute to our broad understanding of the development of culture and communities across prehistory and history in this region. Reports to be submitted to the Federal Highway

Administration (FHWA) and the TDOT will provide a record of over 10,000 years of human occupation in the foothills of the Smoky Mountains in East Tennessee. However, the excavation was limited to the artificial boundaries imposed by the area of potential effect (APE), regardless of cultural material that extended further outside the sanctioned excavation limits.

Initial Phase I archaeological investigations into the APE expanded beyond the original site designation of 40BT89 that was identified in 1998 by TDOT archaeologists, and included a total of six archaeological sites (Driskell 2011:11). However, only four of these sites, 40BT89, 40BT90, 40BT91, and 40BT94, were recommended for further Phase

II and III data recovery. Of these, only sites 40BT89, 40BT90, and 40BT91 contain appreciable Mississippian-period deposits, with the latter two of these containing by far the most significant concentrations of cultural material. Partially in response to the unique linear and narrow geography of Tuckaleechee Cove (see Chapter 3 for a discussion of the natural history of the area) and partially due to the modern constraints placed on prehistoric landscapes, such as the presence of modern driveways, one large site complex was divided into several rather contiguous bounded sites (Figure 4.1).

These multiple designations are useful for research efforts due to the large size of the site

complex; however, it should be noted that the separations are arbitrary divisions and do not represent meaningful divisions to prehistoric people living in the area.

4.2 Artifact and Data Management

To manage the large amount of spatial information and large number of artifacts gathered in the extensive excavations of this multicomponent site complex, personnel involved in subsequent analysis and reports management were divided into teams

(Driskell 2016:3). The division of these “teams” was based on Temporal Cultural

Affiliation (TCA), as well as material type. Thus, the TCA or “period teams” included

Archaic and Woodland, Mississippian, Cherokee, and EuroAmerican, all of which are slated to result in distinct volumes produced for the TDOT. The multiple artifact classes and samples required in-depth data analysis and there was a distinct need for a truly multidisciplinary research team. The interpretive analysis was facilitated by the separation of the work into teams led by specialists including spatial/temporal, subsistence, material culture, and environmental (Driskell 2016:63–70).

The compilation of the spatial and temporal attributes of the sites into a

Geographic Information Science (GIS) and Microsoft Access database provided critical resources to understanding the associations of artifacts and features at the Townsend sites. The GIS and database were made important by the large number of artifacts and features, as well as the comingled nature of the multicomponent deposits.

The Mississippian Period report was also an essential resource that I referred to during the construction of and completion of this dissertation research. Described in detail in this chapter, I largely follow the identifications made by Koerner and Sullivan

(2011) of Mississippian-period deposits, features, and architectural elements. In addition, I further tested their interpretations of the site complex during the

Mississippian period (Section 4.4) using the plant remains, a category of data that is uniquely positioned to address questions of site function, identity, and status across the site complex. Faunal remains are not discussed in detail or included in further analyses due to their generally poor preservation across the sites.

4.3 The Townsend Mississippian Occupation: Sites 40BT89, 40BT90, and

40BT91

The densest occupation of the Townsend Sites dates to the Middle Woodland

Period, 200 BCE to 550 CE. This occupation is represented by a series of 527 postholes,

449 features, and three circle middens that point to the repeated permanent or semi- permanent habitation of the site (Hollenbach and Yerka 2011). Compared to the relatively minimal occupation of the Townsend sites during the Late Woodland period from approximately 550 CE to 900 CE, Tuckaleechee Cove again became an attractive location to live and to build a community in the Mississippian period, from approximately 900 CE to 1300 CE. Over a period of roughly 400 years Mississippian period occupants built homes and communities at this location that were superimposed

on their Woodland predecessor’s investments into the landscape – for example, built on top of Woodland-period architecture including three circle middens at 40BT90. There are three main types of settlements found throughout the site complex: ephemeral, farmsteads, and villages. The diverse settlement layouts are reflective of the ways that people chose to respond to the social conditions they experienced in the Mississippian.

These decisions are particularly important to our understanding of the cultural designation at large because the wide range of responses seen at Tuckaleechee Cove in many ways requires us to consider the complexity and breadth of cultural diversity in the period.

4.3.1 Identification of Mississippian Period Deposits

Mechanical and hand excavation during the Phase III excavations of the

Townsend sites revealed extensive Mississippian-period deposits primarily concentrated at sites 40BT89, 40BT90, and 40BT91. Because of the overlapping nature of the cultural deposits at the Townsend sites, postholes and features received a Temporal Cultural

Affiliation (TCA) mainly based on the presence of temporally diagnostic artifacts, posthole dimensions, and in some cases through the use of radiocarbon dating (Driskell

2011). The artifact classes most heavily used to assign a TCA were projectile points/knives (PP/Ks) and ceramics. In total, 115 features were assigned a Mississippian designation (Table A.1), of which 27 structures were identified as domestic architecture

(Table A.2; Figure 4.2), two structures were designated as community buildings (Table

A.3; Figure 4.2), 17 were considered corncribs/circular buildings (Table A.4; Figure 4.2), and four were primary palisaded enclosures with three bastions (Table A.5; Figure 4.2).

Domestic architecture at the Townsend sites is rectangular to square in form, measures approximately 9-m-x-9-m meters, and does not contain central posts (Koerner and

Sullivan 2011). This style of architecture aligns well with the flexed-pole construction of the Hiwassee Island type. The two community buildings, Structure 1 (designated as M in this dissertation) and Structure 13 (designated as N in this dissertation) contrasted from domestic architecture based on a larger size of 12 m x 12 m, minimal associated features, and rounded corners (Figures 4.2 and 4.3). For the purpose of a comparative analysis of their material remains, they were given classifications of M and N; however, they are considered distinctive from the other domestic architectural remains.

The Mississippian occupation at the larger Townsend site complex dates most clearly to the Hiwassee Island Phase (1100–1300 CE) or Mississippian II (Koerner and

Sullivan 2011:53). I follow, the established temporal division of the site into two periods based on both radiometric dating and artifact attributes: Early Mississippian dating from approximately 900 to 1100 CE and Middle Mississippian dating from approximately

1100 to 1300 CE (Koerner and Sullivan 2011:52). This chronology places the occupation of the Townsend sites within the Martin Farm and Hiwassee Island Phase in the

Tennessee River Valley (Table 4.1) and the Early and Late Pisgah in the Appalachian

Summit Region (Table 4.2).

Figure 4.2. Architectural remains from sites 40BT90 (A) and 40BT91 (B) (Koerner and Sullivan 2011) illustrating palisade and structure outlines designated 1–61 (see Tables A.2–A.5).

Figure 4.3. Two community buildings, Structures 1 (designated as M here) and 13 (designated as N here) located in Village 1 at site 40BT91 (Koerner and Sullivan 2011).

Table 4.1. Mississippian Cultural Phases in East Tennessee.

Period Phase Approximate Date (CE)

Mississippian I Martin Farm 900–1100

Mississippian II Hiwassee Island 1000–1300

Dallas/Mouse

Mississippian III Creek 1300–1600

Mississippian IV Overhill Cherokee 1600–1838

Table adapted from Koerner and Sullivan (2011:Table 3). Based on Kimball and Baden (1985).

Table 4.2. Mississippian Cultural Phases in Western North Carolina.

Period Phase Approximate Date (CE)

Mississippian I Early Pisgah 1000–1250

Mississippian II Late Pisgah 1250–1500

Mississippian III Early Qualla 1300–1500

Table adapted from Koerner and Sullivan (2011:Table 3). Based on Dickens (1970) and Rodning (2004).

4.3.2 Material remains from the Kinzel Springs Site (40BT89)

Site 40BT89, the Kinzel Springs Site, is the most westward site of the complex and is separated from the other Mississippian deposits at 40BT90 and 91 by roughly 280 meters. While the Phase II survey and testing did not uncover exceptional

Mississippian-period deposits, other period deposits that were identified warranted

Phase II testing and data recovery from this location (Driskell 2011:17). Four areas were investigated (Figure 4.4), using a combination of methods including mechanical stripping, mapping, and photography, followed by the acquisition of flotation samples, as well as feature and test unit excavations (Driskell 2011:45). The Phase III excavations at 40BT89 did reveal Mississippian deposits, although they were limited in nature and were only represented by seven pit features, scattered postholes, and 12 burials, with no clearly associated habitation areas (Koerner and Sullivan 2011:70). Four radiocarbon dates obtained from wood and nutshell remains from four features returned dates between 971 CE and 1411 CE (Table 4.3). Out of the seven Mississippian features identified at 40BT89, three were classified as dating to the Early Mississippian, three were Middle Mississippian, and one was General Mississippian (Koerner and Sullivan

2011:72–74). Figures 4.5–4.7 display the spatial distribution of Early, Middle, and

General Mississippian features across site 40BT89 (see Chapter 5 for in-depth feature classification and analysis).

40BT89 appears to be an example of a somewhat ephemeral, short-term, and limited investment in the Cove based on the lack of postholes indicative of structures or 86

fortifications, and the minimal quantity, diffuse nature, and limited type of features at the site. There is no evidence of domestic structures, no clear arrangement of space, and the features are suggestive of only short-term storage. Due to the wide range of time that the site was occupied, the limited quality of the cultural remains, and the distance of the remains from other more clearly defined Mississippian deposits at 40BT90 and

40BT91, the Kinzel Springs Site was interpreted to be a result of the limits of the excavation boundary, the effects of differential preservation and representative of a low- density occupation, or a specialized activity area (Koerner and Sullivan 2011:70).

4.3.3 Material Remains from the Apple Barn Site (40BT90)

40BT90, the Apple Barn Site, was excavated by subareas designated as A, B, and C using block excavations and 2-m-x-2-m units (Figure 4.8). Phase II excavations revealed deep alluvial deposits with Late Archaic, Middle Woodland, Mississippian, Historic

Cherokee, and nineteenth-century Euro-American remains that warranted that the site undergo Phase III data recovery. The Mississippian deposits at this site were extensive and composed most noticeably of palisade remnants associated with features, limited structures, and burials that as a unit appeared to be representative of domestic areas or compounds. Figures 4.9–4.12 illustrate the spatial distribution of Early, Middle, and

General Mississippian features along with postholes representative of Mississippian period structures and palisades across site 40BT91. Seven radiocarbon dates corresponding to ages from 1043 CE to 1288 CE (Table 4.4).

Figure 4.4. Four areas of investigation at site 40BT89 designated as Area A, B, C, and C-Short Creek (adapted from Koerner and Sullivan 2011:Figure 5).

Table 4.3. Radiocarbon Dates from Site 40BT89.

Uncalibrated Mean 95% Cal. Ft. Material Lab. No. 14C Age δ13C Age Age Range (14C yr BP) (CE) (CE) 251 Acer spp., Juglans spp., ISGS-5758 1060±70 -25.7 971 776-794 (2%) Pinus spp. 799-1053 (88%) 1079-1152 (10%) 124 Carya spp. nutshell ISGS-5777 880±70 -25.6 1147 1026-1261 25 Pinus spp., Quercus (red) ISGS-5780 640±70 -25.9 1343 1262-1423 Carya spp. 89 Pinus spp., Acer spp., ISGS-5843 510±70 -26.7 1411 1291-1512 (98%) Juglans cf. nigra, 1601-1615 (2%) Quercus (white) Adapted from Koerner and Sullivan (2011:Table 4) with new mean age and age ranges. Calibration dataset used for 95% (±2 !) calibrated age range (CE) is IntCal 13.14c (Reimer et al. 2013; Stuiver and Reimer 1993).

Figure 4.5. The distribution of Early, Middle, and General Mississippian features from site 40BT89 showing isolated feature in northwest corner of the site (Adapted from Koerner and Sullivan 2011:Figure 21).

Figure 4.6. The distribution of Early, Middle, and General Mississippian features from site 40BT89 moving east (1) (Adapted from Koerner and Sullivan 2011:Figure 22).

Figure 4.7. The distribution of Early, Middle, and General Mississippian features from site 40BT89 moving east (2) (Adapted Koerner and Sullivan 2011:Figure 23).

Figure 4.8. Three areas of investigation, A, B, and C, at site 40BT90 (Adapted from Koerner and Sullivan 2011:Figure 6).

Figure 4.9. The distribution of Early, Middle, and General Mississippian features and postholes representative of structures and palisade remnants from site 40BT90 moving east (adapted Koerner and Sullivan 2011:Figure 10).

Figure 4.10. The distribution of Early, Middle, and General Mississippian features and postholes representative of structures and palisade remnants from site 40BT90 moving east (2) (adapted from Koerner and Sullivan 2011:Figure 11).

Figure 4.11. The distribution of Early, Middle, and General Mississippian features and postholes representative of structures and palisade remnants from site 40BT90 moving east (3) (adapted from Koerner and Sullivan 2011:Figure 12).

Figure 4.12. The distribution of Early, Middle, and General Mississippian features and postholes representative of structures and palisade remnants from site 40BT90 moving east (4) (adapted from Koerner and Sullivan 2011:Figure 14).

Table 4.4. Radiocarbon Dates from Site 40BT90.

Uncalibrated Mean 95% Cal. Ft. Material Lab. No. 14C Age δ13C Age Age Range (14C yr BP) (CE) (CE)

69 Pinus spp., Quercus (red), ISGS-5737 1000±70 -26.6 1043 892-1190 (99%) Carya spp. 1198-1203 (1%) 588 Pinus spp., ISGS-5753 880±70 -26.2 1147 1026-1261 Quercus (white) 1753 Carya spp. and ISGS-5846 880±70 -25.7 1147 1026-1261 Juglans nigra nutshell 729 Pinus spp., Quercus (red), ISGS-5825 860±90 -26 1161 1015-1288 Quercus (white), Carya spp., Diospyros virginiana, Arundinaria gigantea 1610 Jugladnaceae (cf. Carya) ISGS-5760 810±70 -25.8 1209 1040-1109 (14%) nutshell 1116-1291 (86%) 697 Pinus spp. ISGS-5770 750±70 -26 1256 1054-1078 (2%) 1153-1328 (87%) 1341-1395 (11%) 1795 Pinus spp., Quercus spp., ISGS-5744 710±70 -26 1288 1189-1405 Carya spp. nutshell Adapted from Koerner and Sullivan (2011:Table 4) with new mean age and age ranges. Calibration dataset used for 95% (±2 !) calibrated age range (CE) is IntCal 13.14c (Reimer et al. 2013; Stuiver and Reimer 1993).

This site is distinguished by three distinct palisaded enclosures and three

“village areas,” classified by Koerner and Sullivan (2011) as the West Palisade A, Central

Palisade B, and East Village Palisade C and D, plus one additional area of domestic debris that is located on the far eastern boundary of the site (Figure 4.2).

Based on the density of features and the clarity of structure and palisade postholes, the East Village (designated as Village 2 in this dissertation) was interpreted as consisting of a more cohesive community structure although it did contain features spanning a range of time periods: Early, Middle, and General Mississippian.

40BT90 has architectural elements that indicate an occupation representative of greater planning and investment in the site area. A total of five segments of palisaded enclosures, 10 domestic structures, 9 circular structures, and 36 features, some of which are directly associated with distinct households, typify the farmstead settlement of the

Cove at this site.

Nine small, roughly 2 m in diameter, circular structures were identified from posthole patterns at 40BT90 (Table A.4; Creswell 2016). While only two of these contained Mississippian-period artifacts, their association with the period was based on their proximity to Mississippian dwellings, as well as their similarity to other Hiwassee

Island phase structures in East Tennessee (Creswell 2011:184). An example of a similar type of structure can be found in the remains of Structure 2 at the Martin Farm Site. The building at the Martin Farm Site was hypothesized to be the remains of either a storage

facility or “corncrib,” menstrual hut, or shamans hut based on its circular pattern, as well as its lack of associated floor or features (Schroedl et al. 1985:86).

The nine small circular buildings identified at 40BT90 were found in the eastern half of the site, mostly in relatively isolated locations. Apart from Structure 13, which is just south of Structure 18, the circular buildings are not clearly associated with a household. Structures 53, 54, and 55 may be related to Household 58; however, they are located outside of the palisade.

4.3.4 Cultural remains from the Pony Ride Site (40BT91)

Site 40BT91, the Pony Ride Site, was also identified during the initial Phase I excavations and excavated as a series of subareas (A, B, and C) during the Phase II and

III recovery efforts (Figure 4.13). This site had moderate to dense Middle Woodland,

Mississippian, Historic Cherokee, and nineteenth-century Euro-American components

(Driskell 2011:18). The Mississippian component at this site—a collection of structures, features, and burials associated with numerous enclosures that represent multiple domestic and community use areas— was the densest of the Townsend site complex.

Apart from two early dates at 1498 BCE and 866 CE and one late date at 1504 CE, the occupation dates from approximately 1137 to 1387 CE (Table 4.5).

40BT91 has components that date from both the Early Mississippian and Middle

Mississippian designation at Townsend. At the western edge of the site, a densely occupied and fortified village area (designated as Village 1 in this dissertation) that

100

Figure 4.13. Three areas of investigation, Area A, B, and C, at site 40BT91 (adapted from Koerner and Sullivan 2011:Figure 7).

101

Table 4.5. Radiocarbon Dates from Site 40BT91.

Uncalibrated Mean 95% Cal. Feaure. Material Lab. No. 14C Age δ13C Age Age Range (14C yr BP) (CE) (CE) 530 Carya spp. AA55840 1162±45 -27.26 866 728-737 (1%) 768-984 (99%) 138 Pinus spp., ISGS-5734 890±70 -27.4 1137 1023-1259 Carya spp. nutshell 584 Juglandaceae nutshell ISGS-6142 880±70 -25.3 1147 1026-1261 13 Juniperus virginiana, ISGS-5728 740±70 -25.6 1264 1156-1333 (85%) Pinus spp., 1336-1398 (15%) Carya spp. nutshell Pinus spp., Quercus 371 ISGS-5762 700±70 -26.3 1297 1195-1195 (1%) (white), Carya spp. nutshell, 1206-1410 (99%) Quercus (acorn) cotyledon 274 Quercus (red and white) ISGS-5726 680±70 -26.2 1315 1222-1409 Carya spp. PM 710 Pinus spp. ISGS-5754 640±70 -24.8 1343 1262-1423 632 Quercus (white) AA55843 618±38 -26.43 1349 1290-1404 544 Carya spp. AA55841 553±36 -26.1 1387 1306-1363 (48%) 1385-1434 (52%) Str. 9 Quercus (red) AA55838 385±36 -26.02 1504 1440-1527 (64%) 1554-1633 (36%) 517 Carya spp. nutshell ISGS-6122 3220±70 -25.1 -1498A -1662-1377 (96%) -1345-1305 (4%) Adapted from Koerner and Sullivan (2011:Table 4) with new mean age and age ranges. Calibration dataset used for 95% (±2 !) calibrated age range (CE) is IntCal 13.14c (Reimer et al., 2013; Stuiver and Reimer, 1993). A Negative value indicates BCE value for mean age and for age range.

102

Figure 4.14. Village 1 located at the western edge of site 40BT91 (adapted from Koerner and Sullivan 2011:Figure 61).

103

contains domestic and community structures was interpreted to align closely with many cultural attributes associated with the Hiwassee Island phase (Figure 4.14). Based on radiocarbon dates from Features 13 and 584, both located along the southern palisade, this village dates from approximately 1147 to 1264 CE (Figure 4.14; Table 4.5).

However, a second set of radiocarbon dates obtained from Village 1, from

Feature 632 and Postmold 710 that were associated with Structure 22 (Figures 4.14 and

4.15), returned a date of 1343 and 1349 CE (Table 4.5). The presence of a seemingly

Pisgah related household within the fortified village is one of the compelling difference between this site and the others. The only other household that was identified as Pisgah was located in the far southeastern extent of 40BT91 (Figure 4.16). However, based on radiocarbon dating, Koerner and Sullivan (2011) suggested that the Pisgah settlements represent a separate occupation of the Cove. However, the clear spatial arrangement of households, including Structure 22, within Village 1 is suggestive of a relatively contemporary occupation of Structure 22 and the rest of the domestic architecture.

Besides domestic architecture, six of the eight corncribs from 40BT91 are located within this compound (Figure 4.14; Table A.4). Only Corncrib 18 could clearly be associated with a domestic structure (Structure 17) and the remaining were in what could be considered communal areas.

The fortified and densely occupied village area of 40BT91, termed “Skittletown” by the field archaeologists and designated as Village 1 in this dissertation, stands in contrast to much of the cultural material found throughout the rest of 40BT91 and the

104

other sites, which is more dominated by farmsteads and scattered or isolated settlements. The only other settlement that compares to this occupation is the domestic compound, Village 2 in this dissertation, located in the eastern-most extent of 40BT91.

Because of the very distinct palisade walls, the remains of at least five domestic structures, two community structures, and six corncribs, as well as the variety and density of structures, the evidence of rebuilding episodes, and the bastioned palisade,

Koerner and Sullivan (2011:168) interpret the Village 1 enclosure to be the remains of a multiple family compound occupied for approximately 40 to 50 years.

Farther to the east, there are three separate concentrations of postmolds and features representative of possible domestic compounds. Two of these compounds are limited in material and are likely isolated or single family farmsteads (Figures 4.17 and

4.18). Based on feature content alone, the first could only be categorized as General

Mississippian, while the second has Early Mississippian components.

The dense compound in the eastern portion of the site, which I designate as

Village 2, has a mixture of Early and Middle Mississippian features (Figure 4.19), and the two radiocarbon dates that were secured from Features 530 and 544 also display a wide range of occupation (788 and 1397 CE), which raises the possibility that these features and structures may not have been co-habitated. One additional circular structure is located at the eastern edge of Village 2; however, because it is larger than the standard corncrib at the Townsend sites it may represent a misidentified structure.

105

Figure 4.15. Pisgah affiliated Structure 22 in the northern portion of Village 1 at site 40BT91.

106

Figure 4.16. Isolated Pisgah affiliated farmstead in the southeastern portion of site 40BT91 (Koerner and Sullivan 2011:Figure 77).

107

Figure 4.17. One of two isolated farmsteads/domestic debris at site 40BT91 including a single feature, and postholes representative of a single domestic structure (Structure 5), corncrib structures (Structures 14 and 15), and palisade remnants (Koerner and Sullivan 2011:Figure 63).

108

Figure 4.18. Second of two isolated farmsteads at site 40BT91 including features and potential domestic debris (Structure A) (Koerner and Sullivan 2011:Figure 64).

109

Figure 4.19. Domestic compound on eastern boundary of site 40BT91 (designated as Village 2 in this dissertation) including features, a floor midden (Structure 9), and postholes representative of domestic structures (28, 29, 30, 32, and 23/24), corncribs (Structure 25 and 7), and palisade remnants. (Koerner and Sullivan 2011:Figure 65).

110

The one additional isolated Pisgah farmstead in the southeast portion of the site (Figure

4.16) returned dates from Features 530 and 138 of 866 CE and 1137 CE (Table 4.5). Its association with the other settlements at 40BT91 are unclear.

As with the farmstead-type settlements at the site complex, the palisade construction surrounding the Village 1 and Village 2 compounds is unusual for both its early Mississippian age and three apparent circular bastion additions, Bastion S, SE, and

NW (Figure 4.14). Palisades have been documented as early as the Woodstock Late

Woodland period in northwest Georgia, but bastions are rare. These projections from a main palisade wall are first documented in East Tennessee in Late Mississippian Dallas communities dating from approximately 1300 to 1500 CE. If the projections at 40BT91 are indeed bastions, they would predate the earliest and only occurrence of this feature in East Tennessee found at the Toqua Site, which dates to the Dallas Phase (Polhemus

1985:218), by 50 to 300 years. While bastions are commonly viewed as a defensive feature (a tower or lookout area) in communities where warfare is commonly a threat, their function at these upland villages is unexpected. Koerner and Sullivan (2011) raised the possibility of a misinterpretation of these posthole configurations during excavation.

I suggest their distinct form along with associated large, artifact-rich features indicate a special function for these architectural palisade elements.

Although not in a major river system such as the Tennessee River, a combination of rich alluvial soils along the Little River (Larson 1972) and a geographic position along the boundary of multiple environmental zones (Beck Jr. and Moore 2002) may have

111

contributed to a high value placed on this location, particularly for maize farming, which had a high economic importance in the Mississippian period. One example of this situation was proposed by Schroedl et al. (1990:188), who suggested that the sandy loams along the Little Tennessee River, noted for their highly productive qualities for agriculture, may explain why many of the first-order river terraces were relinquished in the Mississippian II to farming rather than habitation. At the Townsend sites, with restricted living space due to the narrow geography of Tuckaleechee Cove, inhabitants may have decided that protection of this fertile landscape was the best option. In the absence of evidence of warfare, an alternate explanation is that these bastion structures may have been ritualized, symbolically charged points of entry such as gateways into the community. Koerner and Sullivan (2011) propose that due to the architectural design of the NW bastion, it appears to be such an entryway.

4.3.4 Summary of Previous Interpretations

The Mississippian-period deposits at the Townsend site complex of 40BT89,

40BT90, and 40BT91 represent a diverse arrangement of structures and features, with artifact assemblages that were interpreted to most closely resemble sites located in the

Tennessee River Valley during the Hiwassee Island phase (Koerner and Sullivan 2011).

This affinity to East Tennessee cultural groups rather than those in nearby western

North Carolina was based on a majority of shell-tempered ceramics rather than sand- tempered, and domestic architecture characterized by single-pole wall trench

112

architecture without a vestibule entrance, which is characteristic of Pisgah domestic architecture. Sand-tempered Pisgah ceramics were present at the sites, but were scarce

(Koerner and Sullivan 2011:86–91). Interestingly, however, based on ceramic composition and a smaller structure size, two possible Pisgah households were identified among and nearby the Hiwassee Island style settlements at the sites.

Although these households appear to date later than the Hiwassee Island structures, they speak to the fluid movement of people and culture in this borderland region over time.

Based on the proximity of the multiple enclosures, on the lack of evidence of rebuilding episodes, the minimal presence of hearths, Koerner and Sullivan (2011) determined the Townsend sites to be predominantly representative of people using the area for intermittent and potentially seasonal habitation over the course of four centuries. Along with these lines of evidence, a scarcity of decorative ornamentation and a minimal number of forms within the ceramic technology, as well as the presence of structures and pits outside the palisade walls were interpreted as potential evidence of a special function of the Townsend sites. The exact function is unclear based on the architectural or cultural material. The lack of a platform mound at the site complex suggests the absence of a chiefly authority and a larger degree of egalitarianism in these communities; however, I would argue that we should not view this absence as evidence of reduced social complexity.

113

4.4 Classifications Used in this Study

The household identifications in this study used the structure identifications based on posthole patterns that were established in Koerner and Sullivan (2011).

However, because of the paired nature of some of the domestic architecture, I considered structures that were roughly side by side to be a single domestic unit. On account of their proximity, I presume that these are instances of a paired structure such as a winter and summer house, structures inhabited by a larger kinship unit, or evidence of sequential rebuilding by a distinct household. Therefore, I suggest that instances of paired structures are indicative of a group of people, likely an extended family unit, working towards shared economic and social goals.

The function of the 17 circular structures is unclear at the Townsend sites. The buildings are identified as “corncribs” in the field and site report (Driskell 2011; Koerner and Sullivan 2011), with the presumed function being above-ground surplus food storage. This classification is conjecture only, as limited data in the form of artifacts or features from these structures are available at the Townsend sites or other sites in the area with similar patterns (Angst et al. 2005:116–118; Creswell 2011; Creswell et al.

2012:75–76; Greene 1991:85–92; Kuttruff and Walling 2002; Polhemus 1987: 311–313;

Polhemus and Simek 1996; Salo 1969b:96; Schroedl 1998; Schroedl et al. 1985). I used distance to household as an indication of a corncribs relationship to a domestic household.

114

In sum, I identified 18 distinct households (A–V) across sites 40BT90 and 40BT91

(Figure 4.20 and Figure 4.21). Many of the households had associated features that I identified as a product of household activity based primarily on distance to domestic structure (Table 4.6).

The geographic locations of all households and their associated features I designated in this dissertation are available in Appendix Figure A.1–A.14. Using feature contexts and radiocarbon dates when available, I list relative and chronological dates of occupation for each of the farmsteads and village settlements in Table 4.7.

Along with the identification of distinct households, I also considered groupings of households, other structures, features, and when present, palisades to be representative of three primary types of settlements at the Townsend sites: ephemeral, farmsteads, and villages (Figure 4.20 and Figure 4.21). Ephemeral settlements referred strictly to features that were located in an isolated geographic location that could not be clearly assigned to any domestic architecture.

I identified a total of eight farmstead settlements (Table 4.7; Figure 4.20 and

Figure 4.21), which were defined based on the presence of one to two domestic structures, a collection of associated features, and in some cases a palisade segment.

Farmsteads were a common settlement type in the Mississippian period and were composed of single-family farms or household clusters scattered across the rural landscape (Emerson 1997:156). At the Townsend Sites, we see a similar pattern, with one domestic structure surrounded by an open area (Koerner and Sullivan 2011:168).

115

Figure 4.20. Location of farmsteads, villages, and households from Townsend site 40BT90.

116

Figure 4.21. Location of farmsteads, villages, and households from Townsend site 40BT91.

117

Table 4.6. All Households with Associated Subterranean Storage Features at the Townsend Sites.

Feature Structures Community Household Site Number Morphology Size TCA Associated Type A 40BT90 212 basin large MSE 2/5 Farmstead 1 267 basin large MSE 2/5 269 pit small MSS 2/5 C 40BT90 1610 pit large MSE 58 Farmstead 2 1622 basin small MSS 58 1795 pit med. MSM 58 1825 pit small MSS 58 2120 basin small MSS 58 Communal 40BT91 88 basin large MSS 32/23/24 Village 2 89 basin small MSS 30/32 92 basin small MSS 30/32 479 pit small MSS 28/29/32 480 pit small MSM 28/29/32 507 basin small MSS 32/23/24 514 basin small MSS 23/24/30/32 516 basin small MSE 32/23/24 517 basin large MSE 32/23/24 D 40BT90 69 pit small MSE 60/61 Farmstead 3 1746 basin small MSE 60/61 1765 basin small MSS 60/61 1767 pit small MSS 60/61 1767A basin small MSS 60/61 E 40BT90 459 pit small MSS 18 Farmstead 4 473 pit med. MSE 18 510 basin small MSE 18 580 basin small MSS 18 588 pit large MSM 18 G 40BT90 561 basin small MSS 6/7 Farmstead 5 H 40BT91 650 basin small MSS 31 Village 1 J 40BT91 629 basin small MSS 22 635 basin small MSM 22 K 40BT91 678 basin small MSS 39/42/43/44 L 40BT91 623 pit med. MSS 36/37 672 basin small MSS 36/37 M 40BT91 599 basin small MSS 1 N 40BT91 582 basin small MSM 13 591 pit small MSS 13 O 40BT91 24 basin large MSS 5 Farmstead 6 P 40BT91 67 basin med. MSM 10 Farmstead 7 71 pit small MSS 10 72 basin large MSE 10

118

Table 4.6 (cont.). All Households with Associated Subterranean Storage Features at the Townsend Sites.

Feature Structures Community Household Site Number Morphology Size TCA Associated Type P 131 pit med. MSE 10 138 pit large MSE 10 139 basin large MSE 10 140 basin large MSM 10 142 basin small MSM 10 157 basin small MSM 10 Q 40BT91 124 pit Xlarge MSE 11/12 Farmstead 8 343 pit Xlarge MSE 11/12 371 pit med. MSE 11/12 U 40BT91 499 basin small MSS 32 Village 2 504 basin small MSS 32 V 40BT91 520 basin small MSE 23/24 528 basin med. MSE 23/24 542 pit small MSM 23/24 706 basin small MSS 23/24

119

Table 4.7. Dates of the Settlements at Sites 40BT90 and 40BT91.

Settlement Radiocarbon dates Feature TCA MSE MSM MSS Farmstead 1 1161 CE; 1256 CE 3 1 2 Farmstead 2 1209 CE; 1288 CE 3 2 4 Farmstead 3 1043 CE 4 0 3 Farmstead 4 1147 CE 3 1 2 Farmstead 5 0 0 1 Farmstead 6 0 0 1 Farmstead 7 1137 CE 4 4 2 Farmstead 8 1297 CE 3 0 0 Village 1 1147 CE; 1264 CE; 1343 CE; 1349 CE 0 7 13 Village 2 1498 BCE; 866 CE; 1387 CE; 1504 CE 6 3 12 Radiocarbon dates adapted from Koerner and Sullivan (2011:Table 4) with new ages (Tables 4.3, 4.4, and 4.5). Radiocarbon dates from feature and postmold contexts at sites 40BT90 and 40BT91.

120

Despite a general likeness to Mississippian farmstead settlements in the Tennessee River

Valley, the fortifications at 40BT90 are unusual at such an early date (1000–1100 CE) in the Mississippian period (Koerner and Sullivan 2011; Schroedl 1998).

The two village settlements are represented by Village 1, along with the dense grouping of domestic debris and the palisade at the eastern extent of 40BT91, Village 2.

The village areas stand apart from the farmstead locations by the number of domestic structures, well-defined palisades, and quantity of features. There is no evidence of mound construction during the Mississippian period at Townsend, despite an undoubted awareness of the nearby mound complexes along the Little Tennessee River and the Tennessee River.

Based on the cultural remains, architectural remnants representative of structures and fortifications, and the residue of past activities and activity areas as demonstrated by features, the Townsend sites generate a surprisingly complex picture of Mississippian settlement configuration and organization and thus, culture and lifeways at a single geographic location over time.

121

CHAPTER 5

ANALYSIS AND INTERPRETATION OF FEATURES

122

5.1 Introduction

A great deal of the information about the Mississippian occupation at Townsend was recovered from feature contexts. Koerner and Sullivan (2011) designated a total of

121 features as belonging to the Mississippian period. These include seven at site

40BT89, 36 at 40BT90, and 78 at 40BT91. The designation of a feature as Mississippian was based on the presence of temporally diagnostic artifacts, usually consisting of ceramic sherds of shell tempering and lithic projectile points, along with radiometric dating of organic material in select cases (Koerner and Sullivan 2011:199). Table 5.1 lists all of Koerner and Sullivan’s (2011:200–202) Mississippian designated features. Table 5.1 also includes the classifications that I developed, consisting of feature morphology, function, and size.

Due to the palimpsest nature of the site, many of the features contained mixed deposits of artifacts from multiple temporal cultural affiliations. In these cases, features were classified by Driskell (2011) as either as General Mississippian (MSS), or Early

(MSE) and Middle Mississippian (MSM) based on whether there was a majority presence of diagnostic ceramics belonging to one period. As a result of these mixed deposits, one must consider the possibility that classifications may have been made in error in some instances. This is especially true in cases where identifications at the sites were made between very similar technological styles such as the Late Woodland

Hamilton Incurvate and the Middle Mississippian Madison projectile points

123

Table 5.1. Mississippian Period Feature Classifications.

Site Feature Number Morphology Function Size 40BT89 25 basin storage large 89 basin storage medium 124 basin storage large 127 pit storage small 164 pit storage small 235 no data n/a n/a 251 pit storage extra-large

40BT90 10 basin cooking small 69 pit storage small 85 basin storage large 106 pit hearth small 212 basin storage large 267 basin storage large 269 pit storage small 294 pit storage extra-large 459 pit storage small 473 pit storage medium 510 basin storage small 561 basin storage small 580 basin storage small 588 pit storage large 697 pit cooking medium 729 pit storage large 873 basin storage extra-large 1109 pit storage medium 1498 basin storage medium 1531 basin hearth large 1590 pit cooking small 1610 pit storage large 1622 basin storage small 1746 basin storage small 1747 basin cooking small 1749 basin cooking small 1765 basin storage small 1767 pit storage small 1767A basin storage small 1795 pit storage medium 1825 pit storage small 1851 pit cooking medium 1856 pit cooking small 2001 pit storage small 2002 pit storage small 2120 basin storage small 124

Table 5.1 (cont.). Mississippian Period Feature Classifications.

Site Feature Number Morphology Function Size 40BT91 13 pit storage extra-large 24 basin storage large 37 basin storage large 48 basin hearth extra-large 67 basin storage medium 68 basin hearth small 71 pit storage small 72 basin storage large 88 basin storage large 89 basin storage small 92 basin storage small 105 pit storage medium 111 basin storage large 124 pit storage extra-large 131 pit storage medium 138 pit storage large 139 basin storage large 140 basin storage large 142 basin storage small 157 basin storage small 248 pit storage small 260 pit storage small 268 pit storage extra-large 272 basin storage extra-large 273 basin storage small 269 basin storage medium 274 basin storage large 294 pit storage small 343 pit storage extra-large 362 pit cooking medium 371 pit storage medium 375 n/a surface rock cluster n/a 479 pit storage small 480 pit storage small 482 basin cooking small 499 basin storage small 504 basin storage small 505 basin hearth small 507 basin storage small 514 basin storage small 516 basin storage small 517 basin storage large 520 basin storage small

125

Table 5.1 (cont.). Mississippian Period Feature Classifications.

Site Feature Number Morphology Function Size 40BT91 528 basin storage medium 530 basin hearth small 542 pit storage small 544 basin hearth large 581 basin storage small 582 basin storage small 584 basin storage large 585 pit storage small 591 pit storage small 599 basin storage small 623 pit storage medium 629 basin storage small 631 basin cooking small 632 pit hearth medium 635 basin storage small 650 basin storage small 672 basin storage small 678 basin storage small 690 pit storage small 705 basin storage small 706 basin storage small 711 basin cooking small 717/718 pit cooking small 753 no data n/a n/a 760 pit storage small 762 basin storage small 782 basin cooking small 805 pit storage small 807 pit cooking medium 812 basin storage medium 813 basin storage small 829 basin cooking small 832 pit storage small 851 basin storage small 861 basin storage small

126

(Driskell 2011:145). Feature 269 from 40BT90 is a good example of the problematic nature of dating some of these deposits. Although it included one Hamilton Incurvate projectile point, it otherwise contained four plain shell-tempered sherds, as well as a collection of Woodland and Middle Woodland sherds. However, founded on the inclusion of the shell-tempered ceramics, the feature was designated as Mississippian.

Due to the mixed nature of many of the deposits, I did not rely on the TCA designations further than to identify a feature or household as Mississippian. Instead, I chose to consider feature association based on distance to a structure as a better indication of the overall domestic household. Because it was difficult to determine the occupation dates of settlements, it is unclear what the relationship, if any, between them was like.

In this chapter I examine and compare the Mississippian features across the

Townsend sites. The analysis and interpretation of the features was an important task for several reasons. First, features and the artifacts within them were the primary way that temporal-cultural affiliation was determined. Second, the number and diversity of features at the Townsend sites are an indication of settlement type. That is, the longer a site is occupied and used for a variety of activities, if people are revisiting or living at a site seasonally or year-round, the more features and feature types we should see. This assumption allowed me to look for differences in how intensively the landscape was used in the Mississippian period across the settlements. Third, just as important as the

127

artifacts within them, the morphology, size, and spatial distribution of the features themselves were a window into the everyday activities in which people participated.

Considering these aspects of features, in this chapter I ask the following questions: Does the form and size of the features relate to function? Is there a correlation between feature type and size and artifact categories? Is there a relationship between settlement type and the organization of space? And, were there differences in the way that people were storing their food that may be related to distinctions in cultural affiliation or status? Because features are the remains of daily subsistence practices— storage, food preparation, processing, cooking, serving, consumption, and discard—they reflect how people’s choices in settlement type, farmstead or village, impacted their daily practices.

5.2 Feature Classification

I identified variation between Mississippian households at the Townsend sites, by first assigning a morphology and function to the Mississippian features. Following the work of Hollenbach and Yerka (2011) and Marcoux (2008), features were assigned a morphology based on profile shape along with field descriptions. Despite the inherent complexity of interpreting the life of a feature, approaches have been successful in identifying the possible extent of use life, the potential function of features, and how features relate to household activities.

128

At the Townsend sites, profile views were photographed and mapped for each individual feature. Profile maps of features recorded a variety of useful information such as dimension, depth, feature stratigraphy or zoning as indicated by soil color and texture, and significant inclusions, whether natural (e.g. root or rodent runs) or human produced (e.g. charcoal lenses, burning, or artifacts). All profile field drawings were digitized and georeferenced into an ArcGIS database. In addition, excavators frequently used a narrative to describe each feature, often citing the overall shape, size, and characteristics that they deemed useful, such as indicators of burning or artifact-rich concentrations or zones.

Based on morphology, I classified 70 of the 121 features (58 percent) as basins, 48

(40 percent) as pits, and one feature as a surface rock cluster with no associated stain or indication of disturbance (Table 5.1). Two additional features that were given a

Mississippian designation by Koerner and Sullivan (2011), Feature 235 from 40BT89 and

Feature 753 from 40BT91, had no corresponding planview or profile maps either from the field forms or in the ArcGIS database. Therefore, they were not given a classification as either basin or pit or considered further in this analysis.

Although there were small quantities (three of each), basins and pits were equally represented at 40BT89. At 40BT90, the distribution of basins and pits was very similar to that at 40BT89; out of the 38 total features from 40BT90, 18 were classified as basins, 19 were classified as pits, and one was a superficial rock cluster. This pattern does not apply to 40BT91, where I found substantially fewer pit features; only 26 of the

129

77 features were classified as pits, while 50 were classified as basins. One feature,

Feature 375, was classified as a surface rock cluster and was not considered further in this analysis.

I defined categories of features following conventions established by Schroedl

(1986) and Marcoux (2008) for Cherokee and Hollenbach and Yerka (2011) for Archaic and Woodland feature assemblages: Features that contain fire-cracked rock (FCR) with no or little associated stain are referred to as “rock concentrations.” Those whose profiles are rounded and demonstrate no definable break are termed “basins.” Features with profiles that show some distinction between walls and a base are “pits.” “Hearths” are features with pronounced zones of fire-reddened earth beneath a layer of fill, and were identified as such in the field. Figures 5.1 to 5.3 depict typical profiles of my definition of a basin, pit, and hearth using morphological shape. Depictions of all profiles that were recorded in the field and GIS database are available at the University of Tennessee’s Archaeological Research Laboratory.

Following the protocol developed by Hollenbach and Yerka (2011), I used size classification, in addition to morphology, to distinguish pit function and use. To create meaningful size classifications, I first used the ArcGIS database to extract a surface area for each feature. The surface area was then multiplied by the feature’s maximum depth

(cm) to estimate approximate feature volume (L). Feature volumes were subsequently examined quantitatively with the aid of histograms, and natural breaks in the dataset were used to define small, medium, large, and extra-large category features. I classified

130

the size of basins as: small (0–275 L), medium (300–425 L), large (575–1,125 L), and extra- large (1,375–1,775 L) (Figure 5.4). Based on the data distribution, a slightly different size classification was established for pits, to include small (0–375 L), medium (400–675 L), large (800–1,000 L), and extra-large (≥1,000 L) (Figure 5.5).

I further classified features as cooking features/earth ovens or storage features based on the quantity of fire-cracked rock (FCR) within them. The contents of storage features likely represents refuse, rather than the contents that were stored. Likewise, basins may have originally been borrow pits to extract daub for houses. Nonetheless,

FCR was frequently used prehistorically as a heating technology, most commonly for the roasting and boiling of food (Homsey 2009). Therefore, it is one measure to allow for a division of cooking versus storge features. Using the calculated FCR density (g) that was recorded in the field, along with the estimated volume of feature fill of each pit (L), I calculated the estimated density of FCR (g/L) of each Mississippian feature. Again, natural breaks in the dataset as visualized in histograms were used to quantitatively define cooking vs. storage features; those that contained a higher density of FCR were classified as cooking features and those with minimal FCR were classified as storage features. Calculations were done separately for both pit and basin features. Basins were classified as cooking if they contained greater than 130 g/L of FCR (Figure 5.6), while pits were classified as cooking if they contained greater than 121 g/L of FCR (Figure 5.7).

I further compared the amount of FCR (g/L) between cooking, storage, and hearths for

131

both basins and pits visually in a boxplot to ensure the validity of the categories (Figure

5.8).

Boxplots are used to compare categories of artifacts and plant taxa from flotation samples across the sites. Boxplots are primarily used as a visual aid to statistically evaluate trends in the datasets. In addition to providing information on the median values as represented by the notch at the center of the box, the hinge spread (the 25th and

75th percentile) and whiskers show the range of the dataset. Statistically significant differences between datasets at the 0.05 level is demonstrated by no overlap between the notched sections of two boxplots. The notches denote the 95 percent confidence interval.

Outliers are also denoted in the graph by asterisks and open circles, and are an important visual signal of unique features (McGill et al. 1978).

5.3 A Comparison of Storage Practices at the Townsend Sites

Koerner and Sullivan (2011:206) noted the relative paucity of large category pit features at site 40BT91. They associated the low number of large pit features with changing patterns in food storage over time in the Southeast—a shift from subterranean to above-ground storage facilities (DeBoer 1988). The presence of multiple potential above-ground storage “corncrib” structures in the Village 1 compound of 40BT91 may be further evidence of this trend.

Multiple explanations have been proposed to explain the shift from subterranean to above-ground storage, ranging from practical to ideological (DeBoer 1988;

132

Zone A

Rock Lining

Legend 00.250.5 Datum A Meters Datum B String Line Ground Surface Rock Rock Lining Zone A

Figure 5.1. Example of the morphology of a cooking basin in profile (top: Feature 10 from 40BT90) compared to a cooking pit in profile (bottom: Feature 590 from 40BT90).

133

Zone B

Zone A

Legend Datum A Datum B String Line 00.250.5 Ground Surface Rock Meters Zone A Zone B

Figure 5.2. Example of a storage basin in profile (top: Feature 580 from 40BT90) compared to a storage pit (bottom: Feature 2002 from 40BT90) in profile.

134

Zone A/B Zone C Zone D

Legend Zone C Datum A Datum B String Line 00.250.5 Ground Surface Meters Rock Zone A/B Zone C Zone D

Figure 5.3. Example of a hearth in profile (Feature 530 from 40BT91).

135

4 Count

0 25 75 125 175 225 275 325 375 425 475 525 575 625 675 725 775 825 875 925 975 1025 1075 1125 1175 1225 1275 1325 1375 1425 1475 1525 1575 1625 1675 1725 1775 Volume (L)

Figure 5.4. Histogram of volume of Mississippian basins.

136

4.5

3.5

2.5

Count 2

1.5

0.5

0 25 150 275 400 525 650 775 900 1025 1150 1275 1400 1525 1650 1775 1900 2025 2150 2275 2400 2525 2650 2775 2900 3025 3150 3275 3400 3525 3650 3775 Volume (L)

Figure 5.5. Histogram of volume of Mississippian pits.

137

20 Count

0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600 FCR Density (L)

Figure 5.6. Histogram of FCR density (L) of Mississippian basins (dashed line represents > 130 L = cooking feature).

138

18 16 14 12 10

Count 8 6 4 2 0 10 30 50 70 90 110 130 150 170 190 210 230 250 270 290 310 330 350 370 390 410 430 450 470 490 510 530 550 570 590 FCR Density (L)

Figure 5.7. Histogram of FCR density (L) of Mississippian pits (dashed line represents > 121 L = cooking feature).

139

Figure 5.8. Boxplots comparing the density of FCR (g/L) by Mississippian

feature function (outliers are not represented in entirety in this graph).

140

Johannessen 1993a; Marcoux 2008). Marcoux (2008:351–352) pointed to the greater suitability of certain crops (corn, beans, squash, and nuts) for aboveground storage. He suggested that at the Townsend sites, the initial shift to above-ground storage during the

Mississippian period helped to protect surpluses from wet conditions, while the shift back to subterranean storage during the English Contact period from 1670 to 1740 AD was a response to an increased use of root crops, which preserve well in cool, damp storage conditions. Alternatively, if as DeBoer (1988) suggested, storage type (above- or below-ground) can relay fundamental ideological and social information, with subterranean pits indicating a desire to conceal and above-ground storage suggesting ostentation, then the shift to above-ground storage at Townsend is a compelling argument for social change.

Johannessen (1993a) used a similar approach towards social change, by investigating how changes in foodways reflected or related to wider social phenomena in the Mississippi Valley during the Mississippian period. She concluded that shifts in storage practices, placement above or below-ground, and the location of features internal or external to the domestic structure, were tied to shifts in the larger social and political structures, as, for instance, a reflection of increasing or decreasing levels of autonomy.

The decline in pit use at 40BT91 is directly related to patterns that are closely linked to the Village 1 community of 40BT91. While this village is the most densely occupied of all the Townsend sites, only one of the five households has an associated

141

storage pit, Feature 632, (Figure 4.2), and it is small. Household J, which contains the pit, was interpreted as an anomaly at this village, because it appeared to be more related to the Pisgah culture and has a later date than the other households in the village, which have been interpreted to be more related to the Hiwassee Island culture. The internal position of this storage pit could indicate intention to hide or it may reflect different foodways traditions. While the five domestic structures do have one to two basin storage features associated with each of them, they are also of very small size. Ten of the

20 features are related to communal use and are positioned along the palisade or near the community buildings.

If this pattern is related to a shift to above-ground storage with the use of corncribs, only two of the corncribs, Structures 16 and 18, appear to be associated directly with individual households and the remaining four, Structures 2, 20, 21, and 34, are related to larger community buildings (Figure 4.2). Instead of individually owned underground storage features, it seems that storage in Village 1 was used as a potential signal of community integration. Based on a location near the community structures and just north of the bastioned palisade segment, these above-ground storage facilities may represent either a strategic place selected for protection of valuable surplus goods or a very visible signal of community wealth or status for visitors entering the village compound.

142

5.4 Differences in the Distribution of Cooking and Hearth Features

The functional categories of features, consisting of storage, cooking/cooking refuse, and hearths, resulted in 95 storage features, 15 cooking/cooking refuse features, 8 hearths, and 1 surface rock cluster (Table 5.1). Regardless of whether they were basin or pit shaped, the largest number of features across the site complex was used for storage

(79 percent). Somewhat perplexing, cooking features and hearths are present, but occur in very low numbers across 40BT90 and 40BT91. Cooking features were distributed almost equally between basins (n=8) and pits (n=7), while most of the hearths (six of the eight) were a classic basin shape.

The small number of cooking features and hearths across the sites is unusual considering the multiple habitation areas at 40BT90 and 40BT91. It is possible that the archaeological signatures of these activities were damaged or erased by post- depositional processes, specifically historic plowing activity (Koerner and Sullivan

2011). But, as Koerner and Sullivan (2011) pointed out, it is unusual that some of the hearths survived, while most did not. The survival of some hearths may also be related to differences in the site topography as different areas of the site have different burial depths. Koerner and Sullivan (2011:64) raised the possibility that the lack of hearths across the sites could indicate that people occupied Tuckaleechee Cove seasonally in the

Mississippian period. However, this explanation does not appear to be supported by the plant remains, palisades, and wide range of features, which argue for more long- term investment in the sites. It also seems unlikely that people would risk abandoning 143

their farmstead or village home—their crops, their stores, or their fertile farmland—for part of the year. Even if the Mississippian occupants of the Cove did only stay for short durations, it is unlikely that prepared fires for warmth, light, and cooking were not needed.

There are few patterns in the size or distribution of hearths that clarify the reasons for the absence of hearths. Out of the two hearths at 40BT90, neither was associated with a household. At 40BT91, six hearths were identified. Two of those were located inside the possible Pisgah households J and P (Figure 4.21), three were associated with two households at Village 2 (U and V) (Figure 4.21), and one extra-large hearth was in an isolated area of the site with no associations. I found no clear difference in the volume of the hearths related to their location; most of the hearths are small (50 percent), while one is medium, two are large, and one is extra-large. Koerner and Sullivan (2011:207–208), using length, width, and depth, found that four of the hearths were much larger than those from a comparative sample of similar sites from the region. However, I found no indications that these four hearths were larger due to a shared use, as none of the larger hearths are in areas that could be considered communal; two are associated with possible Pisgah households, one is in an isolated location, and one is in a farmstead with no association to a household.

Presently, there is no clear reason why so few hearth features were identified across the Townsend sites and the limited use of subterranean cooking features hints at explanations other than restricted use of the sites. Regardless of the presence or absence

144

of underground cooking facilities, people undoubtedly participated in cooking activities across the Townsend sites, especially in habitation areas. No cooking features were identified at 40BT89, indicating that people used this area less intensively, or that this portion of the landscape is a more disturbed context.

At sites 40BT90 and 40BT91, the absence of FCR-dense subterranean cooking features, especially among Mississippian habitation contexts, may relate to shifts in cooking practices over time. Refinement and intensified use of ceramic cooking vessels with well-made coiled pottery technology in the Mississippian period may have negated the need for the subterranean cooking facility. Instead women may have cooked food more frequently in a vessel over an open flame. VanDerwarker et al. (2012) also identified a substantial decrease in earth oven cooking facilities from the Late Woodland to Early Mississippian period in the American Bottom. They associated the decline with an ideological and practical shift from exterior communal cooking to interior individualized cooking.

Another possibility is that the change to open flame cooking was a result of food preparers, women, embracing new maize cooking techniques, with a focus on hominy which is cooked over an open flame for an extended period of time. Hominy processing, or alkali cooking, became a common and vital cooking technique for maize kernels in the Eastern Woodlands (Dezendorf 2013). It not only required significant cook times, but preparation was also intensive and required lengthy soaking and a session of cleaning and processing to remove outer layers and alkali residues (Briggs

145

2016:323). Using health estimates from the Moundville I phase population (1020–1260

AD) Briggs (2016) argued that alkaline treatment of maize kernels emerged and disseminated during the Late Woodland and Early Mississippian periods (1000–1200

AD). Transforming kernels to hominy has several benefits including an increased storage life, higher nutritional content, and more efficient removal of the pericarp (King

1987; Martinez-Bustos et al. 2001). However, this process involves soaking and boiling kernels in a mixture of wood-ash for a period of time, a process that must occur above- ground over an open flame.

If indeed the lack of hearths represents a large-scale shift in foodways practices, implications for daily practice of food preparers would have been substantial. The nixtamalization process requires that maize is boiled for extended periods of time, from

2 to 12 hours (Briggs 2015:120–121; Briggs 2016). If women were cooking in a vessel over an open flame someone may have been required to stay nearby, perhaps reducing the cooks ability to perform other tasks. Understanding “small changes” in food consumption and preparation, specifically the adoption of corn and the transition to hominy preparation, is critical to understanding Mississippian identity. Briggs (2016) argued that identification as a “hominy eater” may have been a pivotal ingredient of

Mississippianization.

146

5.5 A Comparison of the Domestic Yard between Farmsteads and Villages

I examined the spatial distribution of features across the Townsend sites to assess whether the way people organized their space corresponded to the type of settlement they lived in, such as a more isolated farmstead or a village. I assessed the size of the domestic household yard, based on a measure of feature distance (m) to household derived from the Townsend sites ArcGIS database. Only features that could be clearly assigned to a household were considered as part of the domestic space or house yard.

The measure of feature distance to household resulted in two distinct patterns related to spatial location. Histograms of the distance of features to Mississippian households indicate that the household yard size is larger and more variable in farmstead households (Figure 5.9) compared to village households, which are smaller and more narrowly defined (Figure 5.10). For example, the two cooking features associated with a single farmstead were located approximately 18 m away from the structure (Table 5.2).

In contrast, all the cooking features in villages were within 2 m of domestic structures. This pattern remains consistent with storage features as well (Table 5.3). Of the 32 storage features in the two village settings, 23 (72 percent) were within 7 m of a domestic or community structure. In Village 2, all the storage features were very close to households (≤ 7 m) and 13 of the 15 were located between households in an area that could be considered communal. In Village 1, except for the four storage features

147

12 10 8 6

Frequency 4 2 0 0 5 10 15 20 25 30 35 40 More Distance from Closest Household Structure (m)

Figure 5.9. Histogram depicting the distribution of distance of features from the closest household structure in "farmstead" settlement type.

148

10 Frequency 5

0 0 5 10 15 20 25 30 35 40 More Distance to Closest Household Structure (m)

Figure 5.10. Histogram depicting the distribution of distance of features from the closest household structure in "Village" settlement type.

149

Table 5.2. Comparison of Distance of Cooking Features from Associated Households Across Farmsteads and Villages at the Townsend Sites.

Dist. to Storage Feature Household Associated Capacity Community Site Number Morphology Size (m) Household (L) Type 40BT90 1747 basin small 18 D 202 Farmstead 3 40BT90 1749 basin small 17 D 143 Farmstead 3 40BT91 711 basin small 1 K 39 Village 1 40BT91 362 pit med. 1 S 599 Village 2 40BT91 482 basin small 2 Communal 11 Village 2 40BT91 717/718 pit small 0a R 222 Village 2 a 0 m distance to household indicates that the feature was located inside the structure.

150

Table 5.3. Comparison of Storage Feature Distance from Associated Households Across Farmsteads and Villages at the Townsend Sites.

Dist. to Feature Household JLJ Community Site Number Morphology Size (m) A Household B C Type 40BT90 69 pit small 24 D Farmstead 3 212 basin large 0 A Farmstead 1 267 basin large 0 A Farmstead 1 269 pit small 2 A Farmstead 1 459 pit small 4 E Farmstead 4 473 pit med. 11 E Farmstead 4 510 basin small 32 E Farmstead 4 561 basin small 2 G Farmstead 5 580 basin small 1 E Farmstead 4 588 pit large 0 E Farmstead 4 1610 pit large 29 C Farmstead 2 1622 basin small 36 C Farmstead 2 1746 basin small 19 D Farmstead 3 1765 basin small 11 D Farmstead 3 1767 pit small 14 D Farmstead 3 1767A basin small 14 D Farmstead 3 1795 pit med. 22 C Farmstead 2 1825 pit small 3 C Farmstead 2 2120 basin small 13 C Farmstead 2

40BT91 24 basin large 9 O Farmstead 6 67 basin med. 4 P Farmstead 7 71 pit small 7 P Farmstead 7 72 basin large 7 P Farmstead 7 88 basin large 3 Communal Village 2 89 basin small 3 Communal Village 2 92 basin small 2 Communal Village 2 124 pit Xlarge 1 Q Farmstead 8 131 pit med. 18 P Farmstead 7 138 pit large 6 P Farmstead 7 139 basin large 3 P Farmstead 7 140 basin large 5 P Farmstead 7 142 basin small 9 P Farmstead 7 157 basin small 6 P Farmstead 7 343 pit Xlarge 1 Q Farmstead 8 371 pit med. 1 Q Farmstead 8 479 pit small 2 Communal Village 2 480 pit small 2 Communal Village 2 499 basin small 1 Communal Village 2 504 basin small 2 Communal Village 2 507 basin small 3 Communal Village 2 151

Table 5.3 (cont.). Comparison of Storage Feature Distance from Associated Households Across Farmsteads and Villages at the Townsend Sites.

Dist. to Feature Household JLJ Community Site Number Morphology Size (m) A Household B C Type 40BT91 514 basin small 7 Communal Village 2 516 basin small 3 Communal Village 2 517 basin large 2 Communal Village 2 520 basin small 7 Communal Village 2 528 basin med. 7 Communal Village 2 542 pit small 2 V Village 2 582 basin small 6 N Village 1 591 pit small 0 N Village 1 599 basin small 7 M Village 1 623 pit med. 16 L Village 1 629 basin small 1 J Village 1 635 basin small 0 J Village 1 650 basin small 3 H Village 1 672 basin small 0 L Village 1 678 basin small 6 K Village 1 706 basin small 1 V Village 2 A 0 m distance to household indicates that the feature was located inside the structure. A Communal refers to features associated with multiple households. C Household M and N are designated as community buildings rather than domestic structures.

152

associated with the bastions of the palisade, seven of the nine were within 7 m of a domestic structure.

In contrast to the pattern in the two village, there is a great deal of variation in the placement of storage features between the eight farmstead locations at 40BT90 and

40BT91 (Table 5.3). At Farmstead 1, three of the four features are within 2 m of the domestic structure. 40BT91 (Table 5.3). At Farmstead 1, three of the four features are within 2 m of the domestic structure. Alternatively, at Farmstead 2, six of the seven features are at least 13 m or more away from the household.

For both practical and ideological reasons, it is not overly surprising that there is more variation in the placement of features in farmstead communities compared to that expressed in village layouts. First, village areas have restricted and constrained space.

Both Village 1 and Village 2 are bounded by a palisade, which results in limited space for domestic activities and storage. Farmsteads, on the other hand, have much larger living spaces and are not confined to the same extent as villages by other structures or a tightly bound palisade. Second, it is likely that people who chose to be members of a larger village or community were more ideologically constrained than those who lived in more isolated farmstead layouts. Members of communities are usually expected to conform to certain norms of behavior and this undoubtedly applies to the way that space and resources are used.

The feature distance to household analysis also revealed that at both village sites communal storage trumped individual household storage, but each village may have

153

materially expressed this preference in different ways. For instance, in Village 2, despite the presence of 15 storage features, 13 could not be associated with a single household and instead were in communal areas (Table 5.3). Similarly, at Village 1, only six of the

17 storage features could be associated with individual households (Table 5.3). Except for Household J, which had two small storage basins, the other four households had one small storage pit each. The remaining storage features in Village 1 appear to have been for community use; seven are associated with the bastions of the palisade, three are inside or near community buildings, and one was in an isolated area of the complex.

While Village 2 remained invested in subterranean storage, the use of communal corncribs further demonstrates the focus on community storage at Village 1.

5.6 A Comparison of Feature Content

Although the artifacts recovered within a feature are rarely indicative of the primary use of the feature (Schiffer 1987), a comparison of artifact density across features in some instances may contribute to an interpretation of use, or at least should illuminate potential activities performed in the immediate vicinity. At the Townsend sites, I compared the artifact density of ceramics and lithics to examine if there were differences between the features that may help explain their use. Perhaps as consequence of a range of similar activities taking place around them, there were no significant differences in the two most abundant artifact categories, lithic debitage and ceramic sherds, between basins and pits (Figure 5.11). Due to the relatively small

154

quantities of faunal remains recovered from the sites, I did not consider fauna in this analysis. Plant remains are discussed in detail in the following chapter (6). Through this analysis, I determined that just as basins and pits have very similar uses for cooking and storage, there appears to be little difference in their fill patterns that would suggest a specific use for each. Based on this evaluation, basins and pits were combined in the remaining analysis of artifact content based on function.

Besides significant differences in FCR between storage, cooking and hearth features, I found no other differences in the density of artifacts between the functional categories of features (Figure 5.12). As expected, cooking features contained significantly more FCR than other feature categories (Figure 5.8). Illustrated in Figure

5.12, there are contrasting patterns in the density of ceramic sherds and lithic debitage in hearths. Lithic debitage occurred in similar numbers across cooking and storage features, but was very infrequent in hearths. In contrast, and perhaps related to the use of hearths to cook food in vessels over an open fire, sherd density is highest in the hearths.

5.7 A Comparison of the Storage Capacity of Features

Storage capacity of households is often linked to the length of time that a household is occupied, but it may also be one way to interpret differences in status between households (Marcoux 2008). Subterranean storage features are a subtle indicator of status represented as surplus wealth, but are especially important in areas

155

Figure 5.11. Boxplots comparing the density of lithic debitage and ceramic sherds by Mississippian morphological category (the y-axes are scaled logarithmically).

156

Figure 5.12. Boxplots comparing the density of ceramic sherds and lithic debitage by Mississippian functional category (outliers are not represented in entirety within graph).

157

such as the Southern Appalachians where more visible indications of status are often less apparent. This is the case for the Townsend sites, where in the absence of mortuary data, in accordance with the Memorandum of Agreement (MOA), there were few conspicuous signs of status. However, the Townsend sites do demonstrate considerable variation in settlement layout, which serves as an alternative route to investigate differences between settlements and households. Based on an analysis of feature size and distribution, I have previously argued that 1) there is more variation in feature placement in farmsteads than villages, 2) the lack of storage pits in Village 1 may indicate a shift to above-ground storage, and 3) that community storage was more important than household storage in both villages. To further test these hypotheses and to evaluate potential differences in status between these different areas I compare storage capacity between households and settlements.

Of the 22 distinct households, 16 are associated with individual subterranean storage features (Table 4.6). I estimated storage volumes (L) using calculated surface area multiplied by feature depth (cm). Storage capacity was not postulated for the above-ground granaries.

The results of the comparison of storage capacity indicate fundamental differences in storage patterns between people who lived in farmsteads versus those who lived in a village setting. First, there is much more variation in the storage capacity between households that are in farmsteads compared to those in villages (Table 5.4).

Across the eight farmsteads, most of which contain one to two paired domestic

158

structures, storage capacity per household ranged from 0 to 5,182 L, with a standard deviation of 567 L. With the exception of Farmstead 4, which had a communally located corncrib structure, people living in farmstead locations appear to have been relying almost exclusively on subterranean storage, averaging 1,792 L of storage per household.

An above-ground corncrib is associated with Farmstead 4, which may explain the absence of storage features associated with Household F. However, the minimal to no subterranean storage in Household B of Farmstead 1 and Household G of Farmstead 5 may be better explained by a short or limited period of use. An additional explanation is that nearby features lacked diagnostic material and as a result were designated as

“undefined prehistoric” rather than Mississippian.

It is likely that the great range of variation in storage capacity between farmsteads relates to differences in the individual needs of households, such as the number of occupants and composition, as well as the duration of occupation in some cases. Each farmstead or family unit would have had different resource needs, which unregulated by community stores would need to buffer any unexpected pressures. The very low storage capacity of some farmsteads combined with the small number of features seems to suggest that they were occupied for a shorter amount of time.

Farmsteads 5 and 6, for example, have storage capacities of only 58 L and 948 L from single storage features (Table 5.4).

Also demonstrated in Table 5.4, opposite to the patterns seen in the farmsteads of the Townsend sites, the storage capacity of households in village settings is more

159

Table 5.4. List of Storage Capacity of All Households at the Townsend Sites.

Number of Storage Number of Number of Household Domestic Capacity Associated Community Associated A Structures (L) Features Type Corncribs A 2 2098 3 Farmstead 1 0 B 2 0 0 Farmstead 1 0 C 1 1844 5 Farmstead 2 4 D 2 1327 7 Farmstead 3 0 E 1 1646 5 Farmstead 4 1 F 3 0 0 Farmstead 4 1 G 1 58 1 Farmstead 5 0 H 1 187 1 Village 1 1 I 1 0 0 Village 1 1 J 1 539 3 Village 1 1 K 4 73 2 Village 1 0 L 2 473 2 Village 1 0 M 1 51 1 Village 1 4 N 1 198 2 Village 1 4 O 1 948 1 Farmstead 6 2 P 1 4821 10 Farmstead 7 0 Q 2 5182 3 Farmstead 8 0 R 1 222 1 Village 2 0 S 2 599 1 Village 2 0 T 1 0 0 Village 2 0 U 1 189 3 Village 2 0 V 2 94 2 Village 2 1 Communal 5 3287 13 Village 2 0

160

consistent. For instance, the storage capacity of the Village 1 households ranges from 0–

539 L, with an average household capacity of 254.4 L and a standard deviation of 131 L.

At Village 2, there is slightly more variation with storage capacity ranging up to 1,406 L; however, that maximum volume is well below that expressed across farmsteads, which is over 5,000 L. Like Village 1, the average storage capacity of individual households at

Village 2 is relatively small with an average volume of approximately 483 L and a standard deviation of 233 L. Unlike Village 1, there is only one potential corncrib structure associated with the Village 2 palisaded complex; however, its function as a corncrib has been questioned due to its larger size, and it is located on the extreme western edge of the village boundary. In place of the multiple above-ground corncribs that we see at Village 1, at Village 2 the emphasis on community storage is expressed in the centrally located, shared subterranean storage features. Communal storage capacity of subterranean pits varies considerably, from 11 L to 1,526 L. The average storage capacity of community features equals 427 L, with a standard deviation of 512 L.

Based on a comparison of household storage capacities, there are few traces of substantial differences in surplus wealth between the multiple households at Townsend.

This is the case in both the households within each of the villages and compared across the farmsteads. In Village 1, there are no apparent differences in status between the households, where storage capacity is almost equal (Table 5.4). Altogether, instead of storage capacity per households serving as an expression of status, the evidence suggests that at the Townsend sites during the Mississippian occupation, storage is both

161

a concrete symbol of the practical needs of individual farmsteads and a reflection of the choice of people to live in villages—to participate in a community.

5.8 Summary and Conclusions

In this chapter I analyzed the features of the Townsend sites across multiple variables to better understand how the features represent the daily practice of the various occupants of the Townsend site. I compared the shape, size, content, and location of features across the multiple settlements and households at the sites to understand if 1) feature function can be determined by shape, size, and content, 2) the spatial placement or associations between features inform us how people were organizing space and their daily activities, and 3) if differences in storage patterns communicate social meaning, such as ideology or status.

This study determined that few characteristics, other than FCR quantity, are reliable indicators of feature function at the Townsend sites. However, using FCR density to classify features into categories of cooking/cooking refuse, storage, and hearths allowed me to identify patterns in the density and distribution of these facilities that resulted in several important conclusions. First, based on the minimal presence of cooking features across the sites, I suggest that by the Mississippian period in

Tuckaleechee Cove, inhabitants had likely transitioned from an emphasis on subterranean cooking facilities to above-ground, open-flame ceramic pot cooking. The higher density of sherds in hearth features than storage or subterranean cooking features

162

may be further evidence of this trend. Perhaps the low number of hearths at the sites represents a shift to larger above-ground cooking facilities as well. However, the hearths that are associated with a settlement are generally located inside individual households, arguing against communal use.

The distribution, number, and size of subterranean storage facilities across the sites also point to changing patterns in food storage that are tied to the ways that people in different settings organized their space, as well as the ways that they controlled and displayed their resources. The most important factor in determining feature size

(storage capacity) and location was the overarching settlement type—farmstead or village. For inhabitants living within a village, the community dictated that the construction and use of space was much more rigid. Perhaps due to practical or ideological constraints placed on individual households, cooking and storage features were placed in ordered fashion, usually within two to seven meters or inside of the domestic structure. Alternatively, people living in farmsteads appear to have had much more flexibility in the way they organized their space and daily activities. This freedom is displayed in the wide range of locations of both cooking and storage activities, where it was not unusual for features to be placed anywhere from 2 m to 18 m away from the domestic structure.

The results of this study also suggest that although features were not a good indicator of status differences at the Townsend sites in the Mississippian period, collectively they do illuminate some of the different practical and ideological concerns

163

between farmstead and village inhabitants. In farmstead locations, storage capacity is suggestive of the individual needs of families, as well as the duration of site use. Storage was variable and there appears to be little regard for display. In fact, the use of subterranean pits indicates a potential desire to hide surplus goods. Alternatively, in villages the location and storage capacity of features clearly suggests an ideological shift from the household to the community. At Village 2 a large portion of the storage capacity appears to have been shared, mostly in subterranean pits. At Village 1, above- ground granaries seem to have been used communally as an alternative to large individually owned subterranean storage facilities. Located nearby the community stores, a bastioned palisade and two community buildings may further indicate a move to display and protect these stores in a way that was in harmony with this ideology.

Instead of using individual household storage, in most cases at the Townsend sites people in villages appear to have turned to the community to buffer resource needs.

People living at farmsteads and villages in the Mississippian period approached the organization of daily foodways activities very differently, primarily in the way they organized their space and stored their food. In the following Chapters 6 and 7, I use the plant remains as a lens to try to find differences across the sites that are expressed in the more obscure remains of foodways.

164

CHAPTER 6

PLANT REMAINS AND THE MISSISSIPPIAN PEOPLE OF

TUCKALEECHEE COVE

165

6.1 Introduction

Food is one of the primary ways that people express themselves. The rhythm of life in the Mississippian period in Tuckaleechee Cove likely revolved around food production activities–households performing the daily duties of tending the crops, gathering and harvesting, processing, cooking, consuming and sharing meals, and preparing and cleaning for the following day of work. All of this activity takes places in the social milieu; however, in the archaeological record mundane acts are often inseparable from those that can be interpreted as ideological and ritual. Plant remains are often the product of repetitious work, but this work is never mindless and instead is always executed within the constraints of the social structure. Importantly, people perform food production activities in a routine manner, leaving evidence that can relay social information.

In this chapter, I consider the use of plants at the Townsend sites in the

Mississippian period from the botanical analysis of flotation samples from a diverse assemblage of features. This includes features that functioned as storage facilities, cooking pits, and hearths from multiple site areas and households. I qualitatively and quantitatively compare plant use across the sites to identify differences related to feature function, seasonality and phase of occupation, and social circumstance. I further consider the role of corn agriculture at the sites, which is a strong indication of the level of participation of residents of Tuckaleechee Cove in the greater Mississippian farming tradition. 166

6.2 Methods

6.2.1 Field Recovery and Sampling Procedures

All the flotation samples considered in this botanical analysis derive from feature contexts. The recovery of plant remains from the Townsend sites benefited from a robust collection procedure that was initiated in the field excavation of features.

Flotation samples were routinely collected from one-half of each feature in arbitrary 10 cm levels; however, in cases where distinct zones of fill were identifiable, flotation samples were taken by zone and bagged accordingly. Each flotation sample was given a unique Bag Check List (BCL) number that provided secure context information for the sample from field to laboratory and eventually to curation. The volume of flotation samples varied depending on the size of the feature, but if possible excavators collected a standardized 10 L of fill. Ten liters is an amount that is considered appropriate to compensate for preservation inadequacies at open-air sites in the southeastern United

States, which often have moist and acidic soil conditions (Wagner 1988:26).

Archaeological technicians processed a total of 11,375 samples both in the field and laboratory using the technique of water flotation to separate cultural material from the surrounding soil matrix (Driskell 2011:60). There is potential for decreased recovery of wood and fruit parts, but water flotation has been found to greatly improve the recovery rate of seeds in temperate climates (Pearsall 2000; Wagner 1988). Using a modified SMAP-type machine (Watson 1976) allowed for the collection of both heavy

167

fraction (1.64 mm) in the basal screen and light fraction materials with a .05 mm fine mesh, and permitted good recovery of both larger plant materials and small seeds.

I selected flotation samples for botanical analysis based on two primary criteria.

First, I exclusively conducted botanical analysis on a large subset of the features that had been collected from contexts identified as Mississippian by Koerner and Sullivan (2011).

Next, for each Mississippian feature that I included in my analysis I selected the context

(BCL) with the highest estimated botanical percentage as determined by the ARL microartifact analysis team. This initial percentage was extrapolated from a count of natural and cultural material from the heavy fraction of each sample that represented a statistically significant subsample of the dataset as determined by MMCount, a statistical program for microartifact analysis (Driskell 2011:62). Selecting the sample with the statistically highest percentage of plant remains provided the best opportunity to obtain a representative sample of plants from each of the individual features.

6.2.2 Identifications of Plant Remains

I performed botanical analysis of 111 flotation samples recovered from 88

Mississippian features at the Townsend sites (Table 6.1). These include 21 flotation samples from Hollenbach’s (2011) analysis. My analysis covered eight flotation samples from six features at 40BT89, 34 flotation samples from 24 features at 40BT90, and 69 flotation samples from 58 features at 40BT91. All of the flotation samples in this study were analyzed using standard paleoethnobotanical procedures as outlined by Pearsall

168

(2000). One advantage of the modern age of the site excavation was that all of the samples, including those analyzed by Hollenbach (2011), were processed and identified using comparable methods. Each sample was weighed to the nearest 0.01 g. Following weighing, each sample was filtered through nested brass sieves to size-grade the material into the following units that could be scanned more effectively: > 2.0 mm, > 1.4 mm–2.0 mm, > 0.07–1.4 mm, and > 0.05–0.07 mm.

All materials greater than 2.0 mm from Hollenbach’s (2011) samples were sorted into several categories of bone, shell, lithics, and plant material and then weighed. This was done to isolate other categories of artifacts to facilitate other microartifact analyses that were being conducted at the time. Because no further microartifact analyses were being conducted for this dissertation, I focused on botanical material from the ≥ 2.0 mm sieve and considered all other cultural materials as contaminants. I used a stereoscopic microscope at 10 to 40 power magnification to separate any of the plant remains that were pulled from the greater than 2.0 mm sieve into the lowest taxonomic category possible and where definitive identifications could not be made the items were grouped as unidentifiable. The material less than 2.0 mm was scanned for smaller material that was not caught by the larger sieve, such as small seeds, as well as for any material that was not represented in the upper sieve. Because acorn is very fragile and easily fragments, it was pulled from both the 2.0 and 1.4 mm sieve. All plant taxa were then counted and weighed. Because of the insubstantial mass of many of the plant materials there are many occurrences of a 0.00 g recording for plant masses. However, for each

169

taxon, except for wood which fragments easily, I also recorded a count to aid in direct comparisons of plant material.

6.2.3 A Consideration of Biases within the Dataset

There are many factors that influence the recovery of plant materials from archaeological sites that need to be considered to assess the accuracy of qualitative and quantitative interpretations of plant material from individual sites. First, plants are composed of a variety of different materials— the soluble parts of the living cell which are subject to rapid decomposition, and the more resistant parts such as the cell walls and various substances that act to give strength to individual plant parts (Dimbleby

1967:95). Without human intervention, dead plant material is immediately subject to the processes of decay including comminution, digestion, and the loss of soluble substances, but through unique natural and human processes the process of decay can be interrupted. For example, very wet, dry, and compacted soil conditions lack either the oxygen or the water needed for micro-organic decomposition of plant remains

(Dimbleby 1967:96; Miksicek 1987:212). However, at most archaeological sites in the temperate environments of the southeastern United States the primary way that plant remains are preserved is by carbonization in fires. Carbonization occurs when the plant’s organic material is converted into elemental carbon, leaving the plant resistant to microbiologically-induced decay (Dimbleby 1967:100).

170

Mississippian people routinely used plants culturally, including as fuel and food, and through these uses, enhanced their potential for preservation. We can generally assume that the more frequently a plant was used, the more often we should recover that plant in archaeological contexts (Mikisicek 1987). But natural and cultural processes bias the recovery of carbonized plant remains. Many plant taxa including fruits, greens, and seeds are consumed fresh and therefore are never carbonized or preserved. These taxa and their importance are often underrepresented in the archaeological record. On the other hand, taxa that are largely inedible, or constitute refuse that function as a fuel source are frequently introduced to fire and may be overrepresented. Numerous studies have also detailed the importance of other factors including the species and part of the plant, the plant’s physical condition, and the thermal conditions of the fire that also play a role in plant preservation (see Wright 2003:583 for a review of experimental approaches). Wright (2003) cautioned against assuming the frequency of plant use is equivalent to importance and instead suggested a thorough consideration of differential survival rates of plants in fires.

There are inherent difficulties of using incomplete material remains to interpret past human behavior, but this is a challenge that is present across the archaeological endeavor. There are many ways to ensure adequate recovery and comparability of the paleoethnobotanical dataset. To begin, a robust sampling procedure in the field—taking standardized samples of an adequate size from each context—permits meaningful quantitative comparisons and interpretations (Fritz 2005; Pearsall 2000). With few

171

exceptions, fine-mesh recovery of plant remains through the process of water flotation is the most effective way to recover a complete assortment of all size classes. The best way to develop accurate interpretations from plant remains is to be aware of both site-wide problems that may affect the interpretation of contexts such as mixed deposits and variable depositional behavior (Fritz 2005:792), and to address the potential biases of the dataset using appropriate analytical methods.

6.2.4 Statistical Analyses Employed for Interpretation

I apply statistical ratios and indices to determine potential differences between features based on their function and to assess the relative significance of plant taxa over time and across different areas of the Mississippian occupation of the Townsend sites.

Ubiquity, a measure of the percentage of samples in which a taxon is present, is a good indicator of large-scale trends in plant use that occur over time or across space (Fritz

2005:792).

I also use density and relative density ratios to examine differences across the sites. Ratios allow comparisons of material by standardizing the sample by the denominator. In the case of density, the weight or count of a particular taxon or category of plants is divided by the total volume of the sample, which allows for a comparison of samples of different volumes. These data can aid in an interpretation of the changing importance of certain taxa over time, as well as of potential feature function. Relative density standardizes the count or weight of a plant taxon by the total

172

plant weight of a sample. Counts of taxa can also be used to allow for comparisons of plants that have disproportionate weights. Therefore, this measure avoids erroneous conclusions of importance of a plant taxon, by considering how that plant’s abundance compares to other plants in the sample (Fritz 2005; Pearsall 2000; Popper 1988). For instance, if one taxon is very abundant in a sample compared to the other plant material, it is possible to suggest a unique function for that context. If this pattern holds consistently over many samples, this potentially indicates that a certain taxon is more important in one period or for a particular context. Even so, it is still difficult to directly compare taxa because of preservation bias.

In addition to measuring the relative density of both corn kernels and cupules, I use another ratio, a comparison of corn kernel count to corn cupule and glume count, to assess potential areas that may have been used more for consumption as suggested by a high kernel ratio, or for corn processing as indicated by a low kernel ratio (Welch and

Scarry 2005; VanDerwarker and Wilson 2016). By doing so, I hope to highlight any possible social differences in food-related activities, whether they are produced along the lines of gender, kinship, or household.

6.3 Paleoethnobotanical Analysis Results

The combined samples resulted in 770.70 g of plant material, of which wood represents 462.96 g, that were recovered from 1,275 total L of feature fill. The plant remains from the Townsend sites are composed of a variety of wild and domesticated

173

plants that were either available locally in Tuckaleechee Cove or that were planted seasonally in garden plots or agricultural fields. They include nuts, wild fruits, and indigenous cultigens and crop species. The various miscellaneous taxa from the sites, especially a variety of weedy species, also hint at the composition of and human impacts on the surrounding landscape. Forty taxa are represented in the flotation samples, including 21 belonging to a variety of plants that could have been consumed and 19 miscellaneous taxa. These taxa, along with total counts and weights, are reported in

Table 6.2. All plant taxa recovered by feature context across the sites are available in

Table A.6.

By far, nuts comprise the largest component of the plant assemblage across the

Townsend sites in weight, number, and frequency (Table 6.2). The nut category is comprised almost exclusively of acorn, hickory, and black walnut, along with infrequent occurrences of hazelnut and a possible beech fragment. Hickory is almost ubiquitous across the 88 features (Table 6.3), occurring in all but three. Not surprisingly, it is the most numerous by count and weight and contributes to almost half of the total plant weight excluding wood. Hickory nutshell was relatively evenly distributed across the samples; however, 41 percent came from two features at site 40BT90 including from three flotation samples of Feature 697 and six flotation samples of Feature 1795.

Acorn is also present in 70 percent of the features from the combined Townsend sites. A total of 4,672 fragments of acorn shell were recovered in total; however, most of

174

Table 6.1. Analyzed Mississippian Flotation Samples from the Townsend Sites.

Site/ BCL TCA Community Household Function Sample Sample Contam. Residue Plant Wood Feature Number Type Volume Weight Weight Weight Weight Weight Number A (L) (g) (g) (g) (g) (g) 40BT89 25 99-1094-4 MSM Storage 10 14.94 5.08 5.46 2.22 0.54 25 99-1118-6 MSM Storage 10 54.22 6.50 29.23 14.51 4.05 25 99-1126-89 MSM Storage 10 7.34 3.75 3.17 1.00 0.43 89 99-742-253 MSM Storage 10 11.58 3.18 4.23 3.53 2.95 124 99-1197-255 MSE Storage 8 21.93 9.55 6.85 5.33 4.04 127 99-563-1 MSM Storage 10 7.04 4.54 2.33 0.17 0.15 164 99-211-5 MSE Storage 10 29.73 11.85 13.91 1.67 0.01 251 99-1462-1 MSE Storage 10 112.62 58.13 48.76 5.32 1.68 40BT90 10 99-364-1 MSS Farmstead 1 Cooking 10 56.24 25.38 27.96 2.64 2.26 69 99-352-1 MSE Farmstead 3 D Storage 10 236.17 86.52 126.69 21.87 16.93 106 99-9003-1 MSM Hearth 10 96.86 44.70 43.39 8.48 5.36 212 99-1614-1 MSE Farmstead 1 A Storage 10 43.65 12.87 14.35 3.99 2.75 267 99-1703-1 MSE Farmstead 1 A Storage 8 15.95 4.95 2.60 1.46 0.83 269 99-1852-1 MSM Farmstead 1 A Storage 10 52.16 23.36 25.52 0.55 0.25 580 00-289-1 MSS Farmstead 4 E Storage 10 24.20 2.44 9.72 10.24 10.09 697 00-821-252 MSM Farmstead 1 Cooking 14 202.80 34.05 91.71 35.76 21.99 697 00-846-251 MSM Farmstead 1 Cooking 10 33.69 13.22 13.55 1.30 0.37 697 00-872-2 MSM Farmstead 1 Cooking 10 560.65 85.86 302.32 127.27 109.38 873 99-3226-251 MSS Storage 10 58.11 33.72 22.31 2.03 0.92 873 99-3308-256 MSS Storage 10 38.54 21.77 15.73 0.93 0.72 1531 99-4298-251 MSM Farmstead 2 Hearth 5 57.94 7.42 11.64 4.13 2.18 1531 99-4323-252 MSM Farmstead 2 Hearth 20 46.83 15.89 29.90 0.96 0.83 1531 99-4347-1 MSM Farmstead 2 Hearth 8 19.12 7.98 10.14 0.33 0.23 175

Table 6.1 (cont.). Analyzed Mississippian Flotation Samples from the Townsend Sites.

Site/ BCL TCA Community Household Function Sample Sample Contam. Residue Plant Wood Feature Number Type Volume Weight Weight Weight Weight Weight Number A (L) (g) (g) (g) (g) (g) 1610 99-9843-251 MSE Farmstead 2 C Storage 10 385.40 218.92 158.96 6.93 0.76 1746 99-4597-1 MSE Farmstead 3 D Storage 10 60.84 31.08 26.50 1.24 0.69 1747 99-5997-1 MSE Farmstead 3 D Cooking 10 18.67 7.57 6.72 1.54 0.39 1749 99-4782-1 MSE Farmstead 3 D Cooking 10 52.01 16.80 31.40 3.51 2.77 1765 99-5649-252 MSS Farmstead 3 D Storage 10 27.19 13.87 10.86 2.34 1.37 1767 99-5824-1 MSS Farmstead 3 D Storage 10 45.05 15.54 24.05 4.59 1.04 1795 99-5527-256 MSM Farmstead 2 C Storage 10 129.10 42.42 71.55 15.13 11.38 1795 99-5528-1 MSM Farmstead 2 C Storage 10 92.84 33.59 49.13 10.13 5.45 1795 99-5529-1 MSM Farmstead 2 C Storage 10 139.87 41.85 62.44 36.36 14.71 1795 99-5531-257 MSM Farmstead 2 C Storage 6 168.33 56.07 83.29 28.91 13.06 1795 99-5532-255 MSM Farmstead 2 C Storage 10 120.94 53.58 54.27 12.84 6.42 1795 99-5533-1 MSM Farmstead 2 C Storage 10 64.30 26.40 33.71 3.95 2.89 1825 99-4350-1 MSS Farmstead 2 C Storage 20 54.40 22.24 31.33 0.69 0.34 1851 99-5750-1 MSS Cooking 10 4.65 0.84 3.06 0.76 0.63 1856 99-5570-253 MSS Cooking 10 21.77 6.80 14.27 0.51 0.35 2001 99-9014-2 MSM Storage 10 106.07 35.28 51.65 18.29 12.02 2002 99-9240-3 MSM Storage 10 55.06 15.31 27.01 6.20 3.77 2120 99-11062-1 MSS Farmstead 2 C Storage 2 99.23 1.62 28.99 58.86 1.44 1767A 99-5827-1 MSS Farmstead 3 D Storage 10 52.17 20.18 25.06 6.84 3.17 40BT91 13* 99-79-254 MSM Village 1 Storage 20 156.65 79.63 68.01 4.55 0.14 13* 99-82-1 MSM Village 1 Storage 20 287.81 114.17 124.26 27.20 18.50 13* 99-83-258 MSM Village 1 Storage 20 86.99 35.24 24.47 11.66 6.71 13* 99-84-254 MSM Village 1 Storage 20 57.86 22.14 20.41 9.09 5.89 37 99-603-1 MSE Storage 10 67.20 45.10 21.80 0.14 0.14 176

Table 6.1 (cont.). Analyzed Mississippian Flotation Samples from the Townsend Sites.

Site/ BCL TCA Community Household Function Sample Sample Contam. Residue Plant Wood Feature Number Type Volume Weight Weight Weight Weight Weight Number A (L) (g) (g) (g) (g) (g) 37 99-604-2 MSE Storage 10 23.76 21.79 1.54 0.36 0.27 48 99-491-253 MSE Hearth 8 88.04 56.78 29.80 1.45 0.61 68 99-482-1 MSS Farmstead 7 P Hearth 2 28.23 22.08 6.02 0.00 0.00 92 99-99-1 MSM Village 2 Z Storage 10 8.70 5.86 2.16 0.62 0.33 124 99-164-3 MSE Farmstead 8 Q Storage 10 137.19 111.73 22.88 2.21 1.83 131 99-357-252 MSE Farmstead 7 P Storage 6 32.41 10.12 12.65 9.07 0.65 138 99-210-4 MSE Farmstead 7 P Storage 10 8.80 4.65 2.95 0.70 0.17 138 99-221-1 MSE Farmstead 7 P Storage 10 103.76 19.82 46.05 28.92 16.75 139 99-215-4 MSE Farmstead 7 P Storage 10 32.08 8.64 17.00 6.19 4.67 140 99-342-4 MSM Farmstead 7 P Storage 10 27.56 12.69 7.45 6.21 4.02 142 99-236-1 MSM Farmstead 7 P Storage 10 9.76 2.98 4.97 1.71 1.18 157 99-338 MSM Farmstead 7 P Storage 10 11.56 5.48 4.33 1.63 1.05 268 99-546-1 MSS Storage 10 119.50 109.97 9.02 0.28 0.07 273 99-758-1 MSS Storage 4 14.72 5.01 9.61 0.08 0.06 274 99-719-8 MSM Storage 77 330.60 96.60 51.72 12.49 9.05 343 99-2953-1 MSE Farmstead 8 Q Storage 10 123.19 103.57 19.32 0.18 0.06 371 99-2942-4 MSE Farmstead 8 Q Storage 10 65.31 49.03 12.26 3.96 3.83 479 99-2116-3 MSS Village 2 Z Storage 10 16.39 0.67 10.38 0.86 0.25 480 99-2240-3 MSM Village 2 Z Storage 10 11.56 5.38 5.19 0.93 0.68 482 99-2177-3 MSS Village 2 Z Cooking 10 39.98 3.47 27.34 8.95 7.37 499 99-2190-4 MSS Village 2 U Storage 10 28.67 7.45 15.14 6.05 5.11 504 99-2249-251 MSS Village 2 U Storage 10 26.53 6.06 19.42 0.99 0.27 507 99-2215-2 MSS Village 2 Z Storage 6 7.40 1.20 4.57 1.62 0.36 514 99-2241-3 MSS Village 2 Z Storage 10 21.88 7.51 10.77 3.60 0.75 516 99-2470-1 MSE Village 2 Z Storage 10 12.25 3.12 4.88 0.93 0.60 177

Table 6.1 (cont.). Analyzed Mississippian Flotation Samples from the Townsend Sites.

Site/ BCL TCA Community Household Function Sample Sample Contam. Residue Plant Wood Feature Number Type Volume Weight Weight Weight Weight Weight Number A (L) (g) (g) (g) (g) (g) 517 99-2478-5 MSE Village 2 Z Storage 10 13.65 3.31 8.80 0.31 0.16 517 99-2479-252 MSE Village 2 Z Storage 10 10.72 6.00 3.63 0.92 0.63 517 99-2638-253 MSE Village 2 Z Storage 10 29.65 6.24 20.45 0.80 0.28 520 99-2257-2 MSE Village 2 Storage 10 6.30 2.51 3.13 0.62 0.30 528 99-2303-1 MSE Village 2 Storage 10 8.87 5.46 2.05 1.31 0.73 530 99-2703-1 MSE Village 2 V Hearth 10 54.08 30.83 22.12 1.00 0.37 542 99-2081-1 MSM Village 2 V Storage 5 8.13 4.87 2.89 0.34 0.12 544 99-2361-5 MSM Village 2 V Hearth 10 54.56 10.14 40.53 3.46 2.64 581* 99-1329-251 MSM Village 1 Storage 10 4.15 1.63 1.93 0.58 0.46 582** 99-122-3 MSM Village 1 N Storage 10 57.75 19.62 30.57 7.33 4.63 582** 99-1217-2 MSM Village 2 N Storage 10 39.58 11.39 9.07 8.05 6.28 584* 99-1361-6 MSM Village 1 Storage 10 131.45 45.24 68.13 7.49 2.54 584* 99-1374-7 MSM Village 1 Storage 15 101.47 31.92 55.06 4.92 1.93 585* 99-1213-1 MSM Village 1 Storage 10 17.11 4.49 11.77 0.75 0.56 591** 99-1370-251 MSS Village 1 N Storage 10 20.75 6.78 13.10 0.73 0.41 599** 99-1209-1 MSS Village 1 M Storage 10 22.70 8.16 8.67 5.51 0.21 623 99-1346-3 MSS Village 1 L Storage 10 21.68 9.58 12.37 0.52 0.44 623 99-1364-2 MSS Village 1 L Storage 10 50.42 17.71 31.95 0.55 0.44 629 99-2016-4 MSS Village 1 J Storage 70 478.29 111.87 122.67 5.85 4.24 631 99-1965-1 MSS Village 1 Cooking 10 39.30 18.48 19.42 1.36 1.07 632 99-2015-4 MSM Village 1 J Hearth 10 109.28 18.26 62.48 20.89 19.28 635 99-2020-2 MSM Village 1 J Storage 10 168.84 57.36 104.43 6.55 4.16 650 99-1936-3 MSS Village 1 H Storage 10 49.84 11.65 37.65 0.41 0.29 672 99-1871-251 MSS Village 1 L Storage 10 73.29 34.45 35.56 3.21 2.76 678 99-2556-2 MSS Village 1 K Storage 10 50.07 25.52 24.15 0.31 0.22 178

Table 6.1 (cont.). Analyzed Mississippian Flotation Samples from the Townsend Sites.

179

Table 6.2. Plant Remains from Mississippian Contexts at the Townsend Sites.

Weight Common Name Count Seasonality (g) Nuts: Acorn Quercus spp. 4672 9.20 fall Acorn cap Quercus spp. 104 0.29 fall Acorn cap cf. Quercus spp. cf. 12 0.02 fall Acorn cf. Quercus spp. cf. 22 0.01 fall Acorn meat Quercus spp. 22 1.21 fall Acorn meat cf. Quercus spp. cf. 21 0.27 fall Beech cf. Fagus grandifolia cf. 1 0.00 fall Black walnut Juglans nigra 316 5.76 fall Black walnut cf. Juglans nigra cf. 25 0.20 fall Hazelnut Corylus spp. 9 0.01 fall Hickory Carya spp. 11606 140.48 fall Hickory cf. Carya spp. cf. 10 0.07 fall Hickory husk Carya spp. 1 0.00 fall Hickory husk cf. Carya spp. cf. 12 0.12 fall Unidentifiable nutshell 2 0.00 fall Unidentified nutshell 1 0.00 fall Walnut family Juglandaceae 1207 7.81 fall Fruits: Blackberry/raspberry Rubus sp. 7 0.02 summer Blackberry/raspberry cf. Rubus sp. cf. 1 0.00 summer Blueberry Vaccinium sp. 3 0.00 summer Grape Vitis sp. 8 0.06 summer Grape cf. Vitis sp. cf.. 1 0.00 summer Maypop Passiflora incarnata 15 0.07 summer Maypop cf. Passiflora incarnate cf. 1 0.00 summer Maypop uncarbonized Passiflora incarnata 4 0.04 summer Pawpaw Asimina sp. 5 0.22 summer/fall Persimmon Diospyros virginiana 39 0.59 fall Persimmon cf. Diospyros virginiana cf. 13 0.05 fall Persimmon seed cf. Diospyros virginiana cf.. 8 0.00 fall Plum/cherry Prunus sp. 2 0.00 summer Strawberry cf. Fragaria sp cf. 1 0.00 summer Sumac Rhus sp. 2 0.01 fall Unidentifiable fruit 29 0.24 Starchy/Oily Seeds: Bearsfoot Smallanthus uvedalius 5 0.00 summer/fall Chenopod Chenopodium berlandieri 28 0.01 late summer/fall Chenopod cf. Chenopodium berlandieri cf. 3 0.00 late fall/summer

180

Table 6.2 (cont.). Plant Remains from Mississippian Contexts at the Townsend Sites.

Weight Common Name Count Seasonality (g) Starchy/Oily Seeds: Maygrass Phalaris caroliniana 3 0.00 spring/early summer Maygrass cf. Phalaris caroliniana cf. 2 0.00 spring/early summer Sumpweed Iva annua 4 0.00 late summer/fall Sumpweed cf. Iva annua c.f. 2 0.00 summer/fall Sunflower Helianthus annuus 1 0.00 late summer/fall Sunflower cf. Helianthus annuus cf. 2 0.00 late summer/fall Crops: Bean Phaseolus vulgaris 5 0.02 late summer/fall Bean cf. Phaseolus vulgaris cf. 2 0.00 late summer/fall Corn cob frag Zea mays 122 18.27 late summer/fall Corn cupule Zea mays 1691 15.18 late summer/fall Corn cupule cf. Zea mays cf. 40 0.13 late summer/fall Corn cupule/glume Zea mays 3212 38.05 late summer/fall Corn cupule/glume cf. Zea mays cf. 4 0.01 late summer/fall Corn embryo Zea mays 4 0.00 late summer/fall Corn glume Zea mays 139 0.46 late summer/fall Corn glume cf. Zea mays cf. 2 0.00 late summer/fall Corn kernel Zea mays 818 6.20 late summer/fall Corn kernel cf. Zea mays cf. 61 0.22 late summer/fall Cucurbit rind Cucurbitaceae 13 0.04 late summer/fall Cucurbit rind cf. Cucurbitaceae cf. 5 0.01 Cucurbit seed Cucurbitaceae 2 0.01 late summer/fall Tobacco Nicotiana sp. 3 0.00 Tobacco cf. Nicotiana sp. cf. 1 0.00 Tobacco uncarbonized Nicotiana sp. 1 0.00 Miscellaneous: Aster family Asteraceae 2 0.00 Bark 192 1.80 Bark cf. 184 1.47 Bark/pine cone 1 0.01 Bean/hickory Phaseolus/Carya 2 0.03 Bedstraw Galium sp. 4 0.02 Bedstraw cf. Galium sp. cf. 6 0.00 Black walnut/peach Juglans/Prunus persica 9 0.16 Buckwheat family Polygonaceae 3 0.01 Bud 9 0.04 Bud cf. 1 0.00

181

Table 6.2 (cont.). Plant Remains from Mississippian Contexts at the Townsend Sites.

Weight Common Name Count Seasonality (g) Miscellaneous: Bulrush cf. Scirpus sp. cf. 12 0.01 Cane Arundinaria sp. 122 0.97 Carpetweed Mollugo sp. 11 0.00 Catkin 1 0.00 Cheno/am Chenopodium/Amaranthus 4 0.00 late summer/fall Chickweed Stellaria sp. 1 0.00 Dogwood cf. Cornus canadensis cf. 6 0.04 Gall 59 0.07 Grass family Poaceae 7 0.00 Grass family cf. Poaceae cf. 11 0.01 Hickory/cupule 34 0.10 Knotweed Polygonum sp. 6 0.00 Legume cf. Fabacea c.f. 1 0.01 Magnolia Magnolia grandiflora 1 0.01 Monocot stem Poaceae 6 0.04 Monocot stem cf. 1 0.09 Morninglory Ipomoea/Convolvulus 2 0.00 summer/fall Nightshade cf. Solanum sp. cf. 1 0.00 Pine cone Pinus spp. 799 7.34 Pine cone cf. Pinus spp. cf. 4 0.01 Pine seed cf. Pinus spp. cf. 1 0.00 Pink family Caryophyllaceae 1 0.01 Pitch 5646 46.31 Pocket gall 62 0.30 Pokeweed Phytolacca americana 1 0.00 summer/fall Purslane Portulaca sp. 28 0.00 summer/fall Ragweed Ambrosia sp. 15 0.01 Ragweed cf. Ambrosia sp. cf. 9 0.00 Receptacle 1 0.00 Smartweed Polygonum pensylvanicum 5 0.00 Spurge family Euphorbiaceae 12 0.00 Yellow Stargrass Hypoxis sp. 3 0.00 Stargrass cf. Hypoxis sp. cf. 2 0.01 Stem 30 0.33 Thorn 1 0.00 Twig 11 0.19 Unidentifiable 1584 8.08 Unidentifiable seed 235 0.15 Unidentifiable seed cf. 5 0.02 Unidentified 3 0.02 182

Table 6.2 (cont.). Plant Remains from Mississippian Contexts at the Townsend Sites.

Weight Common Name Count Seasonality (g) Miscellaneous: Unidentified seed 108 0.07 Walnut family/corn cupule 2 0.01 Weedy legume Fabaceae 3 0.00 late summer/fall

183

Table 6.3. Ubiquity of Plant Remains Recovered from 88 Features Across the Townsend Sites.

Category Common Name Scientific Name Percentage Nuts Hickory Carya spp. 97 Acorn Quercus spp. 70 Black Walnut Juglans nigra 43 Hazelnut Corylus spp. 5

Crops Corn Zea mays 67 Cucurbit Cucurbitaceae 8 Bean Phaseolus vulgaris 3 Tobacco Nicotiana sp. 2

Fruits Persimmon Diospyros virginiana 7 Blackberry/raspberry Rhubus sp. 5 Grape Vitis sp. 5 Maypop Passiflora incarnata 5 Blueberry Vaccinium sp. 3 Pawpaw Asimina sp. 2 Sumac Rhus sp. 2 Plum/cherry Prunus sp. 1

Smallanthus Starch/Oily Seeds Bearsfoot 5 uvedalius Chenopodium Chenopod 6 berlandia Maygrass Phalaris caroliniana 3 Sumpweed Iva annua 2 Sunflower Helianthus annuus 1

184

the features have relatively small amounts of shell equaling less than 10 fragments. In fact, 86 percent of the acorn shell recovered came from just nine features across the site complex, one from 40BT89, four from 40BT90, and four from 40BT91. All of the features that contained large amounts of acorn shell were categorized for storage, with the exception of one cooking and one hearth feature. Features that contained large amounts of acorn shell, especially Feature 697 which contained 2,695 fragments, are likely the result of processing activities. Most of the high-density acorn features also had large amounts of hickory that were recovered as well. All of the acorn meats came from two flotation samples from Feature 582, a storage feature affiliated with Community

Building N in Village 1.

A variety of fruits were identified from the flotation samples, representing eight taxa (Table 6.2). This category includes both summer-and fall-ripening fruits including blackberry/raspberry, blueberry, grape, maypop, plum/cherry, and strawberry and pawpaw, persimmon, and sumac. Interestingly, Feature 711, a cooking pit associated with Household K and located in Village 1 at 40BT91, had a very large assortment and number of fruit taxa (Table A.6). A total of four taxa including blackberry/raspberry, grape, pawpaw, and persimmon were present, along with an assortment of what appeared to be carbonized fragments of unidentifiable fleshy fruit material. The contents of this feature may represent an unintentional act of spilling.

Corn was recovered from 67 percent of the features across the Townsend sites, and second to nutshell is the most abundant food remain (Table 6.3). It is represented by

185

5,986 combined cob fragments, cupules, glumes, embryos, and kernels weighing 78.16 g,

17 percent of the total plant weight excluding wood. Besides corn, several other crops were identified in small amounts. These include 14 cucurbit rind and seed fragments from seven features, and four bean fragments from three features.

I identified very few edible seeds across the sites. The edible seed category is represented by limited numbers and occurrences of chenopod (Chenopodium berlandieri), maygrass (Phalaris caroliniana), bearsfoot (Polymnia uvedalia), sumpweed (Iva annua), and sunflower (Helianthus annuus). Chenopod was the most abundant with 27 seeds recovered from four features; however, 22 of those came from a single feature, Feature

139, a storage feature associated with Farmstead 7 that may have a Pisgah affiliation. Of note, out of the five features that contained edible seed crops at 40BT91, two are associated with the bastioned palisades of Village 1, Features 13 and 585. In addition to the edible crop species at Townsend, three tobacco seeds were identified, all of which were identified from features that were located inside the bastioned palisade at Village

A variety of taxa belonging to the miscellaneous category were recovered including material that represents the local habitat, as well as plants with potential medicinal, dietary, and economic uses (Table 6.2). Purslane, ragweed, spurge, and carpetweed are the most common of the weedy species, but there are also occurrences of knotweed and smartweed, weedy grasses and legumes, and sedges, bedstraw, and morning-glory. Contributing to the potential ritualistic function of Feature 13 which is

186

associated with the bastioned palisade of Village 1, morning-glory, several spurges, and purslane were recovered alongside the two tobacco seeds.

6.4 Interpretations of Mississippian Plant Use in Tuckaleechee Cove

I begin with an overview of general plant use at the site complex to try to understand how plants contributed to the everyday life of a Mississippian person in

Tuckaleechee Cove. Which plants were relied upon the most? Which food activities did people invest in, did they gather or farm, and how were their efforts likely divided?

What was the seasonal pattern of activities? By the Mississippian period many communities were highly invested in corn agriculture, but there are many instances of different foodways that were less dominated by corn in peripheral geographic areas. I follow with an examination of plant use over time in the Cove. In doing so, I look for changes in the use of plants over time to evaluate whether these communities experienced a level of consistency or unpredictability in their foodways. A large portion of this discussion is spent comparing the plant remains from the numerous feature contexts across the site complex. First, I identify several of the features from the site complex that are noticeably different from their contextual counterparts. Second, I compare the plant remains by context quantitatively over multiple variables, including

1) function, 2) size, and 3) location, to identify the source of the variation, whether it is a function of the type of feature, the differences in the dimensions and storage capacity, or if it is related to space and the greater ideology of the settlement that one lives within.

187

It is probable that a variety of community members other than women—men, elders, and children—participated in some gathering and field preparation activities.

Based on ethnographic accounts of maize production in the Southeast, it appears that men likely helped clear and burn fields, while women bore this as well as the additional horticultural workload including the planting, tending, harvesting, and processing

(Scarry 2008; Scarry and Scarry 2005). Ethnographic accounts of the Cherokee describe a division of labor in which women were aided by men during intensive times of planting and harvesting, but had complete responsibility at other times (Perdue 1998:20). While there is a lack of information on pre-maize farming techniques, the general assumption is that the gendered division of labor we see in three sisters’-style or corn, beans, and squash based agriculture was a continuation of earlier hunting and gathering and horticultural practices. Nonetheless, I proceed with the understanding that women were presumably responsible for coordinating and/or performing gathering and farming activities, and view the management of plants within this gender-based scenario (Fritz

1999; Perdue 1998; Scarry and Scarry 2005; Watson and Kennedy 1991). In the following sections, I discuss the occurrences of and the potential uses of each of the categories of plant remains recovered from this botanical analysis.

6.4.1 Overview of Plant Categories Commonly Found at the Sites

Each of the plant categories, the nuts, crops, fruits, edible seed crops, and miscellaneous, that I recovered from the Mississippian occupation were important

188

resources that likely served occupants in a variety of ways. Many of the items were important practically for the health and maintenance of the community—as food, fuel, and construction materials. Others would have also unquestionably been imbued with social and spiritual meaning.

6.4.1.1 Mast Resources

Hickory trees would have been common in the drier slopes surrounding the sites and likely produced high yields. It is likely that occupants of the Cove intensified collection strategies in the peak months of October and November because of the nuts’ limited period of availability due to competition with other forest animals (Munson

1984:462; Scarry 2003:60; Talalay et al. 1984:345). Hickory nuts are frequently found in large numbers in botanical assemblages across the Eastern Woodlands, a testament to their natural abundance and high yields, nutritional value which is high in fat content, amino acids, calories, easy storability, and ability to be processed rather efficiently into an oil (Scarry 2003:60–61). Hickory nutshell appears to have been a staple resource in the Mississippian period, but I do not want to overestimate its importance over other taxa. Several factors other than use likely contributed to its high recovery at these sites, including the presence of a dense, and thus resilient outer shell and the large amount of inedible by-products that were a convenient fuel source (Fritz 2003:791–792).

Acorn, like many other mast resources, is available seasonally in the fall, most commonly from the months of October and November. Like hickory, oak trees would

189

have been readily available in both the open to mid-slope and gravelly upland environments surrounding the site area (Petruso and Wickens 1984:363). In the Eastern

Woodlands region, acorn has consistently been recovered in large numbers from archaeological sites since the Early Archaic period, a trend that indicates it was a dependable dietary staple (Yarnell and Black 1985).

6.4.1.2 Crops and Farming

By the Mississippian period in Tuckaleechee Cove, inhabitants had adopted a farming way of life centered around corn agriculture. Based on macrobotanical remains, maize was introduced into the Eastern Woodlands between 0 and 200 CE. (Chapman and Crites 1987; Fritz 1987, 1993:52; Reilly et al. 1994). Corn macroremains have been recovered in small numbers from several Middle Woodland sites in East Tennessee, including Icehouse Bottom (Chapman and Crites 1987) and the Townsend sites

(Hollenbach and Yerka 2011:374). Interestingly, both sites have a connection to

Hopewell, which may suggest the plants’ early ritualistic role (Fritz 1993:56;

Johannessen 1993b:74-75; Scarry 1993:90). However, by the Late Woodland and into the

Mississippian period, corn transformed foodways, economies, and social systems across the Southeast.

Although corn may have supplanted acorns as a source rich in carbohydrates during the Mississippian period (Yarnell and Black 1985:97), work in peripherally located sites in western and southeastern Kentucky, the Big Black River Valley, and the

190

Appalachian Summit region indicates that inhabitants continued to rely heavily on nuts, which comprised up to 90 percent of the edible plant assemblage at some sites (Dickens

1976; Jeffries et al. 1996; Lorenz 1992:298–393). Since the Townsend sites are also located in a peripheral geographic location, between two distinct cultural traditions of Hiwassee

Island and Pisgah, I evaluate how the nutshell:maize ratio based on counts compares to other sites regionally. Table 6.4 (adapted from Jeffries et al. 1996:24:Table 9) displays nut shell:maize ratios of the Townsend site as compared to four sites in East Tennessee, the

Croley-Evans site, and a collection of other sites in western Kentucky.

Even with the large amounts of nutshell that were recovered at the Townsend sites, the ratio of nutshell to corn is well within the range of nearby Mississippian sites in

East Tennessee and does not indicate any indication that inhabitants were less invested in corn agriculture than their neighbors. At the Croley-Evans site in Kentucky, Jefferies et al. (1996) found a nutshell:corn ratio of 28.1. They suggested that the site was located on the northern periphery of the “mainstream” Mississippian tradition. Based on the high amount of nut refuse compared to corn remains at this site, they argued that the lack of investment in corn agriculture was due to poor farmland, low populations, and/or a decreased need to invest in the social surplus demands experienced in core

Mississippian localities. The Mississippian occupants at Townsend, perhaps because of an ample supply of farmland, were undoubtedly fully invested in corn agriculture. The question becomes, whether they also invested in the social enterprise of corn agriculture to the same extent.

191

Table 6.4. Comparisons of Nutshell to Maize from Regional Mississippian Sites.

Nutshell/ Nutshell/ Site Maize (ct.) Maize (wt. g) Reference Townsend (40BT89, 40BT90, 40BT91) 2.98 2.09 this dissertation

Tennessee Mississippian Sites: Pittman-Alder (40M15) 0.25 Gremillion & Yarnell (1986) Martin Farm (40MR20) 4.83 Chapman & Shea (1981) Jones Ferry (40MR76) 3.29 Chapman & Shea (1981) Toqua, Dallas Phase (40MR6) 1.65 Chapman & Shea (1981)

Southeastern Kentucky: Croley-Evans (15KX24) 28.15 32.89 Jefferies et al. (1996)

Western Kentucky Mississippian Sites: Twin Mounds, all phases (15BA2) 0.03 Edging (1988) Burcham (15HI5) 39.42 Woodward (1988) Turk (15CE6) 1.81 4.23 Edging (1990) Marshall (15CE27) 0.17 1.25 Woodard (1987b) White (15FU14) 2.24 Woodard (1987a) Adams (15FU4) 1.35 Edging & Dunavan (1986) Wickliffe (15BA4) 0.95 Mackin (1986) Chambers (15ML109) 7.51 22.06 Rossen (1987) Gordon II (15LV151) 0.04 Wymer (1987) Andalex Village (15HK22) 1.05 Crites (1991)

192

Gourds were domesticated early, roughly between 4,500 and 3,500 years ago along with a suite of other indigenous plants (Smith 1995:193). Wild and domesticated

Cucurbita species were used for a variety of purposes, from containers and fishing technology to food. Current evidence indicates that bean was introduced sometime around 1200 CE into farming economies in the Southeast (Hart and Scarry 1999; Smith and Cowan 2003); however, the recovery of one fragment from designated Early

Mississippian (900–1100 CE) Feature 139 from 40BT91 warrants further investigation and could provide more conclusive evidence on the timing of the arrival of the domesticated bean in the region. A current attempt to obtain chronometric dates is in process (Timothy Baumann and Gary Crites personal communication 2017).

The low recovery of squash and bean is not necessarily a good indication of overall use or importance. Besides charred fragments of the cucurbit rind, very little of the plant generally survives as a large portion of the plant, including the rind, seeds, and flesh, is edible (King 1985). Even if it is introduced to fire, much of it does not survive carbonization. Similarly, bean is thoroughly consumed and is also very difficult to identify following carbonization. Experimental evidence suggests that when beans are introduced to fire they shrink by 9 percent, split, and often lose their seed coat and hilum (Schaefer 2017).

In addition to the edible crop species at Townsend, three tobacco seeds were identified. While few in number, their presence is significant and is one of the only direct indications of ritual activity we have at the sites. Their recovery from storage

193

features 13 and 690 located within the bastion structures of the palisade surrounding

Village 1 is compelling evidence that these features served a symbolic or ritualistic function for the community. Tobacco served an important ceremonial role across North

America in prehistory (Swanton 1946).

6.4.1.3 Fruits

The fruits category includes both summer- and fall-ripening fruits; occupants would have accessed a combination of fruits to supplement and enrich their diets in the spring and summer including blackberry/raspberry, blueberry, grape, maypop, plum/cherry, and strawberry. In the fall they would have transitioned to collecting pawpaw, persimmon, and sumac. Persimmon trees are not as common as many of the nut-producing trees in the Eastern Woodlands, but the many streams and river bottoms, along with fallowed fields that were present in close proximity to the Townsend sites, would have given occupants easy access to their preferred habitat (Radford et al.

1964:264; Taylor 1936:600).

Many of these fruits, for example maypop, were likely encouraged in garden plots (Gremillion 1989) or, like blackberry/raspberry, grape, and strawberry, would have done well along the edges of the disturbed habitats surrounding the sites such as abandoned fields, meadows, or forest edges (Lopinot 1982:765; Phillips 1979:25;

Steyermark 1963). The presence of these fruits at the sites, even in small numbers, is an

194

indication of their frequent exploitation and consumption due to their low preservation potential (Lopinot 1991:91).

6.4.1.4 Indigenous Seed Crops

The scarcity of indigenous seed crops from the Townsend sites is somewhat unexpected, especially since they are much more numerous in both the Woodland and

Cherokee occupations (Hollenbach and Yerka 2011:222). A suite of indigenous seed crops, both domesticated and cultivated, were utilized across the Eastern Woodlands beginning at approximately 3,500 years ago (Yarnell 1993). The edible seeds used prehistorically are a mixture of both grains and oily seeds that would have provided carbohydrates, proteins, and fats (Scarry 2003:69). Maygrass and little barley, which ripen in early spring, would have been an early harvest following winter, while others, especially chenopod, could have been collected later into the winter months (Scarry

2003:71).

Despite the benefits of these crops, the limited quantities and frequency of these seeds in features at Townsend suggest that they were of little importance to

Mississippian inhabitants. A focus on corn may have reduced the need for small seed crops (Scarry 2003:88). Corn agriculture combined with a wide range of easily accessible nuts, especially acorns, that could have easily provided adequate amounts of carbohydrates appears to have influenced farmers to mostly forgo growing small seed crops in many areas of the Interior Southeast (Scarry 2003:88).

195

6.4.1.5 Miscellaneous Taxa

A combination of weedy species, of which purslane, ragweed, spurge, and carpetweed are the most common, point to the disturbed nature of the landscape.

Several of the seed species, knotweed, ragweed, and smartweed, along with weedy grasses and legumes, may have also been food sources (Yarnell 1993). Some of the plants, namely sedges, spurges, bedstraw, morning-glory, and purslane, may have also been used for medicinal purposes (Gilmore 1932; Moerman 1998; Powers 1873–74).

Contributing to the potential ritualistic function of Feature 13, which is associated with the bastioned palisade of Village 1, morning-glory, several spurges, and purslane were recovered alongside the two tobacco seeds.

Plant material was frequently used prehistorically for a variety of utilitarian purposes. Moerman (1998:16–17) documented 442 plants used for their fibers, 217 used as dyes, and 1074 species that were used for other utilitarian purposes such as for hunting and fishing supplies, incense and fragrances, fuels, tools, and ceremonial use items. At the Townsend sites, this category is represented by plants including cane, pine cone, pitch, and bark from unidentified tree species. Pitch may represent tree resin, but it could also be the remains of other highly starchy and sugary items such as corn kernels.

196

6.5 Seasonality of the Sites Based on the Plant Remains

Several lines of evidence have been used by Koerner and Sullivan (2011) to suggest that the Townsend sites are representative of a community that was occupied seasonally, perhaps functioning as a short-term nut gathering and processing location.

First, they argued that a scarcity of decorative ornamentation along with a minimal number of forms within the ceramic technology is evidence of a special function of the

Townsend sites. Second, they interpreted the presence of structures and pits outside the palisade as evidence of processing related to the specialized nature of the site. Last, the limited number of hearths is argued as potential evidence of a seasonal occupation of the sites; if the site were abandoned in the cold winter months, the need for hearths for warmth would be negated.

While plant remains cannot stand alone as evidence of the seasonal occupation of a site, as fruits, nuts, and seeds are easily dried and stored for months at a time, the activities surrounding them, the gathering or harvesting and processing, generally take place within the season of ripenin. This is especially true with agricultural crops, such as corn, bean, and squash, which depend on timely planting, care, and harvesting. But this time frame also applies to the collection of many of the wild and encouraged plants that people gathered such as fruits, nuts, and edible weedy seed species that require scheduling of labor, timely collection to prevent infestations and to compete with other consumers, and can sometimes have demanding processing requirements.

197

The variety of plants recovered from across the Townsend sites follow various ripening and harvesting schedules. Strawberry was available in the spring; maypop, blackberry/raspberry, and blueberry were available in the summer; while persimmon, grape, and sumac ripened in the late summer/fall. Some edible seeds were available early in the spring, others were fall ripening, and some were still available into the early winter. Collection of nuts is restricted to the fall, but they can be stored in the shell as needed and the processing of them likely continued through winter. While it is possible that occupants of the Townsend sites brought a variety of resources with them for seasonal consumption, the plant remains in most areas of the sites strongly suggest that people were in the vicinity of the site, actively farming and gathering resources at least from spring through late fall.

6.6 Temporal Patterns of Plant Use

Even with the problematic mixed assignment of time periods for many of the features across the Townsend sites, the large sample size of the Mississippian features considered in this botanical analysis should accentuate any large-scale shift in plant use that may have occurred over time. However, based on the feature TCA designations

(Table 6.1) from 29 MSE (Early Mississippian), 41 MSM (Middle Mississippian), and 41

MSS (General Mississippian) flotation samples, I found a high degree of stability in the types and quantity of plants used by occupants of Tuckaleechee Cove. Based on the density of both wood and plants, it does appear that the sites were used more

198

intensively over time (Figure 6.1). There are no significant changes in the use of nuts, especially of acorn use, which remains constant; however, there is a slight decrease in the use of hickory and black walnut relative to other plants from the Early to Middle

Mississippian that may reflect the increasing importance of corn over time (Figure 6.2).

Figure 6.3 shows the growing influence of corn from the Early to Middle periods.

Along with the greater presence of corn overall in the Middle Mississippian, there is also evidence of increased corn processing activities as indicated by the corn kernel:corn cupule ratio (Figure 6.4). A higher corn kernel ratio indicates more activity associated with consumption, while a higher corn cupule ratio indicates more processing activity.

While there is no strong evidence to suggest the processing of corn for any form of tribute at the Townsend sites, the increased processing activity is an important change to note. This trend and its potential implications for social variation will be explored further in a comparison of the various contexts at the sites.

The minimal amounts of fruits and edible seeds overall preclude extensive quantitative comparisons. However, both categories do increase in abundance from the

Early to Middle periods (Figure 6.5), perhaps another indication of increased use of the sites over time.

6.7 A Comparison of the Various Contexts at the Sites

Facing similar environmental opportunities and challenges, over the course of roughly 400 years, Mississippian people in Tuckaleechee Cove decided to live in

199

Figure 6.1. Comparison of plant and wood density by temporal-cultural affiliation (the y-axes are scaled logarithmically).

200

Figure 6.2. Comparison of the relative densities of nuts (hickory, black walnut, and acorn) by temporal cultural affiliation (the y-axes are scaled logarithmically).

201

Figure 6.3. Comparisons of the relative density of corn kernels and corn cupules by temporal cultural affiliation (the y-axes are scaled logarithmically).

202

Figure 6.4. Comparison of kernel:cupule ratio by temporal cultural affiliation (the y-axes are scaled logarithmically).

203

Figure 6.5. Comparison of relative density of fruits and edible seeds by temporal cultural affiliation (the y-axes are scaled logarithmically).

204

different ways, in farmsteads or small villages, within a palisaded area and in the open.

I compare the plant remains across the multiple areas of the sites to try to understand how the choices people made in how to live affected or were affected by the plant remains. First, to parse out differences in the features themselves that may be influencing the distribution of plants, I compare the plants between the different types of features, cooking, storage, and hearths. Next, I point out unique features that may be skewing the dataset and that are evidence of specific activities at the site, economically and ritually related. I then turn to comparisons to highlight the different ways people used plants to express themselves across the site areas; for example, between farmsteads and villages, and between individual households and communally used areas.

6.7.1 Unique Features from across the Sites

Several of the contexts from across the Townsend sites are unique for the quantities and the combination of plants within them, both of which point to specific activities performed within them. Table A.6 lists all of the plant material that was recovered from each analyzed feature context. Feature 25, from 40BT89, contained a surprising amount and diversity of plant material represented by crops, including a large amount of processed corn, such as byproducts including corn cupules and glumes, squash, and a potential tobacco seed, as well as fruits including maypop and persimmon, many utilitarian and miscellaneous plant materials, and very small quantities of nuts (Table A.6). Using the large number of corn cupules and glumes,

205

Hollenbach (2011:224) interpreted this feature to be an example of a corn processing pit used presumably by women. The potential tobacco seed also points to the symbolism that may have been tied to this otherwise mundane activity.

Plant processing activity is also evident in several other features. Two flotation samples from Feature 697, a cooking feature from Farmstead 1, contained 57 percent of the total acorn from the sites, along with 1,065 fragments of hickory. This feature also contained an assortment of fruits, and a small amount of corn and indigenous seed taxa.

Feature 1795, a storage feature associated with a household at Farmstead 2, contained the largest amount of hickory from the contexts. It also had a relatively diverse assemblage besides nuts, including crops, indigenous seeds, and a generous assortment of miscellaneous weedy seed species.

The largest amount of corn was recovered from Feature 2120, a household storage feature at Farmstead 2. A total of 3,211 cupules, 122 cob fragments, and 41 kernels were identified. With a storage capacity of two liters, this feature is surprisingly small to have such a high density of remains and more likely functioned as a smudge pit. In addition to Feature 2120, processing activity is suggested by the high recovery of large numbers of corn cupules and glumes from Features 25 (previously discussed for nut processing), 582, and 632, all of which contain over 50 cupules. Interestingly,

Feature 582, which contains the third largest amount of processing material, appears to have been communally used and is associated with one of the two community buildings in Village 1.

206

The remains of consumption activity were present in much smaller numbers than were represented by processing activity. However, Features 13 and 1795

(discussed previously) have the largest number of corn kernels. The unique nature of

Feature 13 bears additional comment. Located within one of the bastions in the palisade of Village 1, the quantity and combination of artifacts from the feature indicate that it represents ritualistic activity of the community. Four flotation samples were examined from this context, which produced a remarkable diversity of taxa, some of which are rare from most other areas of the sites. This feature contained nuts, crops, fruits, edible seeds, and a variety of miscellaneous seeds and plants that may have had medicinal uses. Perhaps most compelling to suggest ritualistic activity is the combination of tobacco, morning-glory, and the variety of medicinal plants including purslane and spurges. Besides the plant material, the faunal remains tell a similar story and were interpreted to represent potential feasting due to the recovery of over 100 frog and toad remains (Hollenbach et al. 2011:264). In fact, the majority of the amphibians and reptiles recovered from the sites came from this feature. Both the food remains and the feature’s location within the bastion may hint at the symbolic or ritualistic nature of what is either a point of protection or entrance into the community. Of note, a large number of toads were found in shallow pit at the Warren Wilson site dating to the Pigah phase (Runquist

1979:322) and at the Coweeta Creek site during the Qualla phase (Ward and Davis

1999:171). In these cases, the use of toads has been linked to hallucinogenic or medicinal activity (Runquist 1979:285).

207

6.7.2 A Comparison of Plant remains by Feature Function

I compared plant remains across the various features based on their function to assess whether the plant material was a product of specific activities associated with certain contexts. Additionally, I wanted to assess whether any large-scale differences in food storage or cooking patterns could be identified from the differences between the features.

I found very few quantitative differences in plant remains between hearths, storage, and cooking features across the sites. Plant and wood density is very similar across the three types (Figure 6.6). The same is true for the relative densities of wood and nuts, which, shown in Figure 6.7, only demonstrate a slight increase of wood in cooking and hearth features that was likely a product of the increased fire activity associated with the features’ function. There were significantly more corn cupules in cooking features than in storage features (Figure 6.8), which can be attributed to the frequent use of corn refuse as a fuel source. The notches marking the 95th confidence interval for cooking does not overlap with storage features, but slightly overlaps with hearth features.

208

Figure 6.6. Comparison of plant and wood density by feature function (the y-axes are scaled logarithmically).

209

Figure 6.7. Comparison of the relative density of wood and nuts by feature function (the y-axes are scaled logarithmically).

210

Figure 6.8. Comparison of corn cupules and kernels by feature function (the y-axes are scaled logarithmically).

211

6.7.3 The Importance of Place: Patterns of Plant Consumption Based on Settlement Type

The many different types of settlements at Townsend, the farmsteads, villages, and isolated activities, present an opportunity to look for differences in daily practice that are expressed through foodways. I compare the way that plants are used in each of the settlements at the sites, farmsteads and villages. The results of this analysis are strongly suggestive of a great degree of similarity in the use of plants regardless of the type of settlement one choose to live in. However, the differential distribution of corn suggests that the daily activities surrounding plant use may have differed greatly depending on particular circumstances.

The first trend noted is that, similar to the large amount of variation in the placement and storage capacity of features between farmsteads across the sites, there is also a large amount of variation in the relative importance of different taxa. Although there are no statistically significant differences in the overall density of plants and wood between the farmsteads (Figure 6.9), the clear trend is that there is a lot of variation across farmsteads in quantities of the food remains, particularly acorns and black walnuts, corn, and fruits (Figure 6.10). Hickory does not display the same amount of variation (Figure 6.11), indicating that it was consistently an important resource despite potential different needs across the farmsteads. Interestingly, Farmsteads 1, 2, and 3 show evidence of significantly more corn processing activity, while Farmsteads 7 and 8 show higher rates of consumption (Figure 6.10 and 6.12). This pattern suggests that the women of the farmsteads located in the central portion of 40BT90 may have conducted 212

Figure 6.9. Comparison of plant and wood density by community type (the y-axes are scaled logarithmically).

213

Figure 6.10. Comparison of relative density of acorn, black walnut, corn cupules and kernels, and fruit by community (the y-axes are scaled logarithmically).

214

Figure 6.11. Comparison of the relative density of hickory by community (the y-axes are scaled logarithmically).

215

Figure 6.12. Comparison of corn kernel:cupule ratio by community (the y-axes are scaled logarithmically).

216

the majority of their processing activities near their domestic yard. This may point to the variation in practices at each individual farmstead, where preference or location of cornfields may have dictated activities.

Upon consideration of the differences between the plant remains of the farmsteads and village communities, I found it useful to merge the farmsteads into one category to highlight what I perceived to be fundamental differences that were expressed in the feature contexts (as described in Chapter 5). Although both village sites appear to be communally focused, there were also large-scale differences in the way that food appeared to have been stored, with residents in Village 2, which may have been occupied earlier in time, using subterranean storage pits in a communal fashion and residents in Village 1, which is associated with a later occupation, largely abandoning them for above-ground granaries. Because of the apparent temporal difference, it is useful to evaluate each of the village sites separately. Ephemeral or isolated contexts refer to features that are located in an area that cannot be considered associated with either a particular farmstead or village.

Once again, hickory is a very stable resource, regardless of whether occupants were living in a farmstead, village, or if it was an isolated location (Figure 6.13). There is no clear pattern within the other nut resources, other than that both villages contain higher densities of acorn. However, the differences are not statistically significant

(Figure 6.14). Fruit density was greater in Village 1 (Figure 6.15), although much of that

217

is owed to the large amount of fruits recovered from a single feature, Feature 711, that was located in Village 1.

A comparison of corn remains between the locations, especially between the two village locations, is suggestive of very different practices associated with processing activities. It appears that in every location corn was consumed in relatively similar amounts: there are no significant differences in the abundance of maize kernels (Figure

6.16). However, there are significantly different quantities of corn cupules, which are the byproducts of corn processing activity, between the different types of communities.

Most notable are the differences between Village 1 and Village 2 (Figure 6.16). It appears that the women of Village 1 were shelling significantly more maize than residents at Village 2, and more maize than residents in farmsteads or the ephemeral feature locations. One explanation for this may be a difference in where processing of corn was taking place between the different locations. Shifts in the ratio of corn refuse to corn kernels have been attributed to a move towards an intensification of infield cultivation versus outfield (VanDerwarker 2010:83). If fields are located at some distance from the settlement (outfield), then women might have processed corn near the fields to reduce the bulk of their load to carry back to the settlement. It is reasonable to suggest that corn will be processed more intensively a short distance from field locations. Specifically, it may be that the women of Village 1, tightly circumscribed in the palisaded village, were doing most of their processing inside the village grounds.

With four above-ground storage facilities close at hand, this may have been the most

218

Figure 6.13. Comparison of relative density of hickory by community type (the y-axes are scaled logarithmically).

219

Figure 6.14. Comparison of relative density of acorn and black walnut by community type (the y- axes are scaled logarithmically).

220

Figure 6.15. Comparison of the relative density of fruit by community type (the y-axes are scaled logarithmically).

221

Figure 6.16. Comparison of relative density of corn remains by community type (the y-axes are scaled logarithmically).

222

efficient place to complete the activity. Although there is little evidence to confirm a threat of violence or warfare in the region, if the bastions along the palisade are in fact related to protection, it may have also been safer to complete processing on-site.

A second, less likely explanation for women at Village 1 processing more corn than they were consuming is that some of the consumable material was being sent to another location, perhaps indicating a relationship of tribute with a nearby center such as Toqua or Citico. A good indication of status differences between communities is the amount of processing versus consumption activities occurring at a site (Welch and

Scarry 1995). Based on the premise that shelled kernels would be the preferred form of tribute, Welch and Scarry (1995) argued that despite similar consumption levels across communities, the abundance of processing material in the form of cupules at outlying farmsteads was an indication of these lower-status communities fulfilling tribute requirements, again to reduce the bulk of the load. While there is currently little evidence to confirm this occurrence in the Little Tennessee River Valley, it is not completely unreasonable to think that some of the outlying villages may have paid some form of tribute to the large, conglomerated Mississippian settlements to the southwest on the Little Tennessee River such as Toqua and Citico. Both sites are under 50 kilometers away, have Hiwassee Island occupations, are mound centers, and interestingly, Toqua also has similar defensive-like bastioned architectural elements along its palisade, but it does have a slightly later date than the Townsend sites. The residents of Village 1 were not hiding their surplus underground and instead their

223

above-ground storage could rather be viewed as ostentatious display. If they were paying tribute they were not trying to hide their surplus from more powerful settlements. The display of surplus goods may have been used by residents to reinforce community identity. Similar strategies of reinforcement have been investigated through foodways activities such as feasting (VanDerwarker et al. 2007).

Alternatively, residents at Village 2 have a significantly higher consumption ratio than all of the other areas, but a significantly lower amount of processing refuse (Figure

6.16). Village 2 is characterized by communally-shared subterranean storage. One above-ground storage facility is associated with the village; however, it is located at the east end of the habitation area. The scarcity of shelled corn refuse at Village 2, despite the greater amount of consumption suggests that women in this village chose to process corn off-site. Perhaps space was not an issue and it was both safe and convenient to shell corn elsewhere. Or perhaps they had several dispersed fields to lower the risks of crop failure , with these fields being located farther from the village (Scarry 1993).

6.8 Summary and Conclusions

The data presented in this chapter present a story of plant use at Tuckaleechee

Cove in the Mississippian period that is one of a relatively stable foodways tradition, one that centered around the farming of corn, but that incorporated the many wild foods that could be gathered nearby. The residents of the Townsend sites were highly invested in corn agriculture, and by this period corn likely surpassed nuts in

224

importance. Based on the scarcity of indigenous seed crops, it appears that inhabitants did not need to invest in these resources, and instead focused on higher production offered by the surrounding fertile cropland that was available to them. The presence of a wide diversity of taxa that ripened and were harvested at various times of the year indicates that people likely occupied the sites year-round. The one exception to this trend is the lack of spring-and summer-ripening fruits and seeds at the Village 2 occupation. Their absence could be argued to suggest that this location functioned as a nut-processing camp as suggested by Koerner and Sullivan (2011). However, the lack of increased nut-processing byproducts from this location, along with the obvious investments in the site itself, the various features, buildings, and palisade, would suggest otherwise.

This study finds that there were distinct differences in the activities surrounding the processing and consumption of corn that appear to be related to the characteristics of individual communities. People living within farmsteads likely made decisions based on their individual needs. As far as foodways are concerned, this meant that not only was there a lot of variation in the quantities of foods consumed, but also that processing activities could have occurred near the houselot or in a more distant location near the field. Conversely, the villages’ activities, especially those related to corn, are very patterned. The differences between each are conspicuous: women in Village 1 not only processed a lot of corn, but apparently processed much more than the community consumed. Alternatively, Village 2 consumed large amounts of corn, but there is

225

relatively little evidence of processing. It is apparent that each village was responding to a unique set of circumstances, despite a similar environment and generally similar set of foodways.

Village 1 is interpreted to be one of the later occupations of the sites as it has mostly Middle Mississippian (1100-1300 CE) features (Table 4.6). Village 2 has no clear occupational date, although it may have been occupied earlier than Village 1 based on more pits that date to the Early Mississippian (900–1100 CE) (Table 4.6). The differences in corn practices may reflect the changing social requirements and constraints that came along with the increasing populations and increasing social stratification developing in the nearby Little Tennessee River Valley over time. In the case of Village 1, women appear to have responded by either processing foods closer to home or sending consumables elsewhere, as well as by displaying food stores in above-ground cribs, whether for members or outsiders as a reaffirmation of community health, stability, or wealth. The assemblage recovered from the bastioned pits in Village 1 also point to the use of ritual to enforce community bonds.

Being Mississippian at the Townsend sites meant investing in a farming lifestyle, but the way those activities were carried out appear to be strongly dictated by different needs and perhaps different social forces. The variation in foodways activities seen at the Townsend sites may indicates that even if a community was not directly participating in the status driven social system that was common at core Mississippian centers, they likely felt the reverberations and shifted their daily practices in result.

226

CHAPTER 7

HOUSEHOLD ANALYSIS OF THE TOWNSEND SITES

227

7.1 Introduction

The plant remains at the Townsend sites indicate that there are differences in foodways practices within and across the discrete geographic areas within the sites.

These differences are especially apparent when considering the corn remains associated with the two village sites as demonstrated in Chapter 6. In this chapter, I use a correspondence analysis to assess whether the composition of foodways materials— plants and ceramics— are spatially distinct, indicating a difference in foodways traditions, and potentially suggesting social difference. I first use a correspondence analysis to independently evaluate if the combination of plants was different in nature between the different assemblages. Second, I consider whether the ceramic assemblages, characterized by tempering agents and exterior decoration, pattern according to household or groups of households. Third, I integrate the two datasets to assess if there are any covarying, spatially patterned differences in larger foodways practices.

I begin with a discussion of the advantage of using a correspondence analysis to tease out patterns in foodways, along with my use of the technique at the Townsend sites in the Mississippian period. I then review the spatial areas considered in this analysis, pointing to the overall location of each settlement area along with distinct households within each. I follow with a presentation of my results, along with interpretations of how the patterns in foodways may reflect potential differences in the choices that people made in response to different ecological and social conditions.

228

7.2 Correspondence Analysis: Use in Paleoethnobotanical Assemblages

Correspondence analyses are increasingly being applied to archaeological datasets, including paleoethnobotanical ones to interpret differences in the way a combination of taxa was used within and across sites (e.g. Hollenbach 2009; Hollenbach and Walker 2010; VanDerwarker et al. 2007; VanDerwarker and Peres 2010).

Independent data analyses, such as ubiquity, frequency, densities, and multiple ratios, are often a sufficient quantitative measure to identify general trends in plant data, for example increasing and decreasing use of a taxon. However, multivariate methods are advantageous because they can help to define patterns, as well as the potential factors that may be contributing to those patterns (Fritz 2005). Additionally, multivariate analyses can be used to consider multiple cases and variables, thus allowing a comparison of several datasets from different sites or multiple areas within a single site

(VanDerwarker 2010:75). While the interpretation of multiple datasets through independent data analyses can be problematic for several reasons, including preservation and taphonomic biases and differences in sampling strategies, correspondence analysis is more easily applied as it is a measure of variance. That variance is explained in a visual representation that results in a depiction of differences in both variables plotting away from the origin or expected value and spatial proximity indicating closer relationships between variables (Hollenbach and Walker 2010:235).

These multivariate analyses have successfully been applied to address multiple questions related to plant assemblages—to identify special function or unique 229

assemblages, to find patterns of differential use in multiple datasets, and to examine the spatial variability of a set of data. For instance, VanDerwarker et al. (2007) used principal components analysis to search for archaeological signatures of communal feasting behavior that were otherwise unapparent. Their use of multivariate statistics revealed that feasting menus are not just signaled by luxury or special food items, but instead by the consumption of a unique combination of familiar foods eaten in abundance. VanDerwarker (2010) also demonstrated the effectiveness of correspondence analysis to identify relationships between plant and animal assemblages. By integrating the two datasets, she found close associations of certain plant taxa and fauna types, which she then correlated to large-scale cultural and ecological factors. Fundamentally, multivariate analyses are aimed at understanding the differential use of food resources that can allow an archaeological interpretation of what separates foodways between people and groups across time and space; however, teasing apart those differences to assess whether they are related to social circumstance or if they are a product of environmental change is an interpretive task.

While a number of studies have approached foodways traditions using an independent analysis of plants or ceramic artifacts (e.g. Blitz 1993; Johannessen 1993a;

Marcoux 2008), few studies have combined the datasets to try to understand how each variable relates to one another. Both plant foodways and potting traditions are expressions of distinct choices that people, namely women, made that are a direct result of how they identify themselves—the way they learned to perform an activity, the

230

specific circumstances they experienced, and the manner in which they wanted to be viewed by others. I combine the two datasets at the Townsend sites to evaluate whether there is a pattern in the way that the households used the ceramic vessels to store, cook, and present plant remains. Although there are multiple examples of the effectiveness of integrating plant and faunal assemblages (e.g. Hollenbach and Walker 2010; Peres et al.

2010; VanDerwarker et al. 2010), due to the paucity of faunal remains as a result of poor preservation, animal assemblages were not integrated in this study.

7.3 Correspondence Analysis: The Townsend Sites

7.3.1 Methods

In this analysis, my cases were the contexts of households. At the Townsend sites of 40BT90 and 40BT91, I have suggested in Chapter 4 that there are three types of settlements: 1) ephemeral, 2) farmsteads, and 3) villages. Ephemeral occupations are composed of isolated features that could not clearly be associated with domestic architecture and are not considered further in this analysis. There are eight farmstead settlements at the Townsend sites that are represented by either one to two paired structures, occasional corncrib structures, and palisade segments. Villages are categorically different from the farmsteads due to a combination of factors such as a greater number and density of structures, multiple features, the distinct presence of community buildings at Village 1, and a tightly structured community layout within a

231

circumscribed palisaded compound. I designated two areas as village complexes, which included several households. Each of the farmstead and village areas and corresponding households are illustrated in Figures 4.20 and 4.21.

There are 20 distinct domestic units (A–ZZ) across the Townsend sites: seven at

40BT90 and 13 at 40BT91 (Figures 4.20 and 4.21). Two of these designations, M and N, refer to community buildings in Village 1. Each of the identified households, A–ZZ, are composed of either a single domestic structure or paired domestic structures, features, and in some cases a palisade wall (Figures A.1–A.14.). However, several of the households from the sites could not be associated with feature contexts, including

Households B, F, G, I, O, R, S, T, and were not considered as cases in the correspondence analysis. The designation of Z refers to communal features located at Village 2, and ZZ refers to communal features located at Village 1 of 40BT91.

All statistical correspondence analyses were run using the SYSTAT 13.1 statistical package. The variables used for the correspondence analysis include the following plant categories: acorn, black walnut, hickory, corn refuse (including cupules, glumes, and cob fragments), corn kernel, fruit, and cultigens. These plant variables were selected as they were the most abundant and ubiquitous; however, because of the limited number of individual taxa belonging to the fruits and cultigens categories, multiple taxa were pooled. Ceramic variables include pottery temper and exterior surface treatment, as recorded in the ARL database for each feature context (Driskell 2011). Koerner and

Sullivan (2011) have defined at least two distinct potting traditions during the

232

Mississippian occupation of Townsend, both the Hiwassee Island and Pisgah. They defined Hiwassee Island or Middle Mississippian ceramics based on a predominance of shell-tempering, whereas sand-tempered ceramics, often with curvilinear and rectilinear patterns, are primarily associated with Pisgah cultures. Tempering agents and exterior decorations are readily identified in the Mississippian Townsend contexts, and additionally, have been proven to be good indicators of distinct potting traditions, at least in the later Cherokee occupation of the site (Marcoux 2008).

7.3.2 Results

The correspondence analysis identified several interesting patterns in the household distribution of plant taxa (Figure 7.1). The first two variables explain the greatest amount of variation in the data (62 percent) along two axes (Tables 7.1 and 7.2).

The first visible pattern is the clustering of Households J and K, and Community

Building N, which suggests that these plant assemblages are similar. Community

Building M is loosely associated with this cluster. An examination of the variables contributing most to this trend, from the biplot visual representation (Figure 7.2) and a review of taxa, indicates that all of these households have predominantly large amounts of refuse, both of nuts and maize,

233

Figure 7.1. Biplot depicting the results of the correspondence analysis of plant remains across Mississippian households at the Townsend sites.

234

Figure 7.2. Biplot of plant taxa identified in Mississippian household features from the

Townsend sites.

235

Table 7.1. Diagnostic Statistics for Correspondence Analysis of Plant Taxa from Townsend Mississippian Households.

Factor Eigenvalue Percent Cumulative Percent 1 0.111 34.35 34.35 2 0.089 27.70 62.04 3 0.065 20.09 82.13 4 0.033 10.23 92.36 5 0.017 5.26 97.62 6 0.008 2.38 100.00

236

Table 7.2. Loading of Variables on Factors 1 and 2 of Plant Taxa from Mississippian Townsend Sites. to

Mass Inertia Quality Factor 1 Factor 2 Factor 1 Factor 2 Factor 1 Factor 2 Factor Variable Contribution to to Contribution Contribution Squared Correlation Correlation Squared Correlation Squared Taxon: Acorn 0.169 0.563 0.035 0.342 -0.002 0.178 0.000 0.563 0.000 Black Walnut 0.145 0.434 0.042 -0.294 -0.198 0.112 0.063 0.299 0.136 Corn Kernel 0.145 0.244 0.045 -0.093 -0.258 0.011 0.107 0.028 0.216 Corn Refuse 0.169 0.197 0.023 0.098 -0.132 0.015 0.033 0.070 0.127 Cultigens 0.072 0.848 0.086 -0.963 0.294 0.605 0.070 0.776 0.072 Fruit 0.108 0.886 0.075 0.164 0.766 0.026 0.712 0.039 0.847 Hickory 0.193 0.424 0.017 0.174 -0.082 0.052 0.014 0.346 0.077 Household: A 0.072 0.611 0.022 -0.240 0.36 0.037 0.105 0.188 0.423 C 0.084 0.939 0.008 -0.245 0.185 0.046 0.032 0.598 0.341 D 0.072 0.745 0.017 -0.368 -0.211 0.088 0.036 0.561 0.184 E 0.036 0.827 0.046 0.681 0.76 0.151 0.234 0.368 0.459 H 0.036 0.363 0.04 -0.213 -0.599 0.015 0.145 0.041 0.322 J 0.072 0.361 0.008 0.196 0.053 0.025 0.002 0.337 0.024 K 0.072 0.361 0.008 0.196 0.053 0.025 0.002 0.134 0.024 L 0.048 0.357 0.024 0.240 -0.346 0.025 0.065 0.116 0.241 M 0.06 0.634 0.02 0.411 0.196 0.092 0.026 0.517 0.117 N 0.072 0.361 0.008 0.196 0.053 0.025 0.002 0.337 0.024 P 0.072 0.77 0.024 -0.457 0.218 0.136 0.038 0.628 0.142 Q 0.048 0.623 0.024 0.391 -0.396 0.066 0.085 0.307 0.316 U 0.036 0.487 0.032 0.614 -0.241 0.123 0.024 0.422 0.065 V 0.06 0.938 0.014 0.137 -0.449 0.01 0.136 0.079 0.859 Z 0.072 0.745 0.017 -0.368 -0.211 0.088 0.036 0.561 0.184 ZZ 0.084 0.939 0.008 -0.245 0.185 0.046 0.032 0.598 0.341

237

including hickory, acorn, corn cupules and glumes, along with sporadic occurrences of fruits. The spatial distribution of this cluster is intriguing as Households J, K, and

Community Buildings N and M are located in Village 1; however, despite proximity inside this compound, Household J was originally given a possible Pisgah cultural affiliation based on the presence of a greater proportion of Pisgah-style ceramics than other household features at the sites and a later date of occupation than much of the other material in the compound. Household K and Buildings N and M were suggested to align closely to Hiwassee Island cultures to the west based on their pottery assemblages and architectural design (Koerner and Sullivan 2011). This analysis suggests that the occupations are contemporaneous.

A second cluster of households, V, L, and Q, also have significant quantities of refuse, nut and corn, but do not have the presence of fruits, an attribute that is likely due to differential preservation. Household C and the communal set of features from Village

1, designated ZZ, are differentiated from the other clusters by a large amount of refuse, along with more corn kernels and black walnut. Household A, P, and ZZ are all influenced strongly by the presence of cultigens.

Households E, U, and H have assemblages that are different from V, L, and Q, as suggested by their more distant location from the origin on the graph. However, the variation in both Households E and H are largely the result of their small assemblages with minimal quantities of taxa. Household U is relatively similar to Households V, L, and Q, except that it has a less diverse assemblage that is represented by fewer taxa.

238

The results of the correspondence analysis of ceramic temper and exterior surface treatment support several of the relationships suggested by the plant assemblages across households. The first two factors explain 42 percent of the data (Tables 7.3 and 7.4).

Primarily, the clustering of N and P, along with J, indicates significant household patterning with respect to exterior surface treatments (Figure 7.3). Notably, Household

J, located in Village 1, and P, located in Farmstead 8, were both households that were previously hypothesized to be isolated Pisgah settlements based on ceramic assemblages that included a high proportion of sand-tempered, decorated sherds (Koerner and

Sullivan 2011). What is particularly intriguing is that, as demonstrated in the plant biplot (Figure 7.1), Community Building N clusters with the Pisgah households. The households covary in the recovery of sand-tempered decorated and plain, along with shell-tempered complicated curvilinear ceramics (Figure 7.4).

While most of the households within the farmstead locations, along with those in

Village 2, appear to be represented by similar ceramic assemblages, namely a predominance of shell-tempered plain and eroded sherds, Households C, D, and Q have distinctive assemblages. Household C’s outlying position is apparently correlated to a predominance of shell-tempered plain and eroded sherds. However, Household C appears to be influenced heavily by a limited number of sherds, namely three sand- tempered check stamped and one shell-tempered smoothed over sherd, which are unique to this household context. Households D and Q are also outside the main cluster of households, and are influenced by the presence of limestone-tempered plain and

239

Figure 7.3. Biplot depicting the results of the correspondence analysis of ceramic temper and exterior treatment across Mississippian households at the Townsend sites.

240

Figure 7.4. Biplot of ceramic temper and exterior treatments identified in Mississippian household features from the Townsend sites.

241

Table 7.3. Diagnostic Statistics for Correspondence Analysis of Ceramic Temper and Exterior Treatment from Townsend Mississippian Households.

Factor Eigenvalue Percent Cumulative Percent 1 0.516 23.15 23.15 2 0.431 19.32 42.46 3 0.328 14.71 57.17 4 0.287 12.85 70.02 5 0.24 10.76 80.78 6 0.187 8.38 89.16 7 0.13 5.81 94.97 8 0.067 3.02 97.98 9 0.03 1.36 99.35 10 0.015 0.65 100.00

242

Table 7.4. Loading of Variables on Factors 1 and 2 of Ceramic Temper and Exterior Surface Treatment from Mississippian Townsend Households.

Mass Inertia Quality Factor 1 Factor 2 Factor 1 Factor 2 Factor 1 Factor 2 Factor Variable Contribution to to Contribution to Contribution Squared Correlation Correlation Squared Correlation Squared Temper and Decoration: limestone plain 0.069 0.469 0.121 0.077 -0.904 0.001 0.131 0.003 0.466 sand check stamped 0.017 0.564 0.149 0.080 -2.21 0 0.195 0.001 0.563 sand comp. 0.017 0.897 0.126 2.275 1.164 0.173 0.054 0.711 0.186 sand comp. curvilinear 0.017 0.168 0.126 -0.832 0.727 0.023 0.021 0.095 0.073 sand comp. rectilinear 0.069 0.64 0.129 -0.900 0.624 0.108 0.062 0.432 0.208 sand incised 0.017 0.361 0.183 -1.411 1.356 0.066 0.074 0.188 0.173 sand plain 0.034 0.59 0.137 -1.121 1.041 0.084 0.087 0.317 0.273 shell checked stamped 0.017 0.051 0.126 0.501 -0.343 0.008 0.005 0.034 0.016 shell comp. curvilinear 0.017 0.119 0.233 -1.042 0.724 0.036 0.021 0.080 0.039 shell cordmarked 0.138 0.242 0.076 0.221 -0.291 0.013 0.027 0.088 0.154 shell eroded 0.207 0.036 0.057 -0.091 -0.04 0.003 0.001 0.030 0.006 shell fabric impressed 0.017 0.897 0.126 2.275 1.164 0.173 0.054 0.711 0.186 shell incised 0.017 0.897 0.126 2.275 1.164 0.173 0.054 0.711 0.186 shell plain 0.259 0.03 0.152 -0.119 -0.06 0.007 0.002 0.024 0.006 shell polished 0.017 0.564 0.149 0.080 -2.21 0 0.195 0.001 0.563 shell simple stamped 0.034 0.667 0.108 1.388 0.41 0.129 0.013 0.613 0.054 shell smoothed over 0.034 0.02 0.108 -0.166 0.192 0.002 0.003 0.009 0.012 Household: A 0.072 0.611 0.022 -0.240 0.36 0.037 0.105 0.188 0.423 C 0.084 0.939 0.008 -0.245 0.185 0.046 0.032 0.598 0.341 D 0.072 0.745 0.017 -0.368 -0.211 0.088 0.036 0.561 0.184 E 0.036 0.827 0.046 0.681 0.76 0.151 0.234 0.368 0.459 H 0.036 0.363 0.04 -0.213 -0.599 0.015 0.145 0.041 0.322 J 0.072 0.361 0.008 0.196 0.053 0.025 0.002 0.337 0.024 K 0.072 0.361 0.008 0.196 0.053 0.025 0.002 0.134 0.024 L 0.048 0.357 0.024 0.240 -0.346 0.025 0.065 0.116 0.241 M 0.06 0.634 0.02 0.411 0.196 0.092 0.026 0.517 0.117 N 0.072 0.361 0.008 0.196 0.053 0.025 0.002 0.337 0.024 P 0.072 0.77 0.024 -0.457 0.218 0.136 0.038 0.628 0.142 Q 0.048 0.623 0.024 0.391 -0.396 0.066 0.085 0.307 0.316 U 0.036 0.487 0.032 0.614 -0.241 0.123 0.024 0.422 0.065 V 0.06 0.938 0.014 0.137 -0.449 0.01 0.136 0.079 0.859 Z 0.072 0.745 0.017 -0.368 -0.211 0.088 0.036 0.561 0.184 ZZ 0.084 0.939 0.008 -0.245 0.185 0.046 0.032 0.598 0.341 243

shell-tempered cordmarked and check stamped ceramics. Additionally, designation ZZ, which is a conglomeration of the communal features from Village 1, also has a richer assortment of ceramics than many of the other contexts, composed of predominantly shell-tempered plain and eroded, as well as complicated, cordmarked, fabric impressed, incised, and simple stamped sherds.

The correspondence analysis of the independent datasets, the plants and ceramics, provided relatively similar results suggesting clear associations or relationships between several households and distinctive assemblages in several other households. It is also worthwhile to include the two datasets in a single correspondence analysis to consider the potential differences in larger foodways traditions.

In the integrated dataset (Figure 7.5), once again, Households J, N, and P cluster, indicating a close relationship. The first two factors explain 37 percent of the data

(Tables 7.5 and 7.6). Surprisingly, the communal features from Village 1, designated Z, and the material from Community Building M are very similar in nature to most of the other farmstead and village households. Again, Households C, D, and Q, all of which are relatively isolated farmsteads, have distinctive assemblages. The differences in the ceramic assemblages overshadow the differences in plants and heavily influence the relationships of households, as is demonstrated in the grouping of all plant remains near the origin of the plot (Figure 7.6).

244

Figure 7.5. Combined biplot depicting the results of the correspondence analysis of plant taxa and ceramic temper and exterior treatment identified in Mississippian household features from the Townsend sites.

245

Figure 7.6. Combined biplot of plant taxa and ceramic temper and exterior treatments identified in Mississippian household features from the Townsend sites.

246

Table 7.5. Diagnostic Statistics for Correspondence Analysis of Combined Plant Taxa and Ceramic Temper and Exterior Surface Treatment from Townsend Mississippian Households.

Factor Eigenvalue Percent Cumulative Percent 1 0.263 20.45 20.45 2 0.209 16.26 36.71 3 0.201 15.62 52.32 4 0.149 11.55 63.88 5 0.109 8.50 72.38 6 0.095 7.38 79.76 7 0.079 6.16 85.92 8 0.064 5.01 90.93 9 0.033 2.59 93.51 10 0.032 2.47 95.98 11 0.021 1.63 97.61 12 0.016 1.25 98.85 13 0.011 0.88 99.73 14 0.003 0.27 100.00

247

Table 7.6. Loading of Variables on Factors 1 and 2 of Combined Plant Taxa and Ceramic Temper and Exterior Surface Treatment from Mississippian Townsend Households.

Mass Inertia Quality Factor 1 Factor 2 Factor 1 Factor 2 Factor 1 Factor 2 Factor Variable Contribution to to Contribution to Contribution Squared Correlation Correlation Squared Correlation Squared Taxa: Acorn 0.099 0.049 0.037 -0.082 -0.107 0.003 0.005 0.018 0.031 Black Walnut 0.085 0.018 0.028 -0.068 0.038 0.001 0.001 0.014 0.004 Corn Kernel 0.085 0.01 0.031 -0.060 0.005 0.001 0.000 0.010 0.000 Corn Refuse 0.099 0.058 0.019 -0.091 -0.052 0.003 0.001 0.044 0.014 Cultigens 0.043 0.255 0.041 0.465 -0.18 0.035 0.007 0.222 0.033 Fruit 0.064 0.123 0.062 -0.229 0.258 0.013 0.020 0.054 0.069 Hickory 0.113 0.087 0.026 -0.137 -0.037 0.008 0.001 0.081 0.006 Temper and Decoration: limestone plain 0.028 0.481 0.064 0.238 -1.015 0.006 0.139 0.025 0.456 sand check stamped 0.007 0.424 0.07 0.161 -2.036 0.001 0.140 0.003 0.421 sand comp. 0.007 0.901 0.064 2.466 1.447 0.164 0.071 0.670 0.231 sand comp. curvilinear 0.007 0.143 0.07 -0.691 0.962 0.013 0.031 0.049 0.094 sand comp. rectilinear 0.028 0.526 0.061 -0.823 0.678 0.073 0.062 0.313 0.213 sand incised 0.007 0.351 0.084 -1.416 1.462 0.054 0.072 0.170 0.181 sand plain 0.014 0.525 0.07 -1.053 1.212 0.06 0.100 0.226 0.299 shell checked stamped 0.007 0.144 0.07 0.945 -0.728 0.024 0.018 0.091 0.054 shell comp. curvilinear 0.007 0.107 0.093 -1.000 0.635 0.027 0.014 0.076 0.031 shell cordmarked 0.057 0.303 0.038 0.338 -0.294 0.025 0.023 0.172 0.130 shell eroded 0.085 0.012 0.018 0.036 -0.037 0 0.001 0.006 0.006 shell fabric impressed 0.007 0.901 0.064 2.466 1.447 0.164 0.071 0.670 0.231 shell incised 0.007 0.901 0.064 2.466 1.447 0.164 0.071 0.670 0.231 shell plain 0.106 0.075 0.021 -0.108 -0.052 0.005 0.001 0.061 0.014 shell polished 0.007 0.424 0.07 0.161 -2.036 0.001 0.140 0.003 0.421 shell simple stamped 0.014 0.718 0.06 1.705 0.359 0.157 0.009 0.687 0.031 shell smoothed over 0.014 0.007 0.063 0.127 0.117 0.001 0.001 0.004 0.003 Household: A 0.078 0.061 0.044 -0.095 -0.159 0.003 0.009 0.016 0.045 C 0.092 0.531 0.152 0.083 -0.932 0.002 0.382 0.004 0.527 D 0.092 0.222 0.143 0.485 -0.333 0.082 0.049 0.151 0.071 E 0.021 0.032 0.06 -0.291 0.083 0.007 0.001 0.030 0.002 H 0.028 0.021 0.046 -0.182 -0.025 0.004 0.000 0.021 0.000 J 0.078 0.549 0.139 -0.727 0.669 0.156 0.167 0.297 0.252 K 0.057 0.099 0.022 -0.198 0.005 0.008 0.000 0.099 0.000 L 0.043 0.051 0.031 -0.170 -0.09 0.005 0.002 0.040 0.011 248

Table 7.6 (cont.). Loading of Variables on Factors 1 and 2 of Combined Plant Taxa and Ceramic Temper and Exterior Surface Treatment from Mississippian Townsend Households.

249

7.4 Discussion

The patterns within the correspondence analysis of the plant taxa and exterior ceramic treatment from the Mississippian households across the Townsend sites are revealing in several ways. The analyses indicate the presence of four general groups: 1)

Households P, J, and potentially K, along with Community Building N; 2) Households

A, E, H, L, U, V, as well as Community Building M and the communal features of

Village 2; 3) distinct assemblages from Households C, D, and Q; and 4) a unique assemblage from designation ZZ, the communal features associated with Village 1

(Figure 7.7).

First, the clear association between Households P, J, K, and Community Building

N are highly suggestive of their close affiliation with the Pisgah culture. While

Household P may in fact be representative of a later occupation of the Cove by people from the east, the locations of Households J and K and Community Building N inside palisaded Village 1 speak to the likelihood that they were neighbors to the occupants of

Households H, I, and L. Despite the later date obtained from Household J of 1332 CE, based on its spatial alignment alone, it seems counterintuitive to think that it was not a part of this community. The results of the correspondence analysis provide further evidence that multiple structures in Village 1 may in fact be more closely related to

Pisgah culture.

250

Figure 7.7. Biplot depicting the results of the correspondence analysis suggesting a clustering of Mississippian into four general groups.

251

Although the households at Village 1 possess material assemblages that express difference, this study suggests that the communal features associated with this village, features designated ZZ, are an assertion of similarity. They do not cluster closely to the

Hiwassee Island or the Pisgah household material assemblages, and instead contain unusual and distinct plant and ceramic remains. The location of these features, primarily within the bastioned palisade, further indicate that the bastions may have been ritual or symbolic locations that could have functioned as places used to reinforce a communal identity.

Of further interest is the close association of several of the Village 1 households including H and L, along with Community Building M, to assemblages of Households U and V in the Village 2 complex, as well as several of the farmstead households, namely

A and E. Based on these associations, I argue that this cluster of households is emblematic of a Hiwassee Island tradition. If this is the case, then the dual occupation of

Village 1 by Hiwassee Island and Pisgah affiliated groups may be a product of the particular geographic location of the Townsend sites, on the boundary of two distinct cultural traditions. If Community Building M is more closely related to the Hiwassee

Island residents and Community Building N is more comparable to Pisgah residents, their abutting positions within the village allude to a potential dual kinship structure.

Marcoux (2008) argued, based on distinct potting traditions in the Cherokee period, that the Townsend sites represented a coalescent community; perhaps this is a tradition in

252

the Cove that extends back into the Mississippian period. If so, it is a remarkable occurrence of multiple traditions existing in some form of solidarity.

The distinct assemblages from Households C, D, and Q, all representative of farmstead settlements, are suggestive of either markedly different needs or social circumstances. A high degree of variation between multiple farmstead locations at the

Townsend sites was reflected in the feature locations as demonstrated in Chapter 5 with the variable distance of features from households and the wide range of storage sizes, as well as the plant remains (Chapter 6). It is perhaps not surprising, then, to see residents of farmsteads freely expressing more choice in daily practices and this may have influenced their decision to live in a farmstead instead of in a village.

7.5 Conclusions

Without mortuary data and biometrics from skeletal remains, it is nearly impossible to definitively separate individuals into ethnic categories. However, archaeological materials are correlates of identity and can signal cultural and group affiliations. While the lack of burial goods at the Townsend sites creates difficulty in assessing status differences, the relationships suggested from the correspondence analysis appear to be unrelated to status differences. Instead, the results of this analysis demonstrate that although status is often an initial explanation for differences in material assemblages during the Mississippian period, in the case of frontier locations, ethnic identity and cultural affiliations may need to be examined more closely.

253

CHAPTER 8

SUMMARY AND CONCLUSIONS

254

This dissertation is an exploration of daily practice in a boundary region during the Mississippian period. I use foodways as a lens to understand how the actions of everyday life, the performance of daily tasks, reflect individual and group decisions made in response to specific social and/or environmental circumstances at Tuckaleechee

Cove. While “traditional” definitions of Mississippian culture are becoming increasingly flexible to allow for a growing recognition of variation across Mississippian period sites, peripheral geographic areas and smaller sites are still viewed as subsidiary to larger centers. Instead, I chose to use an agent-oriented approach at the scale of the household to suggest that at the Townsend sites, the daily practice of residents reflects the distinct choices they made, that may have been related to, but were not predetermined by, larger political centers. Placing common people at the center of household and community decisions provides a narrative that is more emblematic of the everyday Mississippian experience.

I considered the variation in settlement practices during the Mississippian period using the multiple settlements, farmsteads and villages, as well as households across the

Townsend sites. In place of luxury items and burial goods, which remain unknown for the Townsend sites, I reconstructed the decisions of daily life across these locations using the form and placement of features, as well as the distribution and spatial patterning of plant taxa and ceramic attributes. I employed qualitative and multiple quantitative techniques including exploratory data analysis to compare the settlements

255

and households across the sites. I also used a multivariate statistical technique to expose patterns in the plants and ceramics datasets that are otherwise difficult to uncover.

The results of the dissertation point to a great deal of stability in larger foodways traditions by multiple groups at the Townsend sites that is indicated by the use of a similar set of foods. In general, the plant remains from the Townsend sites in the

Mississippian period indicate that people in Tuckaleechee Cove were fully invested in corn agriculture. Despite a location outside of the major river valleys, which has been correlated at other sites to a decreased emphasis on farming (Jefferies et al. 1996; Meyers

2002), both farmstead and village residents at this site complex grew, processed, consumed, and stored corn in amounts comparable to nearby larger and less isolated

Mississippian centers (Table 6.4). The scarcity of indigenous cultigens across the sites is perhaps a further testament to the productivity of corn agriculture in this fertile cove.

And, while corn was at the center of their food production strategies, they continued to depend on the wild resources that were readily available nearby, including a variety of nuts and fruits.

This study also addressed site use, whether the sites were occupied year-round or seasonally, which was a possibility that was raised by Koerner and Sullivan (2011).

The diverse range of plant foods identified from across the sites is a good indication of a permanent investment in each of the respective farmsteads or villages. Gathering, in some cases tending, and collecting and harvesting the various plant foods recovered would have in the least required people to be in the cove from spring to late fall.

256

Although limited in quantity, the presence of fruits in most areas of the site complex argues against a limited functional purpose for the settlements at Townsend, such as a nut-processing camp. A cursory glance at the scarcity of fruits from Village 2 may point to a fall occupation; however, if the investment into features, structures, and the palisade are taken in concert with the plant remains, it seems unlikely that residents left the settlement unoccupied for a portion of the year. Therefore, the absence of fruits is likely more related to differential preservation.

The results also point to distinct differences in how residents chose to perform their daily subsistence activities, patterns that I argue are tied to their greater social arena; their choice of settlement type, their cultural affiliation, and the degree of external pressure they were experiencing in their daily life.

There is a high degree of variation in the use of space, in storage practices, and in activities surrounding plant remains that I identified between the different site areas in my analysis of features (Chapter 5) and plant and ceramic remains (Chapters 6 and 7).

First, there are pronounced differences between residents in farmsteads and villages in the choices people made related to how space was used. The activities performed by farmstead residents display a high degree of variation that could be related to environmental, social, and/or political differences. The occupation dates for the farmsteads are difficult to define because of the palimpsest nature of the site complex; however, the lack of patterning is a sign of a high degree of flexibility in the farmstead lifestyle that is lacking in village settlements. There is no apparent spatial patterning of

257

farmstead household features, where features were located close and far from the domestic structure. This contrasts to the villages at the Townsend sites, where most features were located less than 10 m from a domestic structure. A similar spatial distribution to the villages was identified in the later Cherokee occupation of the Cove, where Marcoux (2008) determined that most features were located either inside or within 10 m of a domestic structure. Likewise, the storage capacity between farmstead households varied considerably, while in villages household storage capacity displayed much less variation. There are also no clear patterns in the distribution of foods between individual farmsteads (Chapter 6).

I suggest that these larger differences between the farmsteads and villages are tied to both the more isolated social and spatial locations of the farmstead settlements.

Without a community to contribute to or respond to, it is likely that small groups of one to two families made decisions around foodways activities based more heavily on their respective family’s needs. Thus, the relative densities of plants between farmsteads are highly variable, as are the location and number and capacity of storage features.

Additionally, space was likely not as important of a consideration—people were neither circumscribed by neighbors nor a palisade wall.

What is most striking about the two village sites at Townsend is that while both display what I consider to be a distinct communal identity, such that they are contributing to and using communal storage, they are storing food and performing foodways activities in very different ways that would suggest different social practices

258

and ideologies. For instance, despite Village 2’s similar emphasis on corn agriculture to

Village 1, they, in large part, continued a practice of storing food inconspicuously underground, while people at Village 1 chose to store food in above-ground granaries.

The presence of one large circular structure in Village 2, which could be an unusual form of a corncrib, could be an indication of greater status if associated directly with nearby

Household V. However, since the domestic structure is similar in size to other households in the compound, the ceramic and plant assemblages do not appear significantly different than the other households (Chapter 7), and there are so many shared features in the Village, this storage facility may be have been communally owned and therefore, a singular outward display of community solidarity.

The shift to above-ground storage in Village 1, although potentially a practical choice, is also an ostentatious signal of community identity and status. The multiple corncrib locations adjacent to community buildings and the palisade’s bastioned entrance speak to their “other than strictly utilitarian” purpose. Additionally, other lines of evidence, including the numerous community pits positioned along the palisade, and especially the unusual plant, animal, and ceramic assemblages from

Feature 13, indicate that the occupants of Village 1 were engaged in performative ritual.

Whether this was feasting, ritual cleansing, or some activity related to the either protective or welcoming nature of the bastioned areas is unclear.

The differing ratios of corn kernels and corn refuse between the two village sites

(Chapter 6) also point to the distinct strategies that people employed for corn

259

agriculture, which were likely a response to different sets of social circumstances. The different farming strategies are significant because this plant was at the center of their daily lives; growing, processing, cooking, consuming, and storing corn was an important economic and an important social investment. Although people at both villages consumed similar amounts of corn, there is a significantly lower ratio of corn refuse at Village 2 than at Village 1. I suggest that the differences between the two villages signify infield versus outfield corn cultivation and processing. The low incidences of corn shelling material, cupules and glumes at Village 2 suggests that women were growing and processing corn in an outfield, well away from their domestic living space. I argue that this practice is related to their social environment; they were not overly concerned with the presence of neighboring groups and perhaps placed greater value on diversifying the placement of their fields to lower risk of accidental loss. The presence of palisaded walls does indicate a separation of space, but the lack of bastions suggests that the palisades were not heightened to the same level of defensiveness as the palisade wall at Village 1. In contrast, the increased kernel/cupule ratio at Village 1 may be evidence of infield cultivation, and the shelling and storage of maize by women near their houselots.

Infield cultivation as a farming strategy has been tied to an intensification of maize production, which requires a shorter fallow time, but a potentially greater labor investment to produce sufficient yields (Killion 1987, 1990; VanDerwarker 2010). One reason people may shift to infield cultivation is due to outside pressure. VanDerwarker

260

and Wilson (2016) suggested that people in the Central Illinois Valley moved their fields close to their community so that women working in the fields could be protected by archers stationed at the palisade. Processing would then have been performed within the safety of the palisade.

A further explanation of the variation at the Townsend sites is that people’s decisions reflected their cultural affiliations. In consideration of the patterns of plant taxa and ceramic treatment and surface decoration that I identified using correspondence analysis, I agree with Koerner and Sullivan (2011) that there was a presence of both Hiwassee Island and Pisgah-related peoples occupying Tuckaleechee

Cove during the Mississippian period. A large majority of the households across

40BT90 and 40BT91 have similar assemblages and form a tightly knit cluster (Chapter 7).

The households that Koerner and Sullivan (2011) identified as potentially Pisgah affiliated, J and P, also have a close relationship with regards to plant remains and ceramic style. However, I depart from the suggestion that the Pisgah settlement within

Village 1 is not contemporaneous with the remainder of the community. Based on the close association of the plant remains and ceramic treatments of Households J and K, as well as Community Building N, and to the other isolated Pisgah Household P (Chapter

7), I argue that there were two distinct cultural traditions present in this palisaded complex—Hiwassee Island and Pisgah. I also raise the possibility that there were two community buildings that may have reflected each of these respective groups at the

261

village. If this were the case, then the performance of ritual may have been one way people with seemingly different cultural traditions reinforced community ties.

The question becomes, why did people with potentially different cultural affiliations choose to live in a tightly circumscribed community? One explanation is that the Townsend sites are representative of what researchers in the Contact period suggest are coalescent communities, places of refuge for communities under pressure (Hudson

2002; Kowalewski; Marcoux 2008). Several lines of evidence support a similar strategy at Village 1 in the Mississippian period: 1) the presence of palisades with bastions possibly suggestive of regional violence; 2) the multiple plant and ceramics traditions as identified in Chapter 7 indicative of a heterogeneous community; and 3) the greater amount of corn refuse material likely indicating tending of infields and processing corn within the safety of the palisade. However, I approach this suggestion with caution as there is no direct evidence in the region of frequent violence or warfare and few pronounced vertical status differences at Townsend or nearby sites in the Southern

Appalachians (Boudreaux 2007, 2013; Sullivan 2001; Sullivan and Rodning 2011). The palisades were certainly a marker of space, a distinction of “us versus them”, but this is not a direct indication of endemic physical violence. Nevertheless, outside pressure can be experienced in ways other than direct confrontation, and the residents of Village 1 may have felt the strain of the growing presence of nearby neighbors. Future excavations of similar farmsteads and villages nearby may provide more answers to these questions. Considering the high level of violence in other areas of the Southeast,

262

the interrelations between communities in East Tennessee need to be evaluated in novel ways. I agree with VanDerwarker and Wilson (2016) that approaches which evaluate the effect of violence and warfare on daily practices, such as those surrounding foodways, have much to contribute to understanding prehistoric lifeways in the region.

The main goal of this dissertation was to try to understand what influenced the daily practices associated with foodways during the Mississippian period occupation of the Townsend sites. I suggest that the differences in foodways were a product of distinct historical circumstances that appear to be related to an increasing awareness or threat from neighboring groups and the decisions that residents made concerning how to address those circumstances in daily life. People in Tuckaleechee Cove were undoubtedly aware of and may have been impacted by the activities of their Pisgah neighbors to the east and their Hiwassee Island neighbors to the west, the construction of mounds, participation in trade networks, and ceremonial enterprises, but nonetheless, they were not passive receivers of culture. The diverse ways that Townsend residents chose to operate in the greater Mississippian world are a testament to their agency. In consideration of the Townsend sites, based on the variation that was present across the multiple settlement types at this single geographic location during the Mississippian period, I argue that it is very difficult to make broad generalizations about what it means to be Mississippian. Instead, each individual site is a marker of cultural production and transformation.

263

REFERENCES

264

Allison, Penelope M. 1999 Introduction. In The Archaeology of Household Activities, edited by P.M. Allison, pp. 1–18. Psychology Press, London.

Alt Susan 2001 Cahokian change and the Authority of Tradition. In The Archaeology of Traditions: Agency and History Before and After Columbus, pp. 141–156. University of Florida Press, Tallahassee.

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. 163–216. Research in Economic Anthropology, Supplement 5. JAI Press, Greenwich, Connecticut. 1994 The Savannah River Chiefdoms. University of Alabama Press, Tuscaloosa. 1995 Paleoindian Interaction Networks in the Eastern Woodlands. In Native American Interactions: Multiscalar Analyses and Interpretations in the Eastern Woodlands, edited by Michael S. Nassaney and Kenneth E. Sassaman, pp. 3–26. The University of Tennessee Press, Knoxville. 1996a Fluctuations between Simple and Complex Chiefdoms: Cycling in the Late Prehistoric Southeast. In Political Structure and Change in the Prehistoric Southeastern United States, edited by John F. Scarry, pp. 231–252. University Press of Florida, Gainesville. 1996b 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. 29–45. University of Alabama Press, Tuscaloosa. 1999 Examining Chiefdoms in the Southeast: An Application of Multiscalar Analysis. In Great Towns and Regional Polities in the Prehistoric American Southwest and Southeast, edited by J. E. Neitzel, pp. 215–241. University of New Mexico Press, Albuquerque.

Anderson, David G., and Kenneth E. Sassaman 1996 Modeling Paleoindian and Early Archaic Settlement in the Southeast: A Historical Perspective. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 29–57. The University of Alabama Press, Tuscaloosa and London. 2001 Climate and Culture Change in Prehistoric and Early Historic Eastern North America. Archaeology of Eastern North America 29: 143–186.

265

2004 Paleoindian Occupations in the Southeastern United States. In New Perspectives on the First Americans, edited by B. T. Lepper, and R. Bonnichsen, 119–128. College Station, TX: Center for the Study of the First Americans, Texas A&M University Press. 2005 Pleistocene Human Occupation of the Southeastern United States: Research Directions for the Early 21st Century. In Paleoamerican Origins: Beyond Clovis, edited by R. Bonnichsen, B. T. Lepper, D. Stanford, and M. R. Waters, pp. 29–42. College Station, TX: Center for the Study of the First Americans, Texas A&M University Press. 2012 Recent Developments in Southeastern Archaeology: From Colonization to Complexity. The SAA Press, Washington, D.C.

Anderson, David G., Maasch, Kirk A., Sandweiss, Daniel H., and Paul A. Mayewski 2007 Climate and Culture Change: Exploring Holocene Transitions. In Climate Change and Cultural Dynamics: A Global Perspective on Mid-Holocene Transitions, edited by David G. Anderson, Kirk A. Maasch, and Daniel H. Sandweiss, pp. 1–23. Academic Press, London.

Anderson, David G., O’Steen, Lisa D., and Kenneth E. Sassaman 1996 Environmental and Chronological Considerations. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 3– 15. The University of Alabama Press, Tuscaloosa and London.

Anderson, David G., Smallwood, Ashley M., and D. Shane Miller 2015 Pleistocene Human Settlement in the Southeastern United States: Current Evidence and Future Directions. PaleoAmerica 1(1): 7–51.

Anderson, David G., Stahle, David W., and Malcolm R. Cleaveland 1995 Paleoclimate and the Potential Food Reserves of Mississippian Societies: A Case Study from the Savannah River Valley. American Antiquity 60:258–286.

Angst, Michael G., Creswell, Bradley A., Guymon, Gail L, Howell, Cameron S, Kocis, James J., Marcel, Daniel L., Parmalee, Paul W., and Judith A. Sichler 2005 Phase II Archaeological Evaluation of Sites 40KN45 and 40KN113 for the Proposed Golf Course at the University of Tennessee, Knoxville Experiment Station, Knoxville, Tennessee. The Archaeological Research Laboratory, Department of Anthropology, University of Tennessee. Submitted to the Tennessee Valley Authority.

266

Ashmore, W., and R. Wilk 1988 Household and Community in the MesoAmerican Past. In Household and community in the MesoAmerican Past, edited by R. Wilk and W. Ashmore, pp. 1– 27. University of New Mexico Press, Albuquerque.

Atalay, Sonya, and Christine A. Hastorf 2006 Food, Meals, and Daily Activities: Food Habitus at Neolithic Çatalhöyük. American Antiquity 71(2):283–319.

Bardolph, Dana N. 2014 Evaluating Cahokian Contact and Mississippian Identity in the Late Prehistoric Central Illinois River Valley. American Antiquity 79(1):69–89.

Barfield, T. 1997 The Dictionary of Anthropology. Blackwell, Oxford.

Bass, Quentin R. 1977 Prehistoric Settlement and Subsistence Patterns in the Great Smoky Mountains. Master’s thesis, Department of Anthropology, University of Tennessee, Knoxville.

Beck, Robin A., and David G. Moore 2002 The Burke Phase: A Mississippian Frontier in the North Carolina Foothills. Southeastern Archaeology 21(2):192–205.

Bense, Judith A. 2009 Archaeology of the Southeastern United States: Paleoindian to World War I. Routledge Press, London & New York.

Blitz, John H. 1993 Big Pots for Big Shots: Feasting and Storage in a Mississippian Community. American Antiquity 58(1):80–96. 1999 Mississippian Chiefdoms and the Fission-Fusion Process. American Antiquity 64(4):577–592. 2010 New Perspectives in Mississippian Archaeology. Journal of Archaeological Research 18:1–39.

Boudreaux, Edmond A. III 2007 The Archaeology of Town Creek. University of Alabama Press, Tuscaloosa. 2013 Community and Ritual Within the Mississippian Center at Town Creek. American Antiquity 78(3):483–501.

267

Boyd, C. Clifford Jr., Schroedl, Gerald F., and R.P. Stephen Davis 1983 Early Mississippian Cultural Development at Martin Farm in the Little Tennessee River Valley. Paper Presented at the 40th Annual Meeting of the Southeastern Archaeological Conference, Columbia, South Carolina.

Braun, E.L. 1950 Deciduous Forests of Eastern North America. Hafner Press/Macmillian, New York. Breitburg, Emmanuel, John B. Broster, Arthur L. Reesman, and Richard G. Stearns 1996 The Coats-Hines Site: Tennessee’s First Paleoindian-Mastodon Association. Current Research in the Pleistocene 13:6–8.

Brewer, Andrea J. 1973 Analysis of Floral Remains from the Higgs Site (40LO45). In Excavation of the Higgs and Doughty Sites, I-75 Salvage Archaeology, edited by M.C.R. McCollough and Charles H. Faulkner, pp. 141–144. Miscellaneous Papers No. 12, Tennessee Archaeological Society, Knoxville.

Bridges, P. S. 1989 Changes in Activities with the Shift to Agriculture in the Southeastern United States. Current Anthropology 30:385–94.

Briggs, Rachel V. 2015 The Hominy Foodway of the Historic Native Eastern Woodlands. Native South 8:112–146. 2016 The Civil Cooking Pot: Hominy and the Mississippian Standard Jar in the Black Warrior Valley, Alabama. American Antiquity 81(2):316–332.

Broster, John B., and Mark R. Norton 1996 Recent Paleoindian Research in Tennessee. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 288–297. The University of Alabama Press, Tuscaloosa and London.

Brumfiel, Elizabeth M. 1992 Distinguished Lecture in Archeology: Breaking and Entering the Ecosystem- Gender, Class, and Faction Steal the Show. American Anthropologist, New Series, Vol. 94, No. 3, pp. 551–567. 1995 Heterarchy and the Analysis of Complex Societies. In Heterarchy and the Analysis of Complex Societies, edited by Robert Ehrenreich, Carol Crumley, and Janet Levy, pp. 125–131. Archaeological Papers of the American Anthropological Association no. 6. Washington.

268

Brush, Nigel, and Forest Smith 1994 The Martins Creek Mastodon: A Paleoindian Butchery Site in Holmes County, Ohio. Current Research in the Pleistocene 11:14–15.

Busing, Richard T., Stephens, Luther A., and Edward E.C. Clebsch 2005 Climate Data by Elevation in the Great Smoky Mountain: A Database and Graphical Displays for 1947–1950 with Comparison to Long-Term Data. Great Smoky Mountains Regional Collection, U.S. Department of the Interior, U.S. Geological Survey: Data Series Report DS 115. Reston, Virginia.

Cannon, Michael D., and David J. Meltzer 2003 Early Paleoindian Foraging: Examining the Faunal Evidence for Large Mammal Specialization and Regional Variability in Prey Choice. Quaternary Science Reviews 23:1955–1987.

Chapman, Jefferson 1975 The Rose Island Site and the Bifurcate Point Tradition. University of Tennessee Department of Anthropology Report of Investigations No.14. Tennessee Valley Authority Publications in Anthropology No. 8. 1977 Archaic Period Research in the Lower Tennessee River Valley –1975: Icehouse Bottom, Harrison Branch, Thirty Acre Island, Calloway Island. University of Tennessee, Department of Anthropology Report of Investigations No. 18. Tennessee Valley Authority Publications in Anthropology, No. 13. 1979 Archaeological and Geological Investigations in the Tellico Reservoir. Edited by Jefferson Chapman. The University of Tennessee Department of Anthropology Report of Investigations 29. Tennessee Valley Authority, Publications in Anthropology 24. 1985a Archaeology and the Archaic Period in the Southern Ridge-and-Valley Province. In Structure and Process in Southeastern Archaeology, edited by Roy S. Dickens, Jr. and H. Trawick Ward, pp. 137–153. The University of Alabama Press, Tuscaloosa. 1985b Tellico Archaeology: 12,000 Years of Native American History. Report of Investigations No. 43. Department of Anthropology, The University of Tennessee, Knoxville.

Chapman, Jefferson, and Gary D. Crites 1987 Evidence for Early Maize (Zea mays) from the Ice House Bottom Site, Tennessee. American Antiquity 52:352–354.

Chapman, Jefferson, and Andrea B. Shea 1981 The Archaeological Record: Early Archaic Period to Contact in the Lower Tennessee River Valley. Tennessee Anthropologist 6: 61–84. 269

Clark, Jeffrey 2001 Tracking Prehistoric Migrations: Pueblo Settlers among the Tonto Basin Hohokam. Anthropological Papers No. 65. University of Arizona Press, Tucson.

Clay, R. Berle 2006 Interpreting the Mississippian Hinterlands. Southeastern Archaeology 25(1):48–64.

Cobb, Charles R. 2003 Mississippian Chiefdoms: How Complex? Annual Review of Anthropology 32:63– 84.

Cobb, Charles R., and Brian M. Butler 2002 The Vacant Quarter Revisited: Late Mississippian Abandonment of the Lower Ohio Valley. American Antiquity 67(4):625–641.

Coe, Joffre L. 1964 The Formative Cultures of the Carolina Piedmont. Transactions of the American Philosophical Society 54(5). Philadelphia.

Creswell, Bradley A., Angst, Michael G, Guymon, Gail L., Hollenbach, Kandace D., McDonough, Jennifer, Sichler, Judith A., and Charles D. Susano III 2012 Archaeological investigations and Phase III Data Recovery at Sites 40BT63, 40BT162, 40BT244, 40BT245, and 40BT257 Located at The University of Tennessee, Institute of Agriculture East Tennessee Research and Education Center (ETREC) Little River Animal and Environmental Unit in Blount County, Tennessee. The Archaeological Research Laboratory, Department of Anthropology, University of Tennessee. Submitted to the Tennessee Division of Archaeology and the University of Tennessee, Division of Facilities and Planning.

Crites, Gary D. 1996 Botanical Analysis. In Archaeological Investigations at Site 40BT47: A Multicomponent Site in the Eastern Ridge and Valley of East Tennessee, edited by Charles Bentz, Jr., and Lance Greene, pp. 67–74. Transportation Center, The University of Tennessee, Knoxville. Draft manuscript submitted to Tennessee Department of Transportation, Project Number 25001-2511-96. Manuscript on file at Archaeological Research laboratory, The University of Tennessee, Knoxville.

270

Crumley, Carole 1995 Heterarchy and the Analysis of Complex Societies. In Heterarchy and the Analysis of Complex Societies, edited by Robert Ehrenreich, Carol Crumley, and Janet Levy, pp. 1–6. Archaeological Papers of the American Anthropological Association No. 6. Washington.

Davis, R. P. Stephen, Jr. 1990 Aboriginal Settlement Patterns in the Little Tennessee River Valley. Report of Investigations No. 50. Department of Anthropology, The University of Tennessee, Knoxville.

Delcourt, H.R., and P.A. Delcourt 1987 Quaternary Palynology and Vegetational History of the Southeastern United States. In Pollen records of late Quaternary North American sediments, edited by V.M. Bryant, Jr. and R.G. Holloway, pp. 1–37. American Association of Stratigraphic Palynologists Foundation, Dallas, Texas, USA.

Dezendorf, Caroline 2013 The Effects of Food Processing on the Archaeological Visibility of Maize: An Experimental Study of Carbonization of Lime-treated Maize Kernels. Ethnobotany Letters 4:12–20.

Dickens, Roy S. 1970 The Pisgah Culture and Its Place in the Prehistory of the Southern Appalachians. Ph.D. Dissertation, Department of Anthropology, University of North Carolina, Chapel Hill. 1976 Cherokee Prehistory. University of Tennessee Press, Knoxville.

Dimbleby, G.W. 1967 The Preservation of Plant Material. In Plants and Archaeology, pp. 93–103. John Barker, London.

Dincauze, Dena F. 1993 Pioneering in the Pleistocene Large Paleoindian Sites in the Northwest. In Archaeology of Eastern North America Papers in Honor of Stephen Williams, edited by James B. Stoltman, pp. 43–60. Archaeological Report No. 25, Mississippi Department of Archives and History, Jackson.

Douglas, Mary 1972 Deciphering a Meal. Proceedings of the American Academy of Arts and Sciences 101(1):61–82.

271

Driskell, Boyce 2011 Introduction to the Archaeology of Tuckaleechee Cove, Blount County, Tennessee. Report submitted to the Federal Highway Administration and the Tennessee Department of Transportation Nashville, Tennessee. Archaeological Research Laboratory, University of Tennessee, Knoxville.

Dunbar, James S., and P.K. Vojnovski 2007 Early Floridians and Late Megamammals: Some Technological and Dietary Evidence from Four North Florida Paleoindian Sites. In Foragers of the Terminal Pleistocene, edited by Renee B. Walker and Boyce N. Driskell, pp. 167–202. University of Alabama Press, Tuscaloosa.

Elder, Joe A. 1959 Soil Survey, Blount County, Tennessee. U.S. Department of Agriculture, Soil Conservation Service. Soil Survey, ser. 1953, no. 7. Emerson T.E. 1997 Cahokia and the Archaeology of Power. University of Alabama Press, Tuscaloosa.

1989 Water, Serpents, and the Underworld: An Exploration into Cahokia Symbolism. In The Southeastern Ceremonial Complex: Artifacts and Analysis, edited by Patricia Galloway, pp. 45–92. University of Nebraska Press, Lincoln, Nebraska.

Ezzell, T., Ogle E., and R. Dickson 2005 Little River, Big Future Promoting Water Quality in the Little River Watershed Through Participatory Planning. Publication of the University of Tennessee Community Partnership Center. Available at: http://www.littleriverbigfuture.org/.

Faulkner, Charles H. 1972 The Mississippian-Woodland Transition in the Middle South. Proceedings: Southeastern Archaeological Conference, Bulletin 15:38–45. 1975 The Mississippian-Woodland Transition in the Eastern Tennessee Valley. Proceedings: Southeastern Archaeological Conference, Bulletin 18:19–30. 2002 A History of Tennessee Archaeology. In Histories of Southeastern Archaeology, edited by Tushingham, Shannon, Hill, Jane, and Charles H. McNutt, pp. 172–182. The University of Alabama Press, Tuscaloosa, Alabama.

Fenneman, N.M. 1938 Physiography of Eastern United States. McGraw-Hill Book Company, New York, New York. 1946 Physical Divisions of the United States. United States Geological Survey, Washington, D.C., map. 272

Ferring, R.C. 1992 Alluvial Pedology and Geoarchaeological Research. In Soils in Archaeology: Landscape Evolution and Human Occupation, edited by V.T. Holliday. Smithsonian Institution Press, Washington and London.

Finney, Fred A. 1985 The Carbon Dioxide Site. American Bottom Archaeology, FAI-270 Site Reports 11, Pt. 1. University of Illinois Press, Urbana.

Flatley, William T., Lafon, Charles W., Grissino-Meyer, Henri D., and Lisa D. Laforest 2013 Fire History, Related to Climate and Land Use in Three Southern Appalachian Landscapes in the Eastern United States. Ecological Applications 23(6):1250–1266.

2015 Changing Fire Regimes and Old-Growth Forest Succession along a Topographic Gradient in the Great Smoky Mountains. Forest Ecology and Management 350:96– 106.

Franklin, M. 2004 Foreword. In Household Chores and Household Choices: Theorizing the Domestic Sphere in Historical Archaeology, edited by K.S. Barile and J.C. Brandon, pp. xiii-xi. University of Alabama Press, Tuscaloosa.

Fridley, Jason D. 2009 Downscaling Climate Over Complex Terrain: High Finescale (<1000 m) Spatial Variation of Near–Ground Temperatures in a Montane Forested Landscape (Great Smoky Mountains). Journal of Applied Meteorology & Climatology 48(5):1033–1049.

Fritz, Gayle J. 1993 Early and Middle Woodland Paleoethnobotany. In Foraging and Farming in the Eastern Woodlands, edited by C. Margaret Scarry, pp. 39–56. University Press of Florida, Gainesville. 2005 Paleoethnobotanical Methods and Applications. In Handbook of Archaeological Methods, edited by Herbert D.G. Maschner and Christopher Chippindale, pp. 771-832. Altamira Press, Walnut Creek, CA.

Gilmore, Melvin R. 1932 Some Chippewa Uses of Plants. Papers of the Michigan Academy of Science, Arts, and Letters 17:119–143.

273

Graham, Russell W., Haynes, C. Vance, Johnson, D., and Marvin Kay 1981 Kimmswick: A Clovis-Mastodon Association in Eastern Missouri. Science 213:1115–1117.

Gavrilets, Sergey, Anderson, David G., and Peter Turchin 2010 Cycling in the Complexity of Early Societies. Cliodynamics 1(1):58–80.

Greene, Lance K. 1991 Archaeological Investigations at Site 40BT47: A Multicomponent Site in the Eastern Ridge and Valley of East Tennessee. Transportation Center, The University of Tennessee, Knoxville. Submitted to Tennessee Department of Transportation, Nashville.

Gremillion, K. J. 1989 The Development of a Mutualistic Relationship Between Humans and Maypops (Passiflora incarnata L.) in the Southeastern United States. Journal of Ethnobiology 9:135–155.

Griffith, G.E., Omernick, J.M., and S.H. Azevedo. 2017 Ecoregions of Tennessee. Corvalis, Oregon, US Environmental Protection Agency EPA/600R-97/022.

Hann, J. H. 1986 Translation of Alonso de Leturiondo's Memorial to the King of Spain. Florida Archaeology 2:165–225.

Hart, John P., and C. Margaret Scarry 1999 The Age of Common Beans (Phaseolus vulgaris) in the Northeastern United States. American Antiquity 64:653–658.

Hastorf, Christine A. 2012 Steamed or Boiled: Identity and Value in Food Preparation. Journal for Ancient Studies 2:213–242.

Helmkamp, Richard C. 1985 Biosocial Organization and Change in East Tennessee Late Woodland and Mississippian. Unpublished Ph.D. dissertation, Department of Anthropology, Purdue University, Lafayette Indiana.

Hendon, J. 1996 Archaeological Approaches to the Organization of Domestic Labor: Household Practice and Domestic Relations. Annual Review of Anthropology 25:45–61. 274

Hicks, D.J. 1979 Forest Microclimate in the Great Smoky Mountains. Journal of the Tennessee Academy of Sciences 54:84–88.

Hinkle, C. Ross, McComb, William, C., Safley, and John Marcus Safley Jr. 1993 Mixed Mesophytic Forests. In Biodiversity of the Southeastern United States: Upland. John Wiley and Sons, Inc.

Hirth, K. 1993a The Household as an Analytical Unit: Problems in Method and Theory. In Prehispanic Domestic Units in Western Mesoamerica: Studies of the Household, Compound, and Residence, edited by R. Santley and K. Hirth, pp. 21–36. CRC Press, Boca Raton, Florida. 1993b Identifying Rank and Socioeconomic Status in Domestic Contexts: An example from Central Mexico. In Prehispanic Domestic Units in Western Mesoamerica: Studies of the Household, Compound, and Residence, edited by R. Santley and K. Hirth, pp. 121-146. CRC Press, Boca Raton, Florida

Hollenbach, Kandace D. 2007 Gathering in the Late Paleoindian Period: Archaeological Remains from Dust Cave, Alabama. In Foragers of the Terminal Pleistocene, edited by Renee B. Walker and Boyce N. Driskell, pp. 132–147. University of Alabama Press, Tuscaloosa. 2009 Foraging in the Tennessee River Valley 12,500 to 8,000 Years Ago. University of Alabama Press, Tuscaloosa.

Hollenbach, Kandace D., and Gary D. Crites 2011 The Landscape of Tuckaleechee Cove. In Introduction by Boyce Driskell. Report submitted to the Federal Highway Administration and the Tennessee Department of Transportation Nashville, Tennessee. Archaeological Research Laboratory, University of Tennessee, Knoxville.

Hollenbach, Kandace D., and Renee B. Walker 2010 Documenting Subsistence Change during the Pleistocene/Holocene Transition: Investigations of Paleoethnobotanical and Zooarchaelogical Data from Dust Cave, Alabama. In Integrating Zooarchaeology and Paleoethnobotany: A Consideration of Issues, Methods, and Cases, edited by Amber M. VanDerwarker and Tanya M. Peres, pp. 227–244.

275

Hollenbach, Kandace D., Sichler, Judith A., and Jessica L. Vavrasek 2011 Mississippian Foodways in Tuckaleechee Cove. In The Mississippian Occupation of Tuckaleechee Cove, Blount County, Tennessee by Kandace D. Hollenbach and Stephen J. Yerka. Report submitted to the Federal Highway Administration and the Tennessee Department of Transportation Nashville, Tennessee. Archaeological Research Laboratory, University of Tennessee, Knoxville.

Homsey, Lara K. 2009 The Identification and Prehistoric Selection Criteria of Fire-Cracked Rock: An Example from Dust Cave, Alabama. Southeastern Archaeology 28(1):101–116.

Houk, R. 1993 Great Smoky Mountains National Park. Houghton Mifflin Company, Boston and New York.

Hudson, Charles. M. 1976 The Southeastern Indians. University of Tennessee Press, Knoxville. 2002 Introduction. In The Transformation of the Southeastern Indians, 1540-1760, edited by R. Ethridge and C. Hudson, pp. xi–xxxix. University Press of Mississippi, Jackson.

Jackson, Douglas K. 1980 Final Report on Archaeological Investigations at the Sandy Ridge Farm Site (11-S- 660). FAI-270 Archaeological Mitigation Project Report 20. Department of Anthropology, University of Illinois, Urbana-Champaign.

Jefferies, Richard W., Breitburg, Emanuel, Flood, Jennifer, and C. Margaret Scarry 1996 Mississippian Adaptation on the Northern Periphery: Settlement, Subsistence, and Interaction in the Cumberland Valley of Southeastern Kentucky. Southeastern Archaeology 15(1):1–28.

Johannessen, Sissel 1993a Food, Dishes, and Society in the Mississippi Valley. In Foraging and Farming in the Eastern Woodlands, edited by C. Margaret Scarry, pp. 182–205. University Press of Florida, Gainesville. 1993b Farmers of the Late Woodland. In Foraging and Farming in the Eastern Woodlands, edited by C. Margaret Scarry, pp. 57–77. University Press of Florida, Gainesville.

Johanson, Jessie L. 2010 Plant Remains, Investment Strategies, and Site Processes: Two Sites Along the Nolichucky River in Greene County, Tennessee. Master's Thesis, University of Tennessee, 2012 http://trace.tennessee.edu/utk_gradthes/1253 276

Kidder, Tristram R. 2006 Climate Change and the Archaic to Woodland Transition (3000–2500 cal B.P.) in the Mississippi River Basin. American Antiquity 71:195–231.

Killion, Thomas W. 1987 Agriculture and Residential Site Structure among Campesinos in Southern Veracruz, Mexico: Building a Foundation for Archaeological Inference. Ph.D. Dissertation. Department of Anthropology, University of New Mexico, Albuquerque, NM. 1990 Cultivation Intensity and Residential Structure: An Ethnoarchaeological Examination of Peasant Agriculture in the Sierra de los Tuxtlas, Veracruz, Mexico. Latin American Antiquity 1:191–215.

Kimball, Larry R. 1996 Early Archaic Settlement and Technology: Lessons from Tellico. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman., pp. 149–186. The University of Alabama Press, Tuscaloosa and London.

Kimball, Larry R. and William W. Baden 1985 Quantitative Model of Woodland and Mississippian Ceramic Assemblages for the Identification of Surface Collections. In The 1977 Archaeological Survey: An Overall Assessment of the Archaeological Resources of Tellico Reservoir, edited by Larry R. Kimball, pp. 121–274. Report of Investigations No. 40. Department of Anthropology, University of Tennessee, Knoxville.

King, Adam 2002 Etowah: The Political History of a Chiefdom Capitol. University of Alabama Press, Tuscaloosa.

King, Adam, and Maureen S. Meyers 2002 Exploring the Edges of the Mississippian World. Southeastern Archaeology 21:113– 116.

King, F. B. 1985 Early Cultivated Cucurbits in Eastern North America. In Prehistoric Food Production in North America, edited by R. I. Ford, pp. 73–98. Anthropological Papers No. 75. Museum of Anthropology, University of Michigan, Ann Arbor. 1987 Prehistoric Maize in Eastern North America: An Evolutionary Evaluation. Unpublished Doctoral Dissertation, Department of Agronomy, University of Illinois, Urbana.

277

King, P.B., and A. Stupka 1950 The Great Smoky Mountains, Their Geology and Natural History. Science Monthly 71:31–43.

Knapp, A. Bernard, and Wendy Ashmore 1999 Archaeologies Landscapes: Constructed, Conceptualized, and Ideational. In Archaeologies of Landscape: Contemporary Perspectives, edited by Wendy Ashmore and A. Bernard Knapp, pp. 1–32. Blackwell Publishers, Malden, Massachusetts.

Knight, Vernon J., Jr. 2010 Mound Excavations at Moundville: Architecture, Elites, and Social Order. University of Alabama Press, Tuscaloosa.

Knott, Laura L. 1981 The Botanical Remains from Phipps Bend. In The Phipps Bend Archaeological Project, by Robert H. Lafferty, III, pp. 397–407. Research Series No. 4, Office of Archaeological Research, The University of Alabama. TVA Publications in Anthropology No. 26.

Koerner, Shannon and Lynn P. Sullivan 2011 The Mississippian Period Occupation of Tuckaleechee Cove, Blount County, Tennessee. Report submitted to the Federal Highway Administration and the Tennessee Department of Transportation Nashville, Tennessee. Archaeological Research Laboratory, University of Tennessee, Knoxville.

Kowalewski, Stephen A. 2006 Coalescent Societies. In Light on the Path: The Anthropology and History of the Southeastern Indians, edited by T.J. Pluckhahn and R. Ethridge, pp. 94–122. University of Alabama Press, Tuscaloosa.

Kuttruff, Carl, and Richard Walling 2002 Excavations at 40LD52, an Early Mississippian Hamlet in the Tennessee River Valley, Loudoun County, Tennessee. Paper presented at the 14th Annual Meeting on Current Research in Tennessee Archaeology, Nashville.

Lafferty, Robert H. III 1981 The Phipps Bend Archaeological Project. Research Series 4, Office of Archaeological Research, University of Alabama, and TVA Publications in Anthropology 26, Knoxville, Tennessee.

278

Larson, Lewis H., Jr. 1972 Functional Considerations of Warfare in the Southeast during the Mississippi Period. American Antiquity: 37(3):383–392.

Lengyel, Stacey N., Jeffrey L. Eighmy, and Lynne P. Sullivan 1999 On the Potential of Archaeomagnetic Dating in the Midcontinent Region of North America: Toqua Site Results. Southeastern Archaeology 18:156-172.

Levy, Janet 1995 Heterarchy in Bronze Age Denmark: Settlement Pattern, Gender, and Ritual. In Heterarchy and the Analysis of Complex Societies, Societies, edited by Robert Ehrenreich, Carol Crumley, and Janet Levy. Archaeological Papers of the American Anthropological Association No. 6. Washington.

Lewis, Thomas M. N., and Madeline Kneberg Lewis 1946 Hiwassee Island. University of Tennessee Press, Knoxville. 1961 Eva: An Archaic Site. University of Tennessee Press, Knoxville.

Lewis, Thomas M. N., Madeline D. Kneberg-Lewis, and Lynne P. Sullivan 1995 The Prehistory of the Chickamauga Basin. The University of Tennessee Press, Knoxville.

Lopinot, N. H. 1982 Plant Macroremains and Paleoethnobotanical Implications. In The Carrier Mills Archaeological Project: Human Adaptation in the Saline Valley, Illinois, Vol. 2, edited by R. W. Jeffries and B. M. Butler, pp. 573–860. Research Paper No. 33. Center for Archaeological Investigations, Southern Illinois University at Carbondale. 1991 Archaeobotanical Remains. In The Archaeology of the Cahokia Mounds ICT0II: Biological Remains, pp. 1–253. Illinois Cultural Resources Study No. 13. Illinois Historic Agency, Springfield.

Lorenz, K.G. 1992 Big Men in the Big Black Valley: Small Scale Late Prehistoric Community Organization in Central Mississippi, Ph.D. dissertation, Department of Anthropology, University of Illinois, Urbana.

Lupton, Deborah 1996 Food, the Body, and the Self. Sage, Thousand Oaks, California.

279

Marcoux, Jon B. 2008 Cherokee Households and Communities in the English Contact Period, A.D. 1670–1740. Ph.D. dissertation, Department of Anthropology, University of North Carolina, Chapel Hill.

Martin, W.H. 1992 Characteristics of Old-Growth Mixed Mesophytic Forests. Natural Areas Journal 12:127–135.

Martinez-Bustos, F., Martinez-Flores, H.E., Sanmartin-Martinez, E., Sanchez-Sinencio, F., Chang, Y.K., Barrera-Arellano, D., and E. Rios 2001 Effect of the Components of Maize on the Quality of Masa and Tortillas During the Traditional Nixtamalisation Process. Journal of the Science of Food and Agriculture 81:1455–1462.

Mast, M.A., and J.T. Turk 1999 Environmental Characteristics and Water Quality of Hydrologic Benchmark Network Stations in the Eastern United States, 1963-95: U.S. Geological Survey Circular 1173-A, 158 p.

Maxham M. 2000 Rural communities in the Black Warrior Valley, Alabama: The Role of Commoners in the Creation of the Moundville I Landscape. American Antiquity 65:337–54.

McNab, W.H., Cleland, D.T., Freeouf, J.A., Keys, JE. Jr., Nowacki, G.J, and C.A. Carpenter 2005 Description of Ecological Subregions: Sections of the Coterminous United States. Washington, D.C.: U.S. Department of Agriculture, Forest Service. 80 p.

McCollough, Major C. R., and Charles H. Faulkner 1973 Excavations at the Higgs and Doughty Sites: I-75 Salvage Archaeology. Miscellaneous Paper No. 12. Tennessee Archaeological Society, The Department of Anthropology, The University of Tennessee, Knoxville.

Mehrer MW. 1984 Late Pleistocene Human Adaptations in Eastern North America. Ph.D. dissertation, Department of Anthropology, University of Washington, Seattle. 1988 Late Pleistocene Human Adaptations in Eastern North America. Journal of World Prehistory 2:1–53. 1995 Cahokia’ s Countryside: Household Archaeology, Settlement Patterns, and Social Power. Northern Illinois University Press, DeKalb. 280

Meltzer, David J. 1988 Late Pleistocene Human Adaptations in Eastern North America. Journal of World Prehistory 2:1–53.

Meyers, Maureen S. 2002a The Mississippian Frontier in Southwestern Virginia. Southeastern Archaeology 21(2):178–191. 2002b The Role of Southern Appalachian Mississippian Frontier in the Creation and Maintenance of Chiefly Power. In Archaeological Perspectives on the Southern Appalachians: A Multiscalar Approach, edited by Ramie A. Gougen and Maureen S. Meyers, pp. 219–244. 2008 Political Economy of Exotic Trade: On the Mississippian Frontier: A Case Study of a Fourteenth Century Chiefdom in Southwestern Virginia. University of Kentucky Doctoral Dissertations. Paper 126.

Miksicek, Charles H. 1987 Formation Processes of the Archaeobotanical Record. Advances in Archaeological Method and Theory 10:211-247.

Milner, George R. 1983 The Turner and Demange Sites. American Bottom Archaeology, FAI-270 Reports 4. University of Illinois Press, Urbana.

Moerman, Daniel E. 1998 Native American Ethnobotany. Timber Press, Inc. Portland, Oregon.

Moran, Emilio F. and Eduardo S. Brondizio 2013 Chapter 1: Introduction to Human-Environment Interactions Research. In Human-Environment Interactions: Current and Future Directions, edited by Eduardo S. Brondizio and Emilio F. Moran, pp. 1–24. Springer Netherlands.

Muller, Jon 1997 Mississippian Political Economy. Plenum Press, New York.

Munson, Patrick J. 1984 Comments on Some Additional Species, and Summary of Seasonality. In Experiments and Observations in Aboriginal Wild Food Procurement, Prehistoric Research Series, VI, No.2, pp. 459–473.

Nash, D. J. 2009 Household archaeology in the Andes. Journal of Archaeological Research 17:205– 261. 281

National Park Service 2017 https://www.nps.gov/grsm/learn/nature/index.htm

Pauketat, Timothy. R. 1998 The Archaeology of Downtown Cahokia: The Tract-15A and Dunham Tract Excavations. Studies in Archaeology No. 1. Illinois Transportation Archaeological Research Program, University of Illinois, Urbana-Champaign. 2003 Resettled Farmers and the Making of a Mississippian Polity. American Antiquity 68:39–66. 2004a Ancient Cahokia and the Mississippians. Cambridge University Press, Cambridge. 2004b The Economy of the Moment: Cultural Practices and Mississippian Chiefdoms. In Archaeological Perspectives on Political Economies, edited by G.M. Feinman and L.M. 2005 The Forgotten History of the Mississippians. In North American Archaeology, edited by Timothy R. Pauketat and Diana DiPaolo Loren, pp. 187–211. Blackwell, Oxford. 2007 Chiefdoms and Other Archaeological Delusions. AltaMira Press, Plymouth, United Kingdom. 2008 Founders’ Cults and the Archaeology of Wa-kan-da. In Memory Work: The Archaeologies of Material Practice, edited by B.J. Mills and W. Walker, pp. 61–68. School for Advanced Research Press, Santa Fe, New Mexico.

Pauketat, Timothy, Kelly, Lucretia S., Fritz, Gayle J., Lopinot, Neal H., Elias, Scott, and Eve Hargrave 2002 The Residues of Feasting and Public Ritual at Early Cahokia. American Antiquity 67:257–279.

Pearsall, D.M. 2000 Paleoethnobotany: A Handbook of Procedures. Academic Press, New York.

Perdue, T. 1995 Women, Men and American Indian Policy: the Cherokee Response to 'Civilization'. In Negotiators of Change: Historical Perspectives on Native American Women, edited by N. Shoemaker, pp. 90–114. Routledge, New York

Peregrine, Peter N., and Stephen H. Lekson 2012 The North American Oikoumene. In The Oxford Handbook of North American Archaeology, edited by Timothy R. Pauketat, pp. 64–72. Oxford University Press, Oxford, UK.

282

Peres, Tanya M., VanDerwarker, Amber M., and Christopher A. Poole 2010 The Farmed and the Hunted: Integrating Floral and Faunal Data from Tres Zapatos, Veracruz, edited by Amber M. VanDerwarker and Tanya M. Peres, pp. 287–308. Springer, New York.

Petruso, K. M., and J. M. Wickens 1984 The Acorn in Aboriginal Subsistence in Eastern North America: A Report on Miscellaneous Experiments. In Experiments and Observations on Aboriginal Wild Plant Food Utilization in Eastern North America, edited by P. J. Munson, pp. 360– 378. Prehistory Research Series, Vol. 6, No. 2. Indiana Historical Society, Indianapolis.

Phillips, J. 1979 Wild Edibles of Missouri. Edited by B. Prydon. Missouri Department of Conservation.

Pluckhahn, Thomas J. 2010 Household Archaeology in the Southeastern United States: History, Trends, and Challenges. Journal of Archaeological Research 18(4):331–385.

Polhemus, Richard R. 1985 Mississippian Architecture: Temporal, Technological, and Spatial Patterning of Structures at the Toqua Site (40MR6). Master’s Thesis, The University of Tennessee, Knoxville. 1987 The Toqua Site—40MR6: A Later Mississippian, Dallas Phase Town. Report of Investigations No. 41. Department of Anthropology, The University of Tennessee, Knoxville.

Polhemus, Richard R., and Jan F. Simek 1996 Community Organization at Carden Farm II, Tennessee. Paper presented at the Fifty-third annual meeting of the Southeastern Archaeological Conference, Birmingham, Alabama.

Popper, Virginia 1988 Selecting Quantitative Measures in Paleoethnobotany. In Current Paleoethnobotany: Current Analytical Methods and Cultural Interpretations of Archaeological Plant Remains, edited by C. A. Hastorf and V.S. Popper, pp. 53–71. Chicago University Press, Chicago.

Powers, Stephen 1873–4 Aboriginal Botany. Proceedings of the California Academy of Sciences 5:373–79.

283

Prentice, Guy 1986 An Analysis of the Symbolism Expressed by the Birger Figurine. American Antiquity 51:239–66.

Purrington, Burton L. 1983 Ancient Mountaineers: An Overview of the Prehistoric Archaeology of North Carolina’s Western Mountain Region. In The Prehistory of North Carolina: An Archaeological Symposium, edited by Mark Mathis and Jeffery J. Crow, pp. 83–160. North Carolina Division of Archives and History, Raleigh.

Radford, Albert E., Ahles, Harry F., and C. Ritchie Bell 1964 Guide to the Vascular Flora of the Carolinas. The University of North Carolina, Chapel Hill.

Reimer, P.J., Bard, E., Beck, J.W., Blackwell, P.G., Ramsey, C.B., Buck, E., Cheng, H., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Haflidason, H., Hajdas, I., Hatte, C., Heaton, T.J., Hoffman, D.L., Hoff, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., Manning, S.W., Niu, M., Reimer, R.W., Richards, D.A., Scott, E.M., Southon, J.R., Staff, R.A., Turney, C.S.M., and J. van der Plicht 2013 Intcal13 and Marine13 Radiocarbon Age Calibration Curves 0–50,000 Years cal BP. Radiocarbon 55:1869–1887.

Robin, C. 2003 New directions in Classic Maya Household Archaeology. Journal of Archaeological Research 11: 307-356.

Rodning, Christopher B. 2004 The Cherokee Town at Coweeta Creek. Unpublished Ph.D. Dissertation, Department of Anthropology, University of North Carolina, Chapel Hill.

Rodning, Christopher B, and Jane M. Eastman 2001 Introduction: Gender and the Archaeology of the Southeast, edited by Jane M. Eastman and Christopher B. Rodning, pp. 1–9. University Press of Florida, Gainesville.

Rogers , J. 1985 The Archaeological Analysis of Domestic Organization. In Mississippian Communities and Households, edited by J. Daniel Rogers and Bruce D. Smith, pp. 7–31. The University of Alabama Press, Tuscaloosa, Alabama.

284

Runquist, Jeannette 1979 Analysis of the Flora and Fauna Remains from Protohistoric North Carolina Cherokee Indian Sites. Ph.D. Dissertation, Department of Zoology, North Carolina State University, Raleigh.

Salo, Lawrence V. 1969 Archaeological Investigations in the Tellico Reservoir, Tennessee 1967-1968: An Interim Report. Report of Investigations No. 7. Department of Anthropology, The University of Tennessee, Knoxville.

Sassaman, Kenneth E. 2010 The Eastern Archaic, Historicized. AltaMira Press, Lanham, Maryland.

Saunders, Joe W. 2010 Late Archaic? What the Hell Happened to the Middle Archaic? In Trend, Tradition, andTurmoil: What Happened to the Southeastern Archaic? Proceedings of the Third Caldwell Conference, St. Catherines Island, Georgia, May 9–11, 2008, edited by David Hurst Thomas and Matthew C. Sanger, pp. 237–243. Anthropological Papers 93. American Museum of Natural History, New York.

Scarry, C. Margaret 1993 Variability in Mississippian Crop Production Strategies. In Foraging and Farming in the Eastern Woodlands, edited by C. Margaret Scarry, pp. 78–90. University Press of Florida, Gainesville. 2003 Patterns of Wild Plant Utilization in the Prehistoric Eastern Woodlands. In People and Plants in Ancient Eastern North America, edited by Paul E. Minnis, pp. 50–104. Smithsonian Institution Scholarly Press, Washington, D.C. 2008 Crop Husbandry Practices in North America’s Eastern Woodlands. In Case Studies in Environmental Archaeology, second edition, edited by E. Reitz, C. Scarry, and S. Scudder, pp. 391–404. Springer, New York.

Scarry Margaret C., and John F. Scarry 2005 Native American ‘Garden Agriculture’ in Southeastern North America. World Archaeology 37(2):259–274.

Scarry, C. Margaret and Vincas P. Steponaitis 1997 Between Farmstead and Center: The Natural and Social Landscape of Moundville. In People, Plants, and Landscapes: Studies in Paleoethnobotany, edited by K.J. Gremillion, pp. 107–122. University of Alabama Press, Tuscaloosa, AL.

285

Schaefer, Kimberly 2017 Phaseolus polystachios (L.) BSP. In Laboratory Guide To Archaeological Plant Remains From Eastern North America Advanced Paleoethnobotany Seminar (Anthropology 4214). https://pages.wustl.edu/fritz/phaseolus-polystachios-l.-bsp.

Schiffer, Michael 1987 Cultural Deposition: In Formation Processes of the Archaeological Record, Chapter 4, pp. 47–98. University of New Mexico Press, Albuquerque.

Schroedl, Gerald F. 1975 Archaeological Investigations at the Harrison Branch and Bat Creek Sites in the Tellico Reservoir. University of Tennessee, Department of Anthropology, Report of Investigations No. 10. 1978a The Patrick Site (40MR40), Tellico Reservoir, Tennessee. Report of Investigations No. 25. Department of Anthropology, The University of Tennessee, Knoxville. 1978b Excavations of the Leuty and McDonald Site Mounds. Report of Investigations No. 22. Department of Anthropology, University of Tennessee, Knoxville. 1986 Overhill Cherokee Archaeology at Chota-Tanasee. University of Tennessee, Department of Anthropology Report of Investigations 38, Knoxville. 1998 Mississippian Towns in the Eastern Tennessee Valley. In Mississippian Towns and Sacred Places: Searching for an Architectural Grammar, edited by R. Barry Lewis and Charles Stout, pp. 64–92. The University of Alabama Press, Tuscaloosa.

Schroedl, Gerald F., and Clifford Boyd Jr. 1985 Mississippian Origins in East Tennessee. Paper presented at the 6th Mid-South Archaeological Conference, Starkville, Mississippi.

Schroedl, Gerald F., C. Clifford Boyd Jr, and RP Stephen Davis Jr. 1990 Explaining Mississippian Origins in East Tennessee. In The Mississippian Emergence, edited by Bruce D. Smith, pp. 175–196. Smithsonian Institution Press, Washington, DC.

Schroedl, Gerald F., R. P. Stephen Davis, Jr., and C. Clifford Boyd, Jr. 1985 Archaeological Contexts and Assemblages at Martin Farm. Report of Investigations No. 39, Department of Anthropology, The University of Tennessee. Tennessee Valley Authority Publications in Anthropology No. 37.

Schuldenrein, Joseph 1996 Holocene Landscape History of the Greater Southeast. In Archaeology of the Mid-Holocene Southeast. Edited by Kenneth E. Sassaman and David G. Anderson, pp. 3–27. University Press of Florida.

286

Scott, Elizabeth M. 2008 Who Ate What? Archaeological Food Remains and Cultural Diversity. In Case Studies in Environmental Archaeology, edited by Elizabeth Reitz, C. Margaret Scarry and Sylvia J. Scudder, pp. 357–374.

Shanks, Royal E. 1954 Climates of the Great Smoky Mountains. Ecology 35(3):354–361.

Sharkey, Michael J. 2001 The All Taxa Biological Inventory of the Great Smoky Mountains National Park. Florida Entomologist 84(4):556–564.

Sherwood, Sarah C., Boyce N. Driskell, Asa R. Randall, and Scott C. Meeks 2004 Chronology and Stratigraphy at Dust Cave, Alabama. American Antiquity 69(3):533-554. Smith, Bruce D. 1984 Mississippian Expansion: Tracing the Historical Development of an Explanatory Model. Southeastern Archaeology 3:13–32. 1990 The Mississippian Emergence. Smithsonian Institution Press, Washington, DC. 1995 Seed Plant Domestication in Eastern North America: In Last Hunters, First Farmers: New Perspectives on Prehistoric Transitions to Agriculture, edited by T. Douglas Price and Anne Birgitte Gebauer, pp. 193–214. School of American Research Press, Santa Fe. 2007 Preface to the new edition. In The Mississippian Emergence, edited by Bruce D. Smith, pp. xix–xxxi. Reprint, University of Alabama Press, Tuscaloosa.

Smith, Bruce D., and C. Wesley Cowan 2003 Domesticated Crop Plants and the Evolution of Food Production Economies in Eastern North America. In People and Plants in Ancient Eastern North America, edited by Paul E. Minnis, pp. 105–125. Smithsonian Books, Washington, D.C.

Southworth, Scott, Art Schultz, Danielle Denenny, and James Triplett 2003 Surficial Geologic Map of the Great Smoky Mountains National Park Region, Tennessee and North Carolina. U.S. Geological Survey Open-File Report 03-381 Version 1.0. 2004.

Southworth, Scott, Art Schultz, and Danielle Denenny 2005 Geologic Map of the Great Smoky Mountains National Park Region, Tennessee and North Carolina. U.S. Geologic Survey, Open File Report 2005–1225.

287

U.S. Department of Agriculture (USDA) 2011 United States Department of Agriculture, National Agricultural Statistics Service. Available at: http://nassgeodata.gmu.edu/CropScape/.

Stephens, Luther Allin, Jr. 1969 A Comparison of Climatic Elements at Four Elevations in the Great Smoky Mountains National Park. Master’s Thesis, University of Tennessee.

Steponaitis, Vincas P. 1978 Location Theory and Complex Chiefdoms: A Mississippian Example. In Mississippian Settlement Patterns: Studies in Archeology, edited by Bruce D. Smith, pp. 417–453. Academic Press, New York.

Steyermark, J. A. 1963 Flora of Missouri. Iowa State University Press: Ames, Iowa.

Stuiver, M., and P.J. Reimer 1993 Extended 14C Data Base and Revised Calib 3.0 14C Age Calibration Program. Radiocarbon 35: 215–230.

Sullivan, Lynn P. 1987 Mouse Creek Phase Household. Southeastern Archaeology 6:16–6229. 1989 Cultural Change and Continuity in the Late Woodland and Early Mississippian Occupations of the Mouse Creeks Site. Tennessee Anthropologist 14: 33-63. 1995 Mississippian Household and Community Organization in Eastern Tennessee. In Mississippian Communities and Households, edited by D. R. Rogers and B. D. Smith, pp. 99–123. University of Alabama Press, Tuscaloosa. 2001 Those Men in the Mounds: Gender, Politics, and Mortuary Practices in Late Prehistoric Eastern Tennessee. In Archaeological Studies of Gender in the Southeastern United States, edited by J. M. Eastman and C. B. Rodning, pp. 101- 126. University Press of Florida, Gainesville, FL. 2007 Dating the Southeastern Ceremonial Complex in Eastern Tennessee. In Chronology, Iconography, and Style: Current Perspectives on the Social and Temporal Contexts of the Southeastern Ceremonial Complex, edited by Adam King, pp. 88–106. University of Alabama Press, Tuscaloosa, AL. 2009 Archaeological Time Constructs and the Construction of the Hiwassee Island Mound. In 75 Years of TVA Archaeology, edited by Erin Pritchard. University of Tennessee Press, Knoxville.

288

Sullivan, Lynne P., and Dorothy Humpf 2000 Realm to Realm: A Comparison of Etowah and Mississippian Eastern Tennessee. Paper Presented at the (66th) Annual Meeting of the Society for American Archaeology, New Orleans, Louisiana.

Sullivan, Lynne P. and Christopher B. Rodning 2011 Residential Burial, Gender Roles, and Political Development in Late Prehistoric and Early Cherokee Cultures of the Southern Appalachians. Archaeological Papers of the American Anthropological Association. 20(1):79–97.

Swanton, John R. 1946 The Indians of the Southeastern United States. Bureau of American Ethnology, Bulletin 137. Smithsonian Institution, Washington, D.C.

Talalay, Laurie, Donald R. Keller, and Patrick J. Munson 1984 Hickory Nuts, Walnuts, Butternuts, and Hazelnuts: Observations and Experiments Relevant to Their Aboriginal Exploitation in Eastern North America. In Experiments and Observations on Aboriginal Wild Plant Food Utilization in Eastern North America, edited by Patrick J. Munson, pp. 338–359. Prehistory Research Series Vol. 6, No.2. Indianan Historical Society, Indianapolis.

Taylor, Norman 1936 The Garden Dictionary: An Encyclopedia of Practical Horticulture Garden Management and Landscape Design.

Thomas L. 2001 The Gender Division of Labor in Mississippian Households. In Archaeological Studies of Gender in the Southeastern United States, edited by JM Eastman and CB Rodning, pp. 29–56. University of Florida Press, Gainesville.

Turkon, Paula 2004 Food and Status in the Prehispanic Malpaso Valley, Azcatecas, Mexico. Journal of Anthropological Archaeology 23:225–251.

Twiss, Katheryn C. 2007 We Are What We Eat. In The Archaeology of Food and Identity, edited by Katheryn C. Twiss, pp. 1–15. Center for Archaeological Investigations, Southern Illinois University, Carbondale.

289

Underwood, Christopher A. 2013 Fire and Forest History from Soil Charcoal in Yellow Pine and Mixed Hardwood-Pine Forests in the Southern Appalachian Mountains, U.S.A. PhD diss., University of Tennessee.

VanDerwarker, Amber M. 1999 Feasting and Status at the Toqua Site. Southeastern Archaeology 18(1):24–34. 2010 Correspondence Analysis and Principal Components Analysis as Methods for Integrating Archaeological Plant and Animal Remains. In Integrating Zooarchaeology and Paleoethnobotany: A Consideration of Issues, Methods, and Cases, edited by Amber M. VanDerwarker and Tanya M. Peres. Springer, New York.

VanDerwarker, Amber M., and Greg D. Wilson 2016 War, Food, and Structural Violence in the Mississippian Central Illinois Valley. In The Archaeology of Food and Warfare, edited By A. M. VanDerwarker and G. D. Wilson, pp. 75–105. Springer International, New York.

VanDerwarker, Amber M., Scarry, C. Margaret, and Jane M. Eastman 2007 Menus for Families and Feasts: Household and Community Consumption of Plants at Upper Saratown, North Carolina. In The Archaeology of Food and Identity, edited by Katheryn C. Twiss, pp. 16–49. Center for Archaeological Investigations, Occasional Paper No. 34. Southern Illinois University.

VanDerwarker, Amber M., Wilson, Gregory D., Hoppa, Kristin, and Amy Gusick 2012 Culture Contact, Earth Ovens, and Persistent Foodways: Archaeobotanical Analysis of a Failed Corn Roast from the C.W. Cooper site in the Central Illinois River Valley. Paper presented at the 77th Annual Meeting of the Society for American Archaeology, Memphis, Tennesseee.

Voss, Barbara L. 2005 From Casta to Californio: Social Identity and the Archaeology of Culture Contact. American Anthropologist 107(3):461–474.

Wagner, Gayle 1988 Comparability among Recovery Techniques. In Current Paleoethnobotnay: A Handbook of Procedures, edited by C. Hastorf and V. Popper, pp. 17–35. University of Chicago Press.

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. 99–115. University of Nebraska Press, Lincoln. 290

2007 Foragers of the Terminal Pleistocene in North America. Edited by Walker, Renee B., Detwiler, Kandace R., Meeks, Scott C., and Boyce N. Driskell. University of Nebraska Press, Lincoln. 2001 Berries, Bones, and Blades: Reconstructing Late Paleoindian Subsistence Economy at Dust Cave, Alabama. Midcontinental Journal of Archaeology 26:169– 197.

Ward, H. Trawick, and R.P. Stephen Davis Jr. 1999 Time Before History: The Archaeology of North Carolina. The University of North Carolina, Chapel Hill.

Watson, Patty Jo 1976 In Pursuit of Prehistoric Subsistence: A Comparative Account of Some Contemporary Flotation Techniques. Midcontinental Journal of Archaeology 1(1):77–100.

Webb, S. David, Dunbar, James S., and Ben Waller 1984 A Bison antiquus Kill Site, Wacissa River, Jefferson County, Florida. American Antiquity 49:384–392.

Welch, Paul D. 1991 Moundville’s Economy. University of Alabama Press, Tuscaloosa.

Welch, Paul D., and C. Margaret Scarry 1995 Status-Related Variation in Foodways in the Moundville Chiefdom. American Antiquity 60(3)397–419.

Wells, Edward W., III 2006 Soapstone Vessel Chronology and Function in the Southern Appalachians of Eastern Tennessee. Unpublished Master’s thesis, University of Tennessee, Knoxville.

Wenhold, L. L. 1936 A 17th Century Letter of Gabriel Diaz Vara Calderon, Bishop of Cuba, Describing the Indians and Indian missions of Florida. Smithsonian Miscellaneous Collections, 95(16). Government Printing Office, Washington, DC.

Wesson, C.B. 2008 Households and Hegemony: Early Creek Prestige Goods, Symbolic Capitol, and Social Power. University of Nebraska Press, Lincoln.

291

Whittaker, R. H. 1956 Vegetation of the Great Smoky Mountains. Ecological Monograph 26:1-80.

Wilk, R. W. 1989 Decision Making and Resource Flows Within the Household: Beyond the Black Box. In The Household Economy: Reconsidering the Domestic Mode of Production, edited by R.W., pp. 23–52 Westview Press, Boulder.

Wilk, R., and R. Netting 1984 Households: Changing Forms and Functions. In Households: Comparative and Historical Studies of the Domestic Group, edited by R. Netting, R. Wilk, and E. Arnould, pp. 1–28. University of California Press, Berkeley. Willey, Gordon R. and Jeremy A. Sabloff 1993 History of American Archaeology. Thames and Hudson Ltd, London., Third Edition.

Wills, W.H. 2009 Cultural Identity and the Archaeological Construction of Historical Narratives: An Example from Chaco Canyon. Journal of Archaeological Method and Theory 16:283–319.

Wilson, Gregory D. 2008 The Archaeology of Everyday Life at Early Moundville. The University of Alabama Press, Tuscaloosa. 2010 Community, Identity, and Social Memory at Moundville. American Antiquity 75(1):3–18.

Wilson, Gregory D., Delaney, Colleen M., and Phillip G. Millhouse (In press) The Mississippianization of the Illinois and Apple River Valleys. In Mississippian Beginnings, edited by Gregory D. Wilson, pp. 97–129. University of Florida Press, Gainesville.

Windingstad, Jason D., Foss, John E., and Sarah C. Sherwood 2011 Chapter 5: Quaternary Soils, Geomorphology, and Geoarchaeology. In Introduction to the Archaeology of Tuckaleechee Cove, Blount County, Tennessee. Report submitted to the Federal Highway Administration and the Tennessee Department of Transportation Nashville, Tennessee. Archaeological Research Laboratory, University of Tennessee, Knoxville.

Woffard, B. Eugene, and Edward W. Chester 2002 Guide to the Trees, Shrubs, and Woody Vines of Tennessee. The University of Tennessee Press, Knoxville. 292

Wood, M. Jared 2014 Closely-Spaced Administrative Centers and the Organization of Mississippian Chiefdoms. In Multiscalar Archaeological Perspectives of the Southern Appalachians, edited by Ramie Gougeon and Maureen Meyers. University of Tennessee Press, Knoxville.

Wright, Alice P. 2014 History, Monumentality, and Interaction in the Appalachian Summit Middle Woodland. American Antiquity 79(2):277–294.

Wright, Patti 2003 Preservation or Destruction of Plant Remains by Carbonization? Journal of Archaeological Science 30:577-583.

Yarnell, R. A., and M. J. Black 1985 Temporal Trends Indicated by a Survey of Archaic and Woodland Plant Food Remains from Southeastern North America. Southeastern Archaeology 4:93-106. 1993 The Importance of Native Crops during the Late Archaic and Woodland Periods. In Foraging and Farming in the Eastern Woodlands, edited by C. Margaret Scarry, pp. 13–26. University of Florida, Gainesville.

293

APPENDIX

294

Table A.1. Mississippian Pit Features (n=115) from the Townsend Sites (Table 44 from Koerner and Sullivan 2011:200–202). Site Feature Northing* Easting* Depth* Volume# Category^ 40BT0089 25 209 227 27 1600 1 40BT0089 89 140 142 23 525 2 40BT0089 124 86 124 66 1185 2 40BT0089 127 150 141 14 310 2 40BT0089 164 76 82 67 245 2 40BT0089 235 200 200 40 10 1 40BT0089 251 122 120 182 590 3 40BT0090 10 117 111 18 -- 2 40BT0090 69 54 67 34 140 2 40BT0090 85 214 254 29 270 1 40BT0090 106 69 70 32 144 Hearth 40BT0090 212 161 232 26 822 1 40BT0090 267 715 220 39 1157 4 40BT0090 269 158 129 29 390 2 40BT0090 294 152 180 169 5013 3 40BT0090 459 56 51 30 135 2 40BT0090 473 108 86 62 703 2 40BT0090 510 70 84 30 150 2 40BT0090 561 92 92 10 45 2 40BT0090 580 97 102 22 155 2 40BT0090 588 600 106 122 854 5 40BT0090 697 194 154 37 375 1 40BT0090 729 270 206 28 95 1 40BT0090 873 126 118 130 130 3 40BT0090 1109 114 220 28 651 2 40BT0090 1498 212 140 20 152 1 40BT0090 1531 142 197 38 816 Hearth 40BT0090 1590 81 82 28 334 2 40BT0090 1610 103 396 116 669 5 40BT0090 1622 60 65 25 98 2 40BT0090 1746 126 34 15 175 2 40BT0090 1747 165 117 15 234 2 40BT0090 1749 86 114 19 147 2 40BT0090 1765 106 130 25 174 2 40BT0090 1767 105 96 32 296 2 40BT0090 1767A 98 96 29 175 2 40BT0090 1795 110 109 52 710 2 40BT0090 1825 126 146 23 454 2 40BT0090 1851 139 112 34 405 2 40BT0090 1856 132 101 30 220 2

295

Table A.1 (cont.). Mississippian Pit Features (n=115) from the Townsend Sites (Table 44 from Koerner and Sullivan 2011:200–202). Site Feature Northing* Easting* Depth* Volume# Category^ 40BT0090 2001 68 64 43 187 2 40BT0090 2002 69 68 46 210 2 40BT0090 2120 23 18 5 2 2 40BT0091 13 352 187 56 1190 1 40BT0091 24 237 208 38 1041 1 40BT0091 37 198 150 41 885 1 40BT0091 48 289 134 50 952 Hearth 40BT0091 67 156 173 21 465 1 40BT0091 68 84 205 . . Hearth 40BT0091 71 62 92 31 170 2 40BT0091 72 182 186 25 360 1 40BT0091 88 230 273 20 605 1 40BT0091 89 101 87 19 85 2 40BT0091 92 71 85 17 65 2 40BT0091 105 72 92 75 359 2 40BT0091 111 134 182 40 513 1 40BT0091 124 273 206 70 1767 1 40BT0091 131 98 102 63 622 2 40BT0091 138 143 153 52 643 1 40BT0091 139 161 195 44 733 1 40BT0091 140 198 88 12 1690 2 40BT0091 142 97 154 38 150 2 40BT0091 157 78 66 46 90 2 40BT0091 248 54 50 73 118 2 40BT0091 260 77 79 37 260 2 40BT0091 268 240 158 52 2112 1 40BT0091 272 242 144 11 810 1 40BT0091 273 108 117 10 34 2 40BT0091 269 178 154 26 530 1 40BT0091 274 230 197 24 1077 1 40BT0091 294 125 129 23 350 2 40BT0091 343 230 155 85 2205 1 40BT0091 362 134 139 106 2007 3 40BT0091 371 172 148 24 820 1 40BT0091 375 32 34 . . . 40BT0091 479 114 104 41 370 2 40BT0091 480 90 106 33 225 2 40BT0091 482 53 53 11 25 2 40BT0091 499 76 72 15 55 2 40BT0091 504 107 105 20 125 2 40BT0091 505 32 28 6 17 Hearth 40BT0091 507 64 69 14 55 2

296

Table A.1 (cont.). Mississippian Pit Features (n=115) from the Townsend Sites (Table 44 from Koerner and Sullivan 2011:200–202). Site Feature Northing* Easting* Depth* Volume# Category^ 40BT0091 514 86 88 13 45 2 40BT0091 516 100 292 22 260 2 40BT0091 517 179 210 33 860 1 40BT0091 520 122 139 20 108 2 40BT0091 528 213 127 17 300 1 40BT0091 530 77 103 17 64 Hearth 40BT0091 542 138 34 25 120 2 40BT0091 544 230 150 20 640 Hearth 40BT0091 581 121 114 19 . 2 40BT0091 582 98 105 18 159 2 40BT0091 584 185 248 28 703 1 40BT0091 585 96 113 10 64 2 40BT0091 591 52 63 21 73 2 40BT0091 599 69 83 11 30 2 40BT0091 623 126 136 31 452 2 40BT0091 629 124 131 13 205 2 40BT0091 631 74 80 25 138 2 40BT0091 632 200 146 15 190 Hearth 40BT0091 635 52 44 21 38 2 40BT0091 650 90 106 23 55 2 40BT0091 672 48 75 15 42 2 40BT0091 678 56 67 16 67 2 40BT0091 690 114 127 19 205 2 40BT0091 705 73 128 12 47 2 40BT0091 706 63 55 6 8 2 40BT0091 711 46 50 25 90 2 40BT0091 717 110 79 26 190 2 40BT0091 753 45 50 18 49 2 40BT0091 760 46 46 9 38 2 40BT0091 762 300 250 14 600 1 40BT0091 782 67 27 28 130 2 40BT0091 805 54 120 17 . 2 40BT0091 807 74 99 97 465 2 40BT0091 812 212 90 27 492 2 40BT0091 813 113 109 17 200 2 40BT0091 829 150 239 28 710 1 40BT0091 832 68 65 20 68 2 40BT0091 851 42 52 9 46 2 40BT0091 861 87 144 12 98 2 * measurements in centimeters; # measurements in liters; ^ categories based on statistical groups

297

Table A.2. Domestic Architecture from the Townsend Sites (Table 39 from Koerner and Sullivan 2011:200–202). Site/ Excavation Description Structure Area 40BT90 2 B-OB Rectangular/square, single set post domestic dwelling located just outside Palisade 1 (West Village) on the Office Block. Structure 2 is adjacent to Structure 5. 3/4 B Structure 3 is an irregular outline, single post dwelling. Structure 4 was initially thought to be a separate structure from Structure 3 but after excavation was included as part of Structure 3. 5 B-OB Rectangular/square, single set post domestic dwelling. Southwest of Structure 2. 6 C-2 Rectangular single post dwelling 7 C-2 Rectangular single post dwelling 14 C-1 Square single post dwelling 15 C-1 Square single post dwelling 14/16 & C-1 After excavation the post holes were subsumed into Structures 14 and 15/16 15. 18 C-1 Rectangular/square, single set post domestic dwelling. 57 B-2 L-shaped segment of single set posts. Possible domestic dwelling just southwest of Structure 52. A single shell tempered ceramic sherd from PM 91. 58 B-4 Rectangular single post dwelling 60 B-3 Rectangular/square, single set post domestic dwelling, just northwest of Structure Area 1 from the Archaic/Woodland Volume. 61 B-3 Rectangular/square, single set post domestic dwelling, just northwest of Structure Area 1 from the Archaic/Woodland Volume. 40BT91 5 A-2 Square single post dwelling 9 C-6 Semi-subterranean structure basin with midden. The shape is unknown because postmolds were not preserved. 10 B-4 Square, single post dwelling 11 C-9 Rectangular single post dwelling 12 C-9 Long, arcing segment of posts adjacent to Structure 11. PM 1632 contained shell tempered ceramics. 18/41 C-10 Structure 18 is a square, single post dwelling. After excavation, Structure 41 was determined to be an interior partition of Structure 18. 22 C-10 Square single post dwelling 23 C-6 Square single post dwelling 24 C-6 Rectangular single post portico(?) 28 C-6 Rectangular/square, single set post domestic dwelling. 29 C-6 Rectangular/square, single set post domestic dwelling.

298

Table A.2 (cont.). Domestic Architecture from the Townsend Sites (Table 39 from Koerner and Sullivan 2011:200–202). Site/ Excavation Description Structure Area 31 C-10 Square single post dwelling 36/37 A-1 A palimpsest of posts located in the south-central portion of Skittletown that were initially defined as two separate structures. After excavation was determined most likely to be a single, large rectangular/square structure. PM 5024 contained shell tempered "filleted" lug handle and PM 5017 contained a Small Triangular Cluster PP/K. 39, 42, A-1 A palimpsest of posts located along the western edge of Skittletown 43, & 44 that were initially defined as four separate structures. After excavation of the posts it was exceedingly difficult to determine which posts belonged to each structure. Several of posts in the area contained shell tempered ceramics.

299

Table A.3. Community Buildings from the Townsend Site of 40BT91 (adapted from Koerner and Sullivan 2011:170–171). Excavation Area Location Description Structure 1 Village 1 Rectangular building (with rounded corners). 149 single-set posts including 102 exterior and 47 interior. Dimensions of 8.3 m X 13.0 m. No associated features.

Structure 13 Village 1 Rectangular building (with rounded corners). 109 single-set posts including 72 exterior and 37 interior. Dimensions of 8.8 m X 9.8 m. Associated Feature 591.

300

Table A.4. Corncrib Structures from the Townsend Sites (adapted from Koerner and Sullivan 2011:Table 43 and 44). Major Minor Post Post Exterior Interior Site Structure Axis Axis Diameter Diameter Location Posts Posts (m) (m) (range/cm) (avg./cm) 40BT90 13 2.75 2.71 16 2 13–19.5 17 Farmstead 4 24 2.32 2.26 16 9–22 13.6 Farmstead 2 44 2.80 2.60 11 10–29 22.1 45 2.45 2.38 18 14–23 16.8 46 1.85 1.80 8 17–24.5 19.4 50 2.74 2.56 10 17.5–27.5 22.3 53 2.47 2.39 19 2 17.5–27.5 12.5 Farmstead 2 54 2.15 2.06 17 1 7–21 15.7 Farmstead 2 55 1.52 1.43 8 12–19 16 Farmstead 2 40BT91 2 2.03 1.77 9 2 12–22.5 17.8 Village 1 7 3.10 2.90 18 1 7.3–21 14.7 15 2.30 2.00 6 10–13.5 11.5 16 2.09 1.86 10 11.3–19 16.3 Village 1 17 2.28 2.23 18 12.5–25 17.4 Village 1 20 2.29 2.25 11 12–19.5 15.8 Village 1 21 2.18 2.13 10 11–19.5 15.1 Village 1 34 2.19 1.85 10 13–19 16.4 Village 1

301

Table A.5. Primary Enclosures at the Townsend Sites (adapted from Koerner and Sullivan 2011:Table 38). Post Number Posthole hole Building Palisade of (average Bastions (average episodes Postholes int. cm) dia. cm) West Enclosure: 1/A 181 13.5 50.3 No 1

Central Enclosure: B 135 16.2 39.2 No 1

East Enclosure/Village 2: C/D 269 13.5 40.5 No 2

Skittletown/Village 1: A-H 892 16 10.8 3 4 (?)

302

Table A.6. Mississippian Plant Remains Recovered by Context from the Townsend Sites.

Plant Wood Feature Bag Weight Weight Weight Number Number (g) (g) Common Name Count (g) 40BT89: 25 99-1094-4 2.22 0.54 Acorn 1 0.00 Cane 5 0.03 Corn cupule 174 1.28 Corn cupule/glume cf. 4 0.01 Corn glume 10 0.04 Corn kernel cf. 1 0.00 Cucurbit rind 1 0.00 Grass family 1 0.00 Grass family cf. 3 0.01 Hickory 2 0.01 Maypop 1 0.01 Nightshade cf. 1 0.00 Persimmon 1 0.00 Pine cone 5 0.02 Pitch 7 0.03 Tobacco cf. 1 0.00 Twig 11 0.19 Unidentifiable 10 0.04 Unidentifiable seed 7 0.00 Walnut family 2 0.01 25 99-1118-6 14.51 4.05 Bark 16 0.42 Cane 24 0.34 Carpetweed 2 0.00 Catkin 1 0.00 Corn cupule 668 8.48 Corn cupule cf. 9 0.03 Corn glume 102 0.34 Corn kernel 4 0.07 Cucurbit rind cf. 1 0.00 Grass family 1 0.00 Persimmon seed cf. 1 0.00 Pine cone 45 0.14 Pitch 22 0.12 Stem 16 0.19 Unidentifiable 52 0.46 Unidentifiable seed 16 0.00 Unidentified seed 42 0.01 Walnut family 3 0.01 25 99-1126-89 1.00 0.46 Corn cupule 3 0.00 Pitch 37 0.00

303