Sedentism and Subsistence in the Late Archaic: a Study of Seasonality, Quahog Clam Exploitation, and Resource Scheduling Alexandra L

Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2012 Sedentism and Subsistence in the Late Archaic: A Study of Seasonality, Quahog Clam Exploitation, and Resource Scheduling Alexandra L. Parsons

Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected] THE FLORIDA STATE UNIVERSITY

COLLEGE OF ARTS AND SCIENCES

SEDENTISM AND SUBSISTENCE IN THE LATE ARCHAIC: A STUDY OF SEASONALITY,

QUAHOG CLAM EXPLOITATION, AND RESOURCE SCHEDULING

ALEXANDRA L. PARSONS

A dissertation submitted to the Department of Anthropology in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Degree Awarded: Spring Semester, 2012 Alexandra L. Parsons defended this dissertation on March 1, 2012

The members of the supervisory committee were:

Rochelle A. Marrinan

Professor Directing Dissertation

Daniel J. Pullen

University Representative

Glen H. Doran

Committee Member

Lynne A. Schepartz

Committee Member

The Graduate School has verified and approved the above‑named committee members, and certifies that the dissertation has been approved in accordance with university requirements.

Alexandra L. Parsons

iii

This manuscript is dedicated to my mother, Corinne Royce Dewey.

iv ACKNOWLEDGEMENTS

I would like to thank the National Science Foundation for providing me with a

Dissertation Improvement Grant. This grant permitted me to collect the modern comparative collection in St. Augustine, and funded much of the chemical analyses. The grant also provided funds for general laboratory supplies needed to conduct the research. I would also like to thank the Eisele Foundation for two Eisele Dissertation Research grants. These grants provided travel funds to borrow archaeological collections and to see the Guana shell ring, and also provided funds for equipment and some of the chemical analyses.

I am greatly indebted to my advisor, Rochelle A. Marrinan, for her guidance and assistance throughout my tenure at Florida State University. Rochelle has been an extremely generous advisor and has provided me with many opportunities to further my research and skills. In 2006, Rochelle arranged for me to help on an excavation at the Grand shell ring. This experience resulted in the first of several clam seasonality studies, which have been my primary research focus. Rochelle also helped me to secure an internship on St. Catherines Island, where

I got my first glimpse of shell rings in Georgia. From the planning stages of my research through defense of the dissertation, she has always offered sound advice whenever I encountered a problem. Rochelle has been extremely helpful in every aspect of this research, and she even helped me acquire funds to begin my research. By taking the courses she instructed, conducting research and writing reports with her, and finishing this dissertation under her guidance I have become a better anthropologist and a better scholar. I have greatly enjoyed getting to know her and working with her, and I cannot thank her enough for all her kindness and support over the years.

I would also like to thank the members of my committee, Glen H. Doran and Lynne A

Schepartz, for their thoughtful comments and suggestions. This manuscript has been greatly improved by their participation. Both Glen and Lynne remained on my committee and saw me through until the end, despite all the tumult, changes, and uncertainty for Anthropology graduate students over the past few years. I am extremely grateful to Daniel Pullen, who kindly agreed to serve as my University Representative in the 11th hour before my defense. I am

v very thankful for his participation and for his comments and suggestions. I would also like to thank Joseph Donoghue for his comments and suggestions on this manuscript and my qualifying exams.

I am greatly indebted to Irv Quitmyer of the Florida Museum of Natural History. Irv’s exemplary research on clam seasonality studies was the impetus for this dissertation. I have greatly enjoyed our discussions about clam biology and clam use in prehistory. His guidance on creating a modern comparative clam collection and advice about methodology greatly improved this research. Irv provided feedback on my National Science Foundation Grant proposal that helped me to receive funding. Irv also allowed me to use his equipment to conduct isotopic testing and to cross‑section the clams. I am very grateful to him for all his advice, assistance, and good company.

I would like to extend my sincerest thanks to Mike Russo of the Southeast Archeological

Center. Six years ago I began working on my Master’s thesis, and Mike showed me how to cross‑section and assess clams for seasonality studies. His extensive work on Late Archaic shell rings in the Southeast has generated considerable discussion on the topic, and has served as an important resource for my research. Mike generously allowed me to analyze the clams from his excavations at Guana in 2001. Mike provided information on the excavation and also permitted me to use his maps of the Guana shell ring. I am very grateful for all his advice and assistance over the years.

I would like to thank Becky Saunders of Louisiana State University. Becky allowed me to analyze the clams from her excavations at Guana in 2005, and she prepared the loan for these clams. I would also like to thank Vicki Rolland, who participated in both excavations at Guana.

I first met Vicki during excavations at the Grand shell ring in 2006, and I am always happy to see her or hear from her. Vicki has always been willing to assist me in any way that she could, and for that I am truly grateful.

I would like to thank Dave Thomas for his exceptional research on St. Catherines Island and for all his advice, guidance, and generosity. Working with his crew on the McQueen shell ring and attending two of the Caldwell Conferences on St. Catherines provided me with new insight into a variety of archaeological topics. His research on reservoir effects in the

vi Southeastern U.S. and his advice on correcting and re‑calibrating radiocarbon dates were especially helpful in preparing this dissertation.

I would like to thank Justin Ellenberger, from the Florida Fish and Wildlife Commission for taking my husband and me out to see the Guana shell ring. Even though it was a miserable place with more mosquitoes than I’ve ever seen in my life, going out to the site was very helpful. I am especially appreciative of the bug spray he gave us!

I am forever indebted to Phil Cubbedge for his assistance in procuring the modern clam collection for this dissertation. Phil not only escorted me into the marsh in St. Augustine, he helped me collect clams in almost every month of 2010. When I began this research I had a lot of book‑knowledge about clams, but this information wasn’t very helpful for finding clams in the marsh. When I first started collecting clams, I couldn’t tell the difference between a crab hole and a clam keyhole, but by December of 2010 I had become a proficient clam gatherer, thanks to Phil’s instruction. I have greatly benefitted from Phil’s extensive knowledge, and I have a much broader appreciation for the marsh ecosystem as a result.

I would like to thank my friends in Tallahassee for all their support. Katie Miyar,

Eduardo Miyar, Ian Pawn, Sarah Liko, Ivy Hepp, Guy Hepp, Julie Byrd, Ryan Duggins, Dan

Seinfeld, Josh Englehardt, Ermal Liko, Bridget McDonnell, Giovanna Englehardt, Brew

Schoonover, and Jack Juliet, I am so happy I got to know all of you and I am very grateful for your commiseration and support throughout this process! Many of you were directly involved in this research and were extremely helpful when it came to cleaning (i.e., eating) the quahog clams I brought back for my research. I greatly enjoyed our clam feasts and your great company.

I would like to thank my mom, Corinne Dewey for her endless support and encouragement during this process. I owe my love of books and learning to her, and I wouldn’t have pursued a graduate degree without her support. While I was conducting this research, she always had something positive to say, even when circumstances were difficult. I would also like to thank my father, Dale Dewey, for all his support and encouragement. I couldn’t have done this without them.

vii I would like to thank my sister, Virginia Carr for her hard work and assistance in gathering clams for the modern comparative collection. She bravely followed me into the marsh and became an expert clam gatherer. I would also like to thank her for her untiring sense of humor that got me through many hours of analyzing clam shells.

Finally, I would like to thank my husband, Timothy Parsons. Tim was there during every stage of this research. He accompanied me out to the mosquito‑ridden Guana shell ring and helped me collect live clams from the marsh on several occasions. He even helped me cook them a few times – despite his hatred of clams. I am really sorry about that batch of clams that went bad – the smell of them steaming open out on our porch was pretty horrific! Tim created nearly all of the maps in this dissertation and helped me edit many of the drawings. Tim also found a way to smooth out the cross‑sections of rough cuts on some of the clams from the modern comparative collection, which was extremely helpful. In addition to all this, Tim was always encouraging, even when I didn’t think I could possibly finish this dissertation by the

University’s deadline. I couldn’t have done this without all his love and support. Tim, you are the best part of my life and I am so happy that we are starting a family together!

viii TABLE OF CONTENTS

LIST OF TABLES ...... XIII

LIST OF FIGURES ...... XVI

ABSTRACT ...... XXI

CHAPTER 1 INTRODUCTION ...... 1 PROBLEM ORIENTATION ...... 1 AN OVERVIEW OF FAUNAL SEASONALITY STUDIES ...... 3 SUMMARY OF RESEARCH ...... 7

CHAPTER 2 THEORETICAL BACKGROUND ...... 10 CONCEPTUAL FRAMEWORK OF SEDENTISM ...... 10 OPTIMAL FORAGING THEORY ...... 12 MOLLUSKS AT ARCHAEOLOGICAL SITES ...... 17

CHAPTER 3 ARCHAEOLOGICAL BACKGROUND ...... 20 THEORETICAL FRAMEWORK OF THE LATE ARCHAIC SOUTHEAST ...... 20 Introduction ...... 20 The New Archaic ...... 22 SEA LEVEL RISE AND SETTLEMENT PATTERNS ...... 23 THE EUROPEAN MESOLITHIC ...... 30 THE GUANA SHELL RING ...... 34 Location and Physical Characteristics ...... 34 Previous Research ...... 37 Cultural Affiliations and Chronology ...... 38

CHAPTER 4 CONTEXTUALIZING THE GUANA SHELL RING ...... 43 CHARACTERISTICS OF SHELL RINGS OF THE SOUTHEAST UNITED STATES ...... 43 Introduction ...... 43 Temporality of Shell Rings ...... 44 SHELL RING RESEARCH AND RING FUNCTION ...... 45 SEASONALITY STUDIES AT LATE ARCHAIC SHELL RINGS ...... 50

CHAPTER 5 CLAM BIOLOGY AND SEASONALITY STUDY ...... 56 CLAM BIOLOGY ...... 56 Mercenaria mercenaria, Mercenaria campechiensis, and their Hybrids ...... 56 Life Cycle and Reproduction of Mercenaria spp...... 61 QUAHOG SEASONALITY STUDIES ...... 63

ix Shell Growth and Increment Deposition ...... 63 History of Quahog Seasonality Studies ...... 65 The Modern Comparative Clam Collection ...... 67

CHAPTER 6 METHODOLOGY ...... 71 MODERN CLAM COLLECTION ...... 71 Foraging for Clams ...... 71 Water Temperature and Salinity in the Matanzas River ...... 79 Preparing the Clams for Analysis ...... 80 THE ARCHAEOLOGICAL CLAM SAMPLE ...... 83 CROSS‑SECTIONING THE CLAMS ...... 85 ANALYZING THE CLAMS ...... 87 Macroscopic Analysis ...... 87 Isotopic Analysis ...... 91

CHAPTER 7 RESULTS ...... 94 ISOTOPIC ANALYSIS ...... 94 Modern Clam Number Four ...... 96 Modern Clam Number Three ...... 96 Modern Clam Number 11 ...... 99 Summary of ∂18O Isotopic Profiles for Modern Clams from the St. Augustine Collection ...... 101 Archaeological Clam Number 88 ...... 102 Archaeological Clam Number 222A ...... 104 Archaeological Clam Number 222B ...... 106 Archaeological Clam 336 ...... 108 Summary of ∂18O Isotopic Profiles for Archaeological Clams from the Guana Shell Ring ...... 108 Discussion of ∂18O Profiles for Modern Clams from the St. Augustine Collection and Archaeological Clams from the Guana Shell Ring ...... 109 MODERN COMPARATIVE COLLECTION ...... 113 Gathering Live Clams ...... 113 Water Temperature and Salinity Data ...... 115 Measurements ...... 120 Age ...... 125 Seasonality ...... 127 ARCHAEOLOGICAL CLAMS FROM THE RING INTERIOR ...... 140 Clam Measurements ...... 141 Age ...... 143 Clam Seasonality ...... 144 ARCHAEOLOGICAL CLAMS FROM THE RING DEPOSIT ...... 148

x Clam Measurements ...... 150 Age ...... 151 Clam Seasonality ...... 152 Spatial Comparisons of the Ring Deposit ...... 156 SUMMARY OF RESULTS ...... 158

CHAPTER 8 DISCUSSION ...... 163 ST. AUGUSTINE COLLECTION ...... 163 Seasonal Profiles ...... 163 Age Data ...... 167 CLAM COLLECTION AT THE GUANA SHELL RING ...... 168 Population Structure of Clams from the Guana Shell Ring ...... 168 Patterns of Seasonal Clam Exploitation at the Guana Shell Ring ...... 171 Contextualizing the Seasonal Exploitation of Clams at the Guana Shell Ring ...... 173 EXPLANATORY MODEL FOR THE SEASONAL USE OF CLAMS IN NORTHEAST FLORIDA ...... 178 SEDENTISM AT THE GUANA SHELL RING ...... 180 SUMMARY ...... 182

CHAPTER 9 CONCLUSIONS ...... 184 SUMMARY ...... 184 FUTURE RESEARCH ...... 184 The St. Augustine Collection ...... 184 The Guana Shell Ring Site ...... 185 SUMMARY OF FINDINGS ...... 187 The St. Augustine Comparative Collection ...... 187 The Guana Shell Ring Site ...... 190 SEDENTISM AT LATE ARCHAIC SITES IN THE SOUTHEAST ...... 194

APPENDIX A RADIOCARBON DATES OBTAINED FOR THIS DISSERTATION ...... 196

APPENDIX B LOCATION OF SAMPLES FOR OXYGEN ISOTOPIC TESTING ...... 197

APPENDIX C WATER TEMPERATURE DATA FOR ST. AUGUSTINE ...... 201

APPENDIX D DATA FOR THE MONTH OF OCTOBER FOR THE ST. AUGUSTINE COLLECTION (AQUACULTURE CLAMS) ...... 203

APPENDIX E HINGE LENGTH DATA FOR THE ST. AUGUSTINE COLLECTION AREAS ...... 205

APPENDIX F GROWTH PHASE DATA BY MONTH FOR THE ST. AUGUSTINE COLLECTION ...... 206

xi APPENDIX G DATA FOR THE ARCHAEOLOGICAL CLAMS IN THE RING INTERIOR AND RING DEPOSIT ...... 213

REFERENCES ...... 214

BIOGRAPHICAL SKETCH ...... 234

xii LIST OF TABLES

1.1 Selected Influential Aquatic Faunal Seasonality Studies ...... 4

1.2 Classification of Methodologies to Determine Faunal Seasonality ...... 6

1.3 Russo’s 1998 Classification Methodologies to Determine Faunal Seasonality ...... 7

2.1 Comparison of Calories, Carbohydrates, Fat, and Protein of Oysters, Quahog Clams, and Catfish ...... 18

2.2 Comparison of Yields of Shellfish, Fish and Deer ...... 18

3.1 Corrected and Calibrated Radiocarbon Dates for the Guana Shell Ring ...... 40

4.1 Fauna Used in Seasonality Studies Discussed in this Dissertation ...... 51

4.2 Seasonality Results at Shell Rings in the Southeast ...... 54

5.1 Tidal Predictions for 2010 ...... 69

6.1 Growth Phase Classifications ...... 88

6.2 Growth Phase Characterizations ...... 90

7.1 Statistics for ∂18O Values of Modern and Archaeological Clams ...... 110

7.2 Dates and Times of Clam Collections ...... 114

7.3 Occurrence of Clam Signs ...... 115

7.4 Shell Measurements (in millimeters) for the St. Augustine Collection (excluding October) ...... 120

7.5 Shell Measurement Relationships for the St. Augustine Collection (excluding October) .. 121

7.6 Age Data for the St. Augustine Collection (excluding October) ...... 125

7.7 St. Augustine Collection Age by Collection Area ...... 127

7.8 Increment Distributions for Winter – Spring, Summer – Fall, and Annual ...... 129

xiii

7.9 Outcomes for Increment Assessment by Season for the St. Augustine Collection ...... 132

7.10 Growth Increment Distributions by Season for the St. Augustine Collection ...... 132

7.11 Outcomes of Growth Phase Assessments for the St. Augustine Collection ...... 134

7.12 Growth Phase Distributions by Count for the St. Augustine Collection ...... 134

7.13 Growth Phase Distributions by Percent for the St. Augustine Collection ...... 134

7.14 Outcomes of Growth Phase Assessments for Winter – Spring and Summer – Fall in the St. Augustine Collection ...... 140

7.15 Growth Distributions for Winter – Spring and Summer ‑ Fall ...... 140

7.16 Shell Measurements (in millimeters) for the Ring Interior ...... 142

7.17 Shell Measurement Relationships for the Ring Interior ...... 142

7.18 Hinge Length Measurements (in millimeters) for Ring Interior Levels 1‑3 and 4‑6 ...... 142

7.19 Age Data for the Ring Interior ...... 144

7.20 Age Data for the Ring Interior Levels 1‑3 and 4‑6 ...... 144

7.21 Growth Phases for the Ring Interior ...... 145

7.22 Growth Phases for Levels 1‑3 and 4‑6 in the Ring Interior ...... 146

7.23 Shell Measurements (in millimeters) for the Ring Deposit ...... 150

7.24 Shell Measurement Relationships for the Ring Deposit ...... 151

7.25 Hinge Length Measurements (in millimeters) for Ring Deposit Levels 1‑5 and 6‑12 ...... 151

7.26 Age Data for the Ring Deposit ...... 152

7.27 Age Data for Ring Deposit Levels 1‑5 and 6‑12 ...... 153

7.28 Growth Phases for the Ring Deposit ...... 153

7.29 Growth Phases for Levels 1‑5 and 6‑12 in the Ring Deposit ...... 155

xiv

7.30 Age Distributions for the West Arm, Ring Interior, and East Arm of the Ring ...... 156

7.31 Hinge Length Measurements (in millimeters) for Ring Deposit Areas ...... 157

7.32 Growth Phase Distributions for Areas of the Guana Shell Ring ...... 157

7.33 Distribution of Growth Phases for Shovel Test 440N, 410E ...... 160

8.1 Mean Ages for Modern Comparative Clam Collections ...... 168

8.2 Age Data for Archaeological Clams from the Guana Shell Ring ...... 169

8.3 Size Data for Archaeological Clams from the Guana Shell Ring ...... 170

8.4 Seasonality Assessments of Selected Archaeological Sites in the Georgia Bight ...... 175

8.5 Clam Seasonality Assessments at Selected Sites in the Georgia Bight ...... 175

8.6 Seasonality of Clams, Vertebrate Resources, and Floral Remains at Selected Shell Rings in the Georgia Bight ...... 181

A.1 Radiocarbon Dates from the Guana Shell Ring Obtained for this Dissertation ...... 196

C.2 Monthly Low, High, and Average Water Temperature for 2010 (St. Augustine) ...... 202

C.3 Monthly Low, High, and Average Salinity for 2010 (St. Augustine) ...... 202

D.1 Shell Measurements (in millimeters) for the October St. Augustine Collection ...... 203

D.2 Shell Measurement Relationships for the October St. Augustine Collection ...... 203

D.3 Age Data for the October St. Augustine Collection ...... 204

E.1 Hinge Length for Area A, the Subsample of Area B, and all of Area B ...... 205

F.1 Growth Phase Distributions for the St. Augustine Collection by Percent ...... 206

F.2 Growth Phase Distributions for the St. Augustine Collection by Count ...... 206

G.1 Shell Measurements (in millimeters) for the Ring Interior for Upper and Lower Levels ... 213

G.2 Shell Measurements (in millimeters) for the Ring Deposit for Upper and Lower Levels ... 213

xv LIST OF FIGURES

1.1 Map of the Guana Shell Ring and Selected Modern Comparative Clam Collections in the Georgia Bight ...... 9

3.1 Shell Density Map of the Guana Shell Ring ...... 36

3.2 Approximate Locations of Radiocarbon Dates ...... 41

5.1 Clam Terminology ...... 57

5.2 Injured Clam ...... 59

5.3 Ontogenetic Sequence of a Quahog Clam ...... 64

6.1 Quad Map of the Matanzas River ...... 72

6.2 Ortho Map of Collection Areas ...... 73

6.3 Ortho Map of Collection Area A ...... 74

6.4 Ortho Map of Collection Area B ...... 75

6.5 Satellite Imagery of Collection Area B Sedimentation ...... 76

6.6 Clam Keyhole and Clam ...... 77

6.7 Clam Keyhole ...... 77

6.8 Alternative Shape Keyhole ...... 78

6.9 Shell Length Measurement ...... 81

6.10 Shell Height Measurement ...... 82

6.11 Hinge Length Measurement ...... 82

6.12 Minimum Usable Valve Fragment ...... 83

6.13 Dead Clam Shell ...... 84

xvi 7.1 Profile of ∂18O Vales for Modern Clam Number 4 ...... 97

7.2 Profile of ∂18O Vales for Modern Clam Number 3 ...... 97

7.3 Profile of ∂18O Vales for Modern Clam Number 11 ...... 99

7.4 Profile of ∂18O Vales for Archaeological Clam Number 88 ...... 103

7.5 Profile of ∂18O Vales for Archaeological Clam Number 222A ...... 105

7.6 Profile of ∂18O Vales for Archaeological Clam Number 222B ...... 107

7.7 Profile of ∂18O Vales for Archaeological Clam Number 336 ...... 107

7.8 Boxplots of ∂18O Values for Modern and Archaeological Clams ...... 111

7.9 Average Water Temperatures in 2010 ...... 116

7.10 High and Low Water Temperatures in 2010 ...... 116

7.11 Average Salinity in 2010 ...... 117

7.12 High and Low Salinity Levels in 2010 ...... 118

7.13 Average Salinity and Water Temperature in 2010 ...... 118

7.14 Seasonal Average Water Temperature and Salinity during 2010 ...... 119

7.15 Length and Height Scatterplot for the St. Augustine Collection ...... 122

7.16 Length and Hinge Width Scatterplot for the St. Augustine Collection ...... 123

7.17 Height and Hinge Width Scatterplot for the St. Augustine Collection ...... 124

7.18 St. Augustine Collection Clam Age Distribution (except October) ...... 126

7.19 Scatterplot of Hinge Width and Clam Age in the St. Augustine Collection ...... 128

7.20 Winter – Spring Growth Increment Distribution for the St. Augustine Collection ...... 129

7.21 Summer ‑ Fall Growth Increment Distribution for the St. Augustine Collection ...... 130

7.22 Annual Growth Increment Distribution for the St. Augustine Collection ...... 130

xvii

7.23 Percentage of Opaque and Translucent Growth Increments per Month in 2010 ...... 131

7.24 Growth Increment Distribution by Season for the St. Augustine Collection ...... 132

7.25 Seasonal Temperature Averages and Increment Frequencies ...... 133

7.26 Winter Growth Profile for the St. Augustine Collection ...... 135

7.27 Spring Growth Profile for the St. Augustine Collection ...... 136

7.28 Alternate Spring Growth Profile for the St. Augustine Collection ...... 136

7.29 Summer Growth Profile for the St. Augustine Collection ...... 137

7.30 Fall Growth Profile for the St. Augustine Collection ...... 137

7.31 Annual Growth Profile for the St. Augustine Collection ...... 138

7.32 Winter – Spring Growth Profile for the St. Augustine Collection ...... 139

7.33 Summer – Fall Growth Profile for the St. Augustine Collection ...... 139

7.34 Ring Interior Clam Ages ...... 143

7.35 Growth Profile for Clams in the Ring Interior ...... 146

7.36 Growth Profile for the Ring Interior Levels 1‑3 ...... 147

7.37 Growth Profile for the Ring Interior Levels 4‑6 ...... 147

7.38 Distribution of Ring Deposit Areas, Shovel Tests, and Test Unit ...... 149

7.39 Ring Deposit Clam Ages ...... 152

7.40 Growth Profile for Clams in the Ring Deposit ...... 154

7.41 Growth Profile for the Ring Deposit Levels 1‑5 ...... 155

7.42 Growth Profile for the Ring Deposit Levels 6‑12 ...... 156

7.43 Distribution of Clam Seasonality throughout the Ring Deposit ...... 159

xviii 8.1 Seasonal Growth Profiles for the Kings Bay, St. Augustine, and Indian River Modern Comparative Collections ...... 164

9.1 Map of the Guana Shell Ring and Modern Comparative Clam Collections Discussed in Chapter 9 ...... 189

B.1 Modern Clam Number 4 ...... 197

B.2 Modern Clam Number 3 ...... 197

B.3 Modern Clam Number 11 ...... 198

B.4 Archaeological Clam Number 88 ...... 198

B.5 Archaeological Clam Number 222A ...... 199

B.6 Archaeological Clam Number 222B ...... 199

B.7 Archaeological Clam Number 336 ...... 200

C.1 Water Temperature Data for Fort Matanzas from 2006 to 2010 ...... 201

F.3 Growth Phase Distribution for the Month of January ...... 207

F.4 Growth Phase Distribution for the Month of February ...... 207

F.5 Growth Phase Distribution for the Month of March ...... 208

F.6 Growth Phase Distribution for the Month of April ...... 208

F.7 Growth Phase Distribution for the Month of May ...... 209

F.8 Growth Phase Distribution for the Month of June ...... 209

F.9 Growth Phase Distribution for the Month of July ...... 210

F.10 Growth Phase Distribution for the Month of August ...... 210

F.11 Growth Phase Distribution for the Month of September ...... 211

F.12 Growth Phase Distribution for the Month of October ...... 211

F.13 Growth Phase Distribution for the Month of November ...... 212

xix

F.14 Growth Phase Distribution for the Month of December ...... 212

xx ABSTRACT

This research evaluates sedentism and the seasonal use of resources during the Late Archaic period in the coastal zone of the Southeastern U.S. Specifically, this research examines sedentism at the Guana shell ring near St. Augustine, Florida, which dates to approximately

3400 cal B.P. This study employs the incremental growth technique to determine the season of death of quahog clams (Mercenaria spp.), a common constituent in the ring matrix. To accurately assess season of death for quahog clams, I gathered a modern comparative clam collection from St. Augustine during every month in 2010. It appears that the maximum acceptable distance between the modern collection and archaeological site in question lies somewhere between 45‑80 km. Based on the present results, the maximum acceptable distance may be somewhere around 70 km, but this remains to be tested.

This research demonstrates that Guana occupants gathered clams during the winter and spring. Although clam gathering occurred during half of the year, I have proposed that Guana occupants likely remained at the site throughout the year based on seasonality studies at other

Late Archaic shell rings. Clam seasonality varied slightly throughout the ring deposit, and may indicate differences in clam use by household groups. Mean clam age and size at Guana declined considerably over time. This suggests that Guana occupants relied heavily on quahog clams and intensively exploited nearby clam beds.

Clam collection during the cooler months of winter and spring has been identified at most sites in northeast Florida and throughout the Georgia Bight. Often, clams were gathered exclusively during the cooler months, or they were gathered more intensively during this time.

I have proposed that this pattern of cool‑weather collection is an adaptation to maximize the returns of gathering clams. Based on my observations while assembling the modern comparative clam collection, clams create small holes in the sand, called keyholes, during cool weather. These keyholes mark the location of a clam, which reduces search time and makes clams less costly to procure. This means that clams are less costly to collect during the cooler months of winter and spring, the seasons during which most clam collection occurred in northeast Florida and throughout the Georgia Bight. This adaptation is likely the cause for the

xxi seasonal focus on clam gathering, despite continued occupation at many sites in the Georgia

Bight.

xxii CHAPTER 1

INTRODUCTION

Problem Orientation

Beginning in the late 1970s, a number of archaeologists working in the Southeastern United

States proposed that many coastal Late Archaic (5000 – 2000 B.P.) shell rings and shell middens were the products of semi‑sedentary to sedentary occupations (Colquhoun and Brooks 1986;

DePratter 1975, 1979; Marrinan 1975; Milanich and Fairbanks 1980:150; Price and Brown

1985:11‑12). Archaeologists based these propositions of sedentism on factors such as the apparent adaptation to estuarine and marine resources, large and dense shell middens, ceramic production, and formal cemeteries (Anderson 1996; Doran 2007; Marrinan 1975; Price and

Brown 1985:10). Recently, archaeological research focusing on the Archaic period as a whole has burgeoned, and much discussion has been generated about the nature of Archaic societies.

Specifically, archaeologists have rejected the normative evolutionary view of Archaic societies as relatively mobile, egalitarian groups that fell somewhere between the nomadic Paleoindians and the sedentary agriculturalists of the Woodland period (Russo 1998). Traditional concepts of the Archaic period are being newly explored utilizing methodologies that were unavailable to earlier archaeologists.

Archaeologists are interested in understanding when a society becomes sedentary because the transition to sedentism involves a reorganization of social processes that are necessary to permit extended periods of communal living. Challenges arising from the transition to a sedentary lifestyle would have necessitated organizing subsistence efforts and may have included communal construction and maintenance of weirs and traps. Sedentism also would have required subsistence scheduling and perhaps limiting the collection of particular resources to prevent shortages, determining access to resource patches and hunting

1 grounds and perhaps defending these locations if necessary. Refuse disposal would require some type of management or rules, and social problems resulting from long periods of group living would have to be mitigated. The transition to a sedentary lifestyle required more than a mere cessation of movement on the landscape – it required a reorganization of social rules and norms to allow people to remain in one location for an extended period of time.

Previously, archaeologists resorted to models of seasonal transhumance to explain a presumed aversion to coastal living during certain times of the year. Early scholars expected that Native Americans would not have consumed shellfish during the warmer months of the year based on the extant Euro‑American aversion to shellfish during warm weather. This aversion is not because shellfish are inedible or unsafe to consume during the warmer months; it is probably a result of spoilage that occurred in historic times while the shellfish were being transported inland during warm weather (Waselkov 1987:110‑111). One reason that modern people often avoid consuming bivalves during the warmer months is because the risk of exposure to waterborne diseases, such as paralytic shellfish poisoning, is higher during the summer. Shellfish poisoning is transmitted by consuming shellfish that have accumulated toxic amounts of dinoflagellates and diatoms, which occur during red tides (Parsons 2008). Many locations documented their first cases of shellfish poisoning during the 1970s and 1980s

(Claassen 1998:33), indicating that red tides may have been less widespread in prehistory (i.e., shellfish were safer to consume during warmer months in prehistory). Studies of many prehistoric Southeastern sites have shown that indigenous people consumed bivalves during warm months (e.g., Parsons in prep; Quitmyer 1995; Quitmyer et al. 1985 a and b, 1997).

Some archaeologists presumed that Native Americans would only resort to eating certain shellfish, such as coquina, during times of starvation or seasonal food shortage. Yet even today coquina is consumed with regularity, and an Internet search for coquina soup yields numerous recipes and directions on how to prepare coquina. Furthermore, shellfish provide a source of protein and other beneficial nutrients and are easy to gather. Recent zooarchaeological research has demonstrated that shellfish were an important part of the coastal diet, and has dispelled notions of the coast as uninhabitable during portions of the year.

In particular, seasonality studies have played an important role in identifying sedentary

2 occupations at a number of Late Archaic coastal sites in the Southeast (e.g., Colaninno 2010;

Marrinan 1975; Piatek 1992; Quitmyer and Jones 2012; Quitmyer et al. 1985, 1997; Russo 1991,

1998; Russo and Heide 2002; Russo et al. 1993; Russo and Quitmyer 1996, Russo and Ste. Claire

1992; Saunders 2002; Thompson 2006; Thompson and Andrus 2011). This research indicates that sedentary occupation of the coast occurred at least as early as the Late Archaic period.

An Overview of Faunal Seasonality Studies

Although seasonality studies are relatively new to archaeology (beginning in the 1960s and becoming more common in the 1980s), they have a long history in other sciences such as biology, geology, and paleontology. Table 1.1 lists a number of important influential seasonality studies on archaeological materials. In particular, shellfish have been extensively studied for their potential as seasonal indicators. Shellfish were used in seasonality studies by ecologists and biologists since the beginning of the twentieth century (Claassen 1986:21), and were among the first animals to be employed in archaeological seasonality studies in the late

1960s and early 1970s (e.g., Coutts 1970; Coutts and Higham 1971; Weide 1969). In the

Southeast United States seasonality studies became a significant research concern following the seminal publication of Quitmyer and colleagues’ (1985a) study of quahog clam seasonality at several sites in Georgia.

Seasonality studies are typically based on biological knowledge of animal behavior, morphology, or growth characteristics. Seasonality studies must assume that past climatic conditions (and animal responses to them) were similar to those we have observed in modern times. These assumptions limit the usefulness of most seasonality studies to times of relatively modern climatic conditions during the Holocene. Even in the wake of considerable recent climate change, however, isotopic analyses of a number of species have shown that retrodiction of the season of death for archaeological remains is possible using a variety of methods, as long as valid modern comparative collections are employed. I will discuss this in further detail in

Chapter 5.

3 Table 1.1. Selected Influential Aquatic Faunal Seasonality Studies

Year Fauna Element Method Location Source 1969 Pismo clam Valve Increment California Weide 1969 1970 Chione clam, Valve Increment New Zealand Coutts 1970 Protothaca clam 1971 Chione clam, Valve Increment New Zealand Coutts and Protothaca clam Higham 1971 1973 Limpet Valve Isotopes Africa Shackleton 1973 1979 Quahog clam Valve Increment Georgia (US) Clark 1979 1981 Rangia clam Valve Increment Texas Aten 1981 1981 California mussel Valve Isotopes California Killingley 1981 1983 Limpet, Top Shell Valves Isotopes Spain Bailey et al. 1983 1983 Clams Valve Increment Britain Deith 1983 1985 Quahog clam Valve Increment Georgia (US) Quitmyer et al. 1985 1986 Fish (multiple) Otolith Increment, size‑ North Simons 1986 class America 1986 Clams (multiple) Valve Increment Southeast US Claassen 1986 1988 Clams (multiple) Valve Isotopes Franchthi, Deith 1988 Greece 1989 Fish (multiple) Otolith, Scales Increment North Wheeler and Jones America 1989 1991 Quahog clam Valve Increment Florida Russo 1991 1991 Fish (multiple) Multiple Size‑class Florida Russo 1991 1991 Oyster via Boonea Gastropod Size‑class Florida Russo 1991 impressa 1996 Quahog clam Valve Isotopes, Southeast US Jones and Increment Quitmyer 1996 1997 Quahog clam Valve Isotopes, Southeast US Quitmyer et al. Increment 1997 1998 Oyster Valve Increment Maryland Herbert and Steponaitis 1998 2000 Cod Otolith Isotopes New Zealand Higham and Horn 2000 2005 Coquina clam Valve Isotopes Florida Quitmyer et al. 2005 2008 Quahog clam Valve Isotopes, Georgia OʹBrien and Increment Mayer 2008

4 When examining site seasonality, one cannot use negative evidence to infer seasons of site use. In other words, the absence of a seasonal indicator cannot be used to infer the absence of people at a site (Monks 1981; Reitz 2008; Quitmyer et al. 1985a; Russo 1998). Cultural factors such as seasonal food preferences or resource scheduling may make available food items undesirable or less desirable than other available foods for portions of the year. A group may choose to focus on other, more preferable or more readily available food resources that are less visible in the archaeological record or that cannot be evaluated to determine their season of death. Many subsistence strategies include foods without identifiable seasonal markers; for example, many mammalian species and non‑migratory birds do not provide any indication of their season of death. For these reasons one can identify the minimum seasons of human occupation at a site, but one can rarely identify seasons of human absence from the site.

Seasonality studies rely on the death of the animal to indicate when people occupied a site. One must use caution to ensure that commensal species (animals that were introduced to the deposit by non‑human means) are not included in seasonality assessments. This is relatively easy with aquatic species such as large‑bodied fish and shellfish, as it is somewhat unlikely that they would be introduced into a midden without human intervention (but see

Erlandson and Moss 2001; Wing and Quitmyer 1992). By using species that are common throughout the assemblage and that were probable food items, one can be more confident in the application of seasonality results.

Seasonality studies are very sensitive to sample sizes (Monks 1981). Because there is variability in the onset of seasonal indicators within species, large samples are necessary to make correct determinations for seasons of death. Quitmyer et al. (1997:837) argued that samples greater than 30 individuals are generally agreed to be acceptable. It is therefore imperative that seasonality studies detail the samples size and types of seasonal indicators used

(Monks 1981:183).

Seasonality determinations of different animals (or different parts of the same animal) require distinct methodologies based on their growth characteristics. In some cases, more than one methodology is appropriate for an animal; typically, the least expensive technique is the one most widely applied. Scholars have divided seasonality methodologies into a variety of

5 classifications (e.g., Aten 1981; Monks 1981; Russo 1998). Aten (1981) and Monks (1981) use very similar classifications despite differing terminology. Table 1.2 provides a summary of their classifications. Russo (1998) characterizes seasonality studies in a similar fashion as Aten and

Monks, but he describes only three types of studies and ranks them according to reliability

(Table 1.3). This dissertation primarily utilizes the incremental growth technique, which Russo

(1998:149) identifies as “potentially the most rigorous measures of seasonality.” In Chapter 4, I provide a discussion of specific seasonality studies at various shell rings, and I provide a detailed methodology of incremental growth studies for quahog clams in Chapters 5 and 6.

Table 1.2. Classification of Methodologies to Determine Faunal Seasonality

Aten (1981) Monks (1981) Examples Presence/Absence Presence/Absence Presence of migratory waterfowl (Aten) Presence of migratory cod (Monks) Demography Population Structures Size‑class ‑ Allometric scaling of fish (Aten) Age and sex composition, size class, size of remains (Monks) Morphology Physiological Events Changes in shell contour through annual cycle – Rangia cuneata (Aten) Epiphyseal fusion, tooth eruption, antler growth, medullary bone in mammals and birds (Monks) Structure Incremental Structures Changes in shell microstructure correlated with seasonal changes ‑ Growth increments of Mercenaria spp. (Aten) Self‑contained additions to previous growth ‑ annuli in fish otoliths, molluscan annuli (Monks) Chemistry Oxygen Isotope Analysis Changes in shell composition ‑ ratios of oxygen and carbon isotopes in Mercenaria spp. (Aten) Studies the isotopic composition of oxygen ‑ carbonate samples from molluscan shells (Monks)

6 Table 1.3. Russo’s (1998) Classification of Methodologies to Determine Faunal Seasonality

Method Example Reliability Presence/Absence Presence or absence of migratory fauna Poor Modal Size Classes Measured through population structures ‑ fish and Good shellfish Incremental Growth Growth increments that form annually ‑ in annuli of Best Structures bivalves, fish scales, otoliths, etc.

Summary of Research

This dissertation focuses on the Guana shell ring (8SJ2554), a Late Archaic site just north of St.

Augustine, Florida. Shell rings, as the name implies, are circular piles of refuse comprised primarily of shell. These rings are found along the estuaries and coasts of the Southeast from

South Carolina to Mississippi, although they are concentrated in South Carolina, Georgia, and

Florida. The Guana shell ring is located in the Georgia Bight, an embayment along the Atlantic coast from Cape Fear, North Carolina to Cape Canaveral, Florida (Hubbard et al. 1979; Reitz et al. 2009). The Georgia Bight is characterized by generally similar salt marsh and estuarine habitats along its length, which are home to a similar suite of resources. This dissertation focuses primarily on archaeological sites located in the portion of the Georgia Bight that extends from St. Catherines Island, Georgia to St. Augustine, Florida.

Archaeologists have identified evidence for sedentary or at least semi‑sedentary occupation of several shell rings and ring complexes in the Southeast (e.g., Colaninno 2010;

Marrinan 1975; Thompson and Andrus 2011; Quitmyer and Jones 2012; Russo 1991). I hypothesize that Late Archaic people occupied most shell rings on a sedentary basis.

Specifically, I hypothesize that the Guana shell ring is the product of a sedentary occupation. I propose that this level of sedentism was made possible by exploiting estuarine resources that are available throughout the year and could be caught with mass‑capture techniques. It is likely that at the end of the Middle Archaic period Southeastern estuaries stabilized after substantial sea level rise and marsh transgression. Following sea level stabilization circa 5300‑6000 B.P., estuaries stabilized or formed for the first time, resulting in a resource base that was both

7 plentiful and predictable (Miller 1998; Rollins and Thomas 2011; Yesner 1984). Shellfish and schooling fish flourished in the newly formed or newly stabilized estuaries and lagoons. These resources could be gathered in large quantities by net fishing or collecting from shellfish beds throughout the year. These changes provided Late Archaic inhabitants with the opportunity to remain in a single location for multiple seasons or throughout the entire year.

To evaluate the degree of sedentism at the Guana shell ring, I examine seasons of quahog clam exploitation using the incremental growth technique. The name quahog refers to clams in the genus Mercenaria. The genus is comprised of M. mercenaria, M. campechiensis, and their hybrids. Incremental growth analysis evaluates incremental growth structures that are visible in cross‑section of the shell, and has proven to be an effective technique to determine the season of death for each species of Mercenaria and their hybrids (Arnold et al. 1998). Regional differences in the temporal onset of seasonal growth structures make it necessary to compare archaeological clams to modern clams collected from the same area (Jones and Quitmyer 1996).

To identify the seasons of clam exploitation at Guana, I gathered a modern comparative clam collection from St. Augustine, Florida. Although two modern collections bracket the St.

Augustine area, these collections are rather distant and the ecosystems of these locations are quite different (Figure 1.1). St. Augustine lies in a transitional zone between the salt marshes and tide‑dominated coasts of Georgia and the mangrove swamps and wave‑action coasts of

Central Florida. I chose to collect a modern sample in St. Augustine to 1) ensure the accuracy of the comparison between modern and archaeological clams, 2) to examine the variability among three collections at gradually increasing latitudes, and 3) to evaluate an acceptable distance between the modern collection and the archaeological site under investigation. To assemble the

St. Augustine collection I gathered living clams from the Matanzas River during every month in

2010. The St. Augustine collection provides not only a baseline for identifying seasonal profiles of clam growth with which to evaluate archaeological clam seasonality, but also provides important environmental data for the area and relevant biological information about quahog clams. Lastly, the St. Augustine collection permits archaeologists to conduct seasonality studies at the numerous prehistoric and historic shell bearing sites in St. Augustine.

Figure 1.1. Map of the Guana Shell Ring and Selected Modern Comparative Clam Collections in the Georgia Bight. This map shows the location of the Guana Shell Ring, the Kings Bay collection (Quitmyer et al. 1985a and b; Quitmyer et al. 1997), the St. Augustine collection (this dissertation), and the Indian River collection (Quitmyer 1995; Quitmyer et al. 1997).

9 CHAPTER 2

THEORETICAL BACKGROUND

Conceptual Framework of Sedentism

To discuss the expected mobility patterns of Late Archaic foragers, I utilize Murdock’s (1967) simple, but useful divisions of mobility. I consider a group sedentary if they occupy a single location for four seasons throughout the year. I consider a group semi‑sedentary if they occupy a single location for two or more consecutive seasons of the year (i.e., they live at a site for much of the year but depart for the remainder of the year). Cultures that occupy a single location for less than two seasons of the year are considered mobile. Several anthropologists have pointed out the distinction between group mobility and individual or family‑based mobility (Bar‑Yosef and Rocek 1998; Kelly 1998; Rafferty 1985). The nature of the Guana shell ring deposit and the excavation methodology employed make individual or family‑based mobility difficult to observe. Instead, I focus on group mobility, not only because it is easier to detect archaeologically but because of its relationship with cultural restrictions. I believe that group mobility is primarily governed by cultural values, as it is a group activity that must be agreed upon by its members. Group mobility is the primary focus of this dissertation because I want to examine cultural and environmental factors that led to the development and maintenance of sedentism.

An important factor in becoming sedentary is population change, which appears to be both a cause and effect of becoming sedentary. Kelly (1998:19) has proposed that particularly in areas where daily foraging time is minimal, population will grow from increased reproduction and decreased infant mortality. Once population grows sufficiently, if the group does not fission, movement would become limited as the cost of moving increases. It is also likely that in a sedentary lifestyle with abundant available resources, population would continue to grow

10 (Kelly 1998). This is especially relevant in estuarine/marine environments, where resource diversity and abundance are high and resources are typically available throughout the year

(Bailey and Parkington 1988; MacMahon and Marquardt 2004; Quitmyer 1985; Yesner 1980).

Plentiful estuarine and coastal resources may have led to increased population, resulting in higher costs associated with mobility and a need to remain in a single location for a longer duration. Reliable and predictable resources in the coastal zone would have provided the opportunity for groups to become more sedentary. Binford (2006) found that ethnographic groups with an aquatic‑based subsistence often demonstrate reduced mobility and higher population densities than groups with terrestrial‑based subsistence. This suggests that the nature of estuarine and marine environments may encourage the development of sedentism if cultural preferences permit it.

Because the transition to sedentism requires a reorganization of social structures and the way people live in their natural and cultural environments, the study of sedentism addresses a number of anthropological interests. While some have viewed sedentism as the evolutionary goal of hunter‑gatherers, sedentism includes more than a simple cessation of movement. One of the most obvious results of becoming sedentary involves a reorganization of both individual and group foraging to ensure that resources are not depleted (Kelly 1998). Sedentary groups must devise methods of preventing resource depletion, which may involve broadening the resource base (Reitz and Wing 1999:253) and/or developing a schedule for gathering particular resources. A number of coastal sites have demonstrated seasonal resource scheduling, despite the year‑round availability of the resource (e.g., Claassen 1986; Parsons 2008; Quitmyer et al.

1985a and b; Quitmyer et al. 1997; Russo et al. 1993). Resource scheduling can occur for reasons other than resource conservation, including cultural valuations of seasonal suitability, taste preferences, or to maximize returns from prey due to seasonal fluctuations of biomass (Monks

1981; Parsons 2008).

Becoming sedentary also significantly affects divisions of labor, both across gender and age. Kelly (1992) argued that the foraging activities of women should highly influence when groups move. Women often gather resources that have lower returns than large game; therefore, their foraging distance is shorter than that of a hunter (Kelly 1998). Women’s

11 resources have the potential to be exhausted more quickly due to a shorter foraging radius.

Estuarine and aquatic settings, however, offer a number of different resources that women and children can gather. Water transport in canoes certainly extended their access to estuarine resources and allowed people to bring larger quantities of unprocessed food to the residential base (Thomas 2008a). Estuarine and coastal resource gathering typically requires low energy expenditure and non‑complex technology (e.g., net fishing, rakes, and hands). These resources are found in predictable locations and can be gathered by people of all ages. During the Late

Archaic period, women and children could have procured shellfish with simple gathering techniques. Small‑scale harvesting of shellfish beds can actually increase bed productivity by thinning out the population and hence providing less competition for nutrients among remaining shellfish (Crook 1992:494‑495; Thomas 2008a:96‑98). If the beds were not over‑ harvested, women and children could collect a variety of shellfish resources from the same locations with little diminishment over a long period of time (Yesner 1980), reducing the need to move the group in order to find food. Furthermore, Binford (1980) argued that in areas where resources were unevenly dispersed (or patchy), a residential move would not solve the problem of food shortage. In an estuarine setting, for example, many food sources are found in patches and concentrated within the estuary, but food is significantly sparser on the surrounding land.

A seasonal flow of fish, birds, and other animals move in and out of the estuary, resulting in a variety of species that are available at different times of the year. Due to this seasonal rotation of animals, seasonal food shortage (not caused by natural disaster or significant human impacts) would typically be brief in nature, and moving the group might not improve the situation. This setting would obviate the usefulness of mobility; in cases of seasonal food shortage moving to a new location would not necessarily provide more food.

Optimal Foraging Theory

Optimal foraging theory came to anthropology from biology and human behavioral ecology, and has been applied in anthropological study since the early 1980s (Bettinger 1980, 1991;

Thomas 2008b). Anthropologists have operationalized optimal foraging theory (OFT) as a

12 model to evaluate choices about food procurement. OFT assumes that people make calculated decisions about what they eat, where and how they procure food, how long to spend foraging, and where they live in order to maximize the amount of energy gained from their subsistence regime (Bettinger 1991). In other words, humans make rational decisions about food based on the energy expended and the energy gained. The cost/benefit ratio is analyzed by identifying available prey and examining characteristics such as density, distribution, mobility, behavior, fecundity, and size (Waselkov 1987:118). OFT emphasizes the relationship between the environment and human behavior: behavior is assumed to be adaptive and ultimately related to genetic fitness (Jochim 1998:13). OFT assumes that by making effective foraging decisions, a forager will either provide more food or decrease foraging time (Thomas 2008b:63). A decrease in time spent foraging allows more time for other activities, such as childcare and reproduction, craft production, exchange, warfare, mate procurement, and so forth. Because hunter‑gatherers have extensive interaction with the environment, an ecological approach to the foraging behavior of hunter‑gatherers makes sense (Jochim 1998:2).

Early anthropological applications of OFT were often highly specific and focused on caloric intake, ranking of food based on nutritional content, and a direct application of the mathematics of OFT. Early anthropological use of OFT considered larger prey more efficient than smaller prey and focused primarily on large game. These models did not consider the density of small prey or the division of labor: some small prey, such as shellfish, could be gathered by women, children, and older individuals who typically did not attempt to obtain large prey (Erlandson 2001:305). Often, the ranking of foods was based almost exclusively on terrestrial resources, which were larger and offered more biomass than small resources such as shellfish and small fish (Erlandson 2001).

More recently, anthropologists have used OFT in a broader sense, rather than trying to apply the direct mathematics and predictions of human behavioral ecology and biology because of the role that culture plays in food choices and general subsistence patterns. Bettinger (1991) argued that food values or preferences are often justified by cultural terms. For example, many hunter‑gatherers who lived in areas with extensive water‑borne pathogens did not consider shellfish to be a food item, probably because gathering shellfish put them at risk for contracting

13 pathogens (Parsons 2007). Although shellfish could provide important nutrients, risk lowers the overall return, making the resource suboptimal. In this case, cultural values of the unsuitability of shellfish discouraged people from exploiting this resource for a logical reason.

OFT predicts which foods a person should forage, but people do not always behave in optimal ways. People can ignore higher ranked foods or include lower ranked foods based on cultural or personal preference. Cultural factors such as food suitability, food taboos, taste preferences, and value of certain foods often influence the foraging pattern in ways that are not optimal. People do not necessarily know the nutritional value of food or how long it will take to procure and process the food (Bettinger 1991). Recently, anthropologists have recognized the variability between ecosystems and individual foraging abilities (Jochim 1998), and that the aggregation of prey and use of mass capture technologies can make small prey more optimal than large prey (Jochim 1998:16). Bettinger (1991) describes the relationship of energy gained and spent as a generalization that can be used to determine whether foraging is optimal. OFT can be used as a set of testable hypotheses about subsistence behavior in a variety of habitats

(O’Connell and Hawkes 1981:116). Current use of OFT in archaeological research does not precisely apply the mathematics of caloric intake and energy output, but rather uses OFT as a tool to generally understand the subsistence base and settlement patterns (Jochim 1998).

OFT is comprised of several models, including diet breadth, patch choice, the marginal value theorem, and central place foraging. Of these, diet breadth and central place foraging are most commonly employed in hunter‑gatherer research. I briefly describe each of the models below.

The diet breadth model focuses on which foods people forage. Available food items vary in the amount of energy expended and gained by acquiring and processing (Bettinger

1991:84). The diet breadth model focuses exclusively on search, capture, and handling times, and assumes that all resources are randomly distributed (i.e., are not located in patches)

(Thomas 2008b). Hunter‑gatherers must have a diet that is broad enough to provide for all their nutritional needs and that requires a reasonable amount of energy expenditure. The diet breadth model assumes that the costs and benefits of procuring a food item are fixed because the model assumes infinite abundance once the resource is located (Bettinger 1991:85). The

14 model focuses on the combination of foods to create the most optimal diet. Once one encounters a resource, one must decide to exploit it or continue pursuing other food items. As more kinds of food are encountered and added to the diet, diet breadth increases. A decrease in food availability will often result in an increase in diet breadth because lower ranked foods must be added to the diet. Also, increase in population will increase the diet breadth (Bettinger

1991).

The patch choice model recognizes that resources are not infinitely abundant or equally distributed. Patch choice is particularly relevant for subsistence strategies based on aquatic resources, especially shellfish, because they are often located in patches with varying density.

Bettinger (1991:97) describes patch choice as a special case of the diet breadth model; similar to the way that people decide which foods to include in the diet, they also choose where to forage and which patches to exploit. Some patches provide optimal energy gain, whereas other patches do not. A patch can be partially exploited and returned to again for further exploitation, or it can be exploited fully in one foraging trip. Depletion of the patch from increased use causes an immediate decrease in return rates, and if return rates fall below acceptable levels the patch will not be exploited. When population increases, patch selectivity declines (Bettinger 1991). This decline in patch selectivity may result in overexploitation of a patch or resource. It is important to note, however, that not all exploitation negatively impacts the overall or long‑term productivity of the patch. Limited exploitation of shellfish beds can actually improve the rate of growth for shellfish by lowering competition for food (Crook

1992:494‑495; Thomas 2008a:96‑98). For example, removing regulation size oysters from beds

(but leaving small oysters) lessens food competition and allows the smaller oysters to grow rapidly. Within a year, the oysters that were too small to legally harvest will reach harvestable size (Phil Cubbedge, personal communication 2010). Therefore, patches with lowered return rates (resulting from exploitation) may eventually rebound and once again provide high return rates.

The marginal value theorem focuses on how long to stay in a single location while foraging. This is predicted by the quantity and quality of the resources in a location, and the rate at which these resources are depleted (Bettinger 1991). As soon as a forager removes

15 resources from the patch, the rate of return begins to decline; once the rate of return equals that which can be expected by moving from patch to patch, the forager should move on (Bettinger

1991:92). When population increases, the overall time spent in a single patch increases

(Bettinger 1991).

Central place foraging examines a spatial component that involves a departure and return to a single location (Bettinger 1991; Orians and Pearson 1979). This model focuses on a home base (the central place), from which travel time to and from resource locations is an important consideration. Prey choice is based on return rate of prey, time spent handling the prey, and the distance traveled to and from the central place (whereas diet breadth ranks resources based on only energy gained and handling time) (Bettinger 1991:96). As travel time increases, time spent in a patch and acceptable prey size (or abundance) should increase

(Bettinger 1991:94). When traveling time is not costly, handling time is more important

(Bettinger 1991). If the patch is very close, one should consider taking smaller or less abundant items with lower returns; but if the patch is farther, one must take larger or more abundant prey with higher returns (Bettinger 1991). When population increases and foraging return rates decline, the benefit of travel decreases because diet breadth and processing time increase.

One factor that must be taken into account in central place foraging models is technology related to travel, such as watercraft. The finds at Newnans Lake, Florida revealed that sophisticated canoe technology was in place in the Southeastern U.S. by the Late Archaic period (Wheeler et al. 2003) and was very likely present before the Late Archaic. Canoes significantly decrease travel time to aquatic patches as well as the physical labor required to bring resources back to the central place (Thomas 2008a), hence relaxing restrictions on prey size or abundance. Canoes also reduce the need to process resources in the field, allowing for mass‑processing (such as opening and cooking bivalves at the same time) to occur at the central place. The effects of population increase on foraging, such as increased diet breadth, decline in patch selectivity, increased time spent in patches, and decreased benefit of travel (Bettinger

1991) were likely the outcome of increased sedentism as well. Many of these effects can be mitigated to some extent by use of the canoe to reduce overall costs of food items.

16 Mollusks at Archaeological Sites

Mollusk shells are the most common invertebrate remains recovered from archaeological sites

(Bar‑Yosef Mayer 2005:1), and often represent a substantial part of the overall subsistence strategy at coastal sites. The earliest shells in an archaeological context were recovered from a habitation at Terra Amata, France (Claassen 1998), but shell gathering does not become evident in other places until much later. The oldest shell middens (sites in which shell is the primary component of the matrix) are found in coastal South Africa dating from 130,000 to 30,000 years ago (Waselkov 1987:125). Although shellfish were exploited during the Paleolithic by coastal people, it was not until the Mesolithic and Archaic periods that mollusks were exploited on a large scale. In the southeastern U.S., freshwater mollusks accumulated as large shell middens along the St. Johns River beginning at approximately 7300 B.P. (Randall 2008). It is possible that coastal shell middens may have existed at this time but currently, evidence for these early shell middens is lacking. While it is possible that rising sea levels destroyed evidence for early coastal middens, the ability to locate inundated Paleoindian sites off the coast suggests that if these sites existed we should be able to find evidence for them.

The phylum Mollusca includes both bivalves and gastropods, including clams, oysters, scallops, snails, whelks, squids, and octopuses among other animals (Abbot and Morris 1995).

The term shellfish is typically used for bivalves, gastropods, crabs, lobster, shrimp, and other crustaceans (Evans 1969:480). Shellfish were a valuable food resource to many prehistoric cultures, and were one of the most abundant resources available (Quitmyer 1985:28). In comparison to other protein sources, gathering of shellfish requires low energy expenditure and simple technology (such as digging sticks, rakes, or simply the hands). Shellfish can be obtained and cooked in mass quantities, are found in predictable locations, and can be gathered by segments of the population who cannot otherwise procure much protein (such as children and aged people). Shellfish, while low in calories, are high in protein, calcium, iodine, electrolytes, and other minerals (Yesner 1980:733). Shellfish also offer carbohydrates that other animal foods lack, which must typically be acquired through plants. Table 2.1 shows a comparison of the nutritional value of oysters and clams harvested on St. Catherines Island to

17 Table 2.1. Comparison of Calories, Carbohydrates, Fat, and Protein of Oysters, Quahog Clams, and Catfish. This table presents data on the nutritional value of an oyster and clam from St. Catherines Island and catfish estimates from the USDA (after Thomas 2008a:75).

Species Energy kcal/100 g Carbohydrates % Fat % Protein % Quahog 102 4.1 1.2 18.7 Oyster 49‑81 0‑4.6% 0.6‑1.5 6.9‑12.9 Catfish 95 0 2.8 16.4

Table 2.2. Comparison of Yields of Shellfish, Fish, and Deer This table provides a comparison of the yields of shellfish, fish, and deer (compiled from Waselkov 1987:120‑121).

Meat Yield Edible Kcal Protein Fat Carbohydrates Resource (g) Portion (%) (per 100 g) (g per 100 g) (g per 100 g) (g per 100 g) Clam 18‑30 15 68 8.4‑10.7 1.2‑1.8 2.7‑3.4 Oyster 5 15 66 8.4‑12.0 1.8‑2.5 3.4‑6.5 Catfish 500 67 103 17.6 3.1 0 Deer 32,500 58 126‑198 20.0‑35.0 4.0‑6.4 0

USDA nutrition estimates for catfish. Table 2.2 shows a comparison of the nutritional value of clams, oyster, deer, and catfish (from Waselkov 1987). Although the two tables present slightly different values for resources, both demonstrate relatively high caloric value, protein content, and carbohydrates in shellfish when compared to other resources. This variability may be the result of different nutritional values at different times of the year (see Thomas 2008a:82, 90‑91).

Shellfish can withstand higher rates of human predation than most mammalian food sources

(Yesner 1980:729) (although not without consequence), and exploiting shellfish beds occasionally creates more productive beds (Thomas 2008a:96‑98). Shellfish can also be used as protein source when hunting forays fail (Meehan 1982; Waselkov 1987). All of these qualities mean that shellfish can serve as valuable resources capable of sustaining relatively large populations for a long period of time.

Coastal resources played a prominent role in the economic and cultural development of coastal groups in Georgia and Florida. Most coastal areas have high resource biomass and

18 diversity and are extremely productive – primary productivity in coastal zones is approximately 2,000 kcal per square meter (Yesner 1980:728). An estuary can produce four to ten times the amount of organic matter produced by a cornfield of the same size; although humans do not directly consume the organic matter, it provides food and nutrients for species they do eat (MacMahon and Marquardt 2004:8). In fact, the estuary provides food and shelter for more than 90 percent of the marine species consumed today (MacMahon and Marquardt

2004).

In general, coastal regions yield large quantities of resources, which can permit high population densities that otherwise would necessitate reliance on domesticated plants.

Plentiful coastal resources enabled some prehistoric cultures to become highly complex, such as the Calusa of Southwest Florida. The Calusa subsisted primarily on coastal resources and achieved a socially complex chiefdom without reliance on agriculture (Widmer 1988). The need for agriculture in the coastal zone appears to have been minimal – many coastal groups did not adopt agriculture until relatively late in prehistory. This is likely a result of the myriad coastal resources, especially shellfish, which were capable of sustaining large populations over a long duration.

19 CHAPTER 3

ARCHAEOLOGICAL BACKGROUND

Theoretical Framework of the Late Archaic Southeast

Introduction

Since Willey and Phillips’ (1958) use of the term “Archaic” our notions about what characterizes the Archaic period have changed significantly. Willey and Phillips (1958:107) described the

Archaic as “the stage of migratory hunting and gathering cultures continuing into environmental conditions approximately those of the present.” The Archaic period (10,000 –

2500 B.P.) was considered a transitional phase between the nomadic Paleoindians and sedentary agricultural people in the Woodland period (Russo 1998:143). For many early archaeologists, the Archaic was primarily interesting for the changes that took place before and after it – the

Archaic period itself was not especially interesting.

Most archaeologists agreed that Archaic people were egalitarian, non‑complex foragers who did not engage in more advanced societal activities such as mound‑building. Archaic mounds and rings were viewed as uncommon (Gibson and Carr 1998) or were thought to belong to the Woodland period. Despite these views, several archaeologists hypothesized that people living in the Southeastern coastal zone were capable of living a rather sedentary lifestyle

(e.g., Colquhoun and Brooks 1986; DePratter 1975, 1979; Marrinan 1975; Waring 1968a; Waring and Larson 1968). However, the dominant paradigm was one of high mobility, and traditional markers of sedentism, such as storage pits, ceramics, substantial structures, mound construction, large village area, and a dependable subsistence economy were often considered insufficient evidence to alter the overall conception of the Late Archaic period (Russo 1998:144).

Storage pits, ceramics, and mound complexes were originally thought to be characteristics of sedentary Woodland societies (Russo 1998), and were not considered to be part of the Archaic

20 repertoire. However, these characteristics have now been identified in the Late Archaic period, and may be associated with decreased mobility during this period.

The large shell middens in the Southeast were an important catalyst in the realization that Archaic people were capable of living a sedentary lifestyle. Several scholars who worked on shell middens and rings recognized the potential of the coast – abundant shellfish and fish remains suggested that the coast was capable of providing year round sustenance. Two of the first archaeologists to note this potential were Antonio Waring, Jr. and Lewis Larson, Jr. at the

Sapelo Island shell ring (work beginning in 1949). They saw no reason why shellfish could not be gathered throughout the year, despite extant notions of the dangers of eating shellfish during warmer months. Waring (1968a:244) commented, “In these favored locations, however, our hunters and gatherers were in hog‑heaven.” He argued that the extensive shell deposits reflect large populations who had ample leisure time given the large amounts of food available to them in the nearby estuaries.

DePratter (1975) argued that shellfish collecting adds an element of residential stability, resulting in greater midden deposition and more variable artifact assemblages. He proposed that while seasonal scheduling was important, it would not require relocation of the group because plentiful resources were nearby in the estuary. He later proposed that coastal shell middens were the products of relatively sedentary populations dependent primarily on marine resources that were available throughout the year (DePratter 1976). Marrinan (1975) also recognized the potential role of estuarine resources, particularly mollusks, in a sedentary lifestyle. She proposed that the reliance on mollusks influenced the settlement pattern and permitted a more sedentary lifestyle (Marrinan 1975:15). Based on a combination of floral and faunal indicators, Marrinan (1975, 2010) identified deposition throughout most of the year at the shell rings on St. Simons Island.

Throughout the 1970s and 1980s arguments for relatively sedentary Late Archaic societies along the Southeastern coasts continued to appear in the literature. In 1979, DePratter claimed that coastal groups occupied the barrier islands year‑round. Milanich and Fairbanks

(1980:150) described the preceramic Late Archaic in Florida as the time when “sporadic occupation became more sedentary and village life began.” (Milanich and Fairbanks [1980]

21 specifically referred to these characteristics during the Mount Taylor period, which is defined as a preceramic portion of the end of the Middle Archaic and beginning of the Late Archaic periods.) In 1985, Price and Brown (1985a:11‑12) stated that shell middens of the Archaic period provide an example of extensive settlement and are likely associated with sedentism. In 1986,

Colquhoun and Brooks (1986:279) stated that early shell middens along the coast might indicate rather intensive, multi‑season occupation.

Beginning in 1985, seasonality studies of invertebrates became a viable option to examine seasons of resource use, which could then be used to evaluate group mobility. The publication of a comparative clam collection (Quitmyer et al. 1985a, b) permitted assessments of quahog clam seasonality studies at dozens of sites in the Georgia Bight. Publication of comparative clam collections in other locations facilitated quahog clam seasonality at sites throughout Florida and South Carolina as well (e.g., Quitmyer 1995; Quitmyer and Jones 1992;

Quitmyer et al. 1997). These data led to a number of studies suggesting more sedentary lifestyles based on demonstrated multi‑seasonal occupation at Late Archaic sites. Faunal seasonality studies continue to amass further evidence of decreased mobility in the coastal zone during the Late Archaic period.

The New Archaic

An edition of the SAA Archaeological Record (2008, Volume 8, Number 5) entitled “The New

Archaic” demonstrates the substantial theoretical transformations that have taken place in our paradigm of the Archaic period. The presence of mounds and formal cemeteries demonstrate more complex behaviors than originally expected of the Archaic period. Radiocarbon dating of these mounds and burials forced scholars to reevaluate social complexity in the Archaic (Gibson and Carr 1998). Our current paradigm of the Archaic period encompasses considerable variation and has abandoned much of the normative ideology regarding the nature of Archaic societies.

The issue of monument building is a relatively new one to be examined at Late Archaic

(and some Middle Archaic) sites. A number of scholars (e.g., Sassaman 2008; Russo 2004a, 2008;

Saunders 2002, 2004) have identified shell rings as the remains of ceremonial activity and/or

22 monument building by groups engaged in the beginnings of hierarchy. Other archaeologists are more utilitarian in their estimations of site function (e.g., Marrinan 1975; Thompson 2007;

Trinkley 1985). The issue of shell ring function will be addressed further in Chapter 4.

It has become clear that we cannot base our interpretations of the archaeological record solely on hunter‑gatherers in the ethnographic record. It appears that ethnographic hunter‑ gatherers are not adequate analogs to those of the Archaic period of the Southeast U.S. or the

Mesolithic period of Europe (Binford 2006; Greaves 2006; Price 1991). Price (1991:230‑231) argued that most ethnographic evidence on hunter‑gatherers comes from inland, marginal locations where population density is low and mobility is more frequent – clearly a different scenario than that of Archaic period people exploiting abundant coastal resources. Overall, hunter‑gatherers in the ethnographic record do not represent the full range of variability that existed in prehistory. We must therefore draw our evidence for archaeological hunter‑gatherers primarily from the archaeological record, and not allow the ethnographic record of hunter‑ gatherers to limit our interpretations.

Sea Level Rise and Settlement Patterns

The development of coastal occupation in the Georgia Bight is entwined with global sea level rise as well as regional variations in sea level rise, and the development of sustainable populations of shellfish. It is my belief that the sedentary or semi‑sedentary coastal settlement pattern did not develop until sea level stabilized sometime around 6000 B.P. At this time, estuarine conditions occurred that favored the sustainable growth of oysters, clams, and other shellfish that permitted a more sedentary lifestyle.

The coastline of the Georgia Bight is characterized by a broad, shallow continental shelf.

This translates into rapidly moving shorelines during times of changing sea levels. At its lowest level during the Wisconsinan glaciation (circa 17,000 B.P.), sea level is estimated to be 90 m below its present level (Donoghue and White 1995:651), corresponding to a coastline approximately 100 km seaward of its current location (Miller 1998:43). Two distinct periods of world‑wide sea level change since the Pleistocene have been identified: 1) from 18,000‑6000 B.P.

23 when glacial meltwater caused significant sea level rise throughout the world, and 2) from 6000

B.P. to more recent times, when sea level changes were dominated by regional and local processes, including subsidence, uplift, erosion, release of sediments, and reclamation

(Jelgersma and Tooley 1995). In North America, Fairbanks (1989) identified two major sea level rises: one occurred around 12,000 B.P. when sea level rose approximately 24 mm per year and the other was around 9500 B.P. when sea level rose approximately 28 mm per year. Both events resulted in exceedingly rapid sea level rise with rapid sea level transgression.

Two terms should be defined when discussing sea level changes. The first is eustatic sea level change, which has a global expression, is related to global climate change, and is a result of changes in ocean water or basin volume (Colquhoun and Brooks 1986:275). The second term is isostatic change, which occurs in relation to regional or local geophysical characteristics

(Colquhoun and Brooks 1986:275), such as post‑glacial land rebound (once the weight of glaciers is removed, the Earth’s surface actually rises).

As a result of glacial melting, eustatic sea levels began rising rapidly throughout the world around 12,000 B.P. and lasted until about 6000 B.P., after which time sea level rise was more moderate (Colquhoun and Brooks 1986; Donoghue and White 1995; Miller 1998). The effect of eustatic sea level rise is influenced by regional characteristics of the local topography and bathymetry. For example, some areas of the world, such as northern Europe, experienced isostatic uplift as glaciers retreated (Price 1991), and to some extent this mitigated the effect of rising sea level. The southeastern U.S., however, did not undergo isostatic land uplift (as glaciers were absent from the Southeast), and the shallow continental shelf meant that as sea level rose in the Holocene, coastlines retreated landward (coastline transgression) at a rapid pace throughout the Early and Middle Archaic.

Anderson (1996) examined settlement patterns throughout the Archaic period based on the presence of 32,428 known sites with Archaic components. He found that Early Archaic

(10,000‑7000 B.P.) sites tend to occur along or near major river systems and near major lithic outcroppings. Populations moved over large areas, primarily along river drainages and employed a mixture of logistical and residential mobility (Anderson 1996). Moderate occupation of the interior of the coastal plains is evident over much of the region, but habitation

24 along the modern coast is evident in a few areas Anderson 1996:162). In Florida, Early Archaic components are rare compared with the rest of the Southeast, and tend to occur in the same areas as Paleoindian sites (Anderson 1996).

Middle Archaic (7000‑5000 B.P.) sites suggest seasonally mobile groups concentrated along river drainages. The coastal plain is characterized by low site incidence, lower even than the Early Archaic (Anderson 1996). Anderson posits that environmental changes, particularly changes in vegetation, may have prompted groups living near the coast to move farther inland at this time. Site concentrations east of the Mississippi River are present in areas where early mound complexes are known, while concentrations west of the Mississippi River may be the products of highly mobile groups (Anderson 1996:165).

During the Middle Archaic period freshwater shell middens along the St. Johns River appeared and proliferated. At this time, there was a dramatic increase in the exploitation of riverine aquatic fauna, including mollusks (Smith 1986:22). At some point in the Middle

Archaic, the water table in Florida rose significantly (as a result of rising sea levels) (Miller

1992). This resulted in dramatic changes in the St. Johns River drainage. It became more similar to a freshwater estuary or lake than a river (Miller 1998) because of rising sea levels, which moved the mouth of the river and changed the river gradient to a more gradual drop (Miller

1998). These changes created a tidal influence throughout the river; elevated water tables changed the flow of the river and made it a more broad, shallow flowing expanse of water, rather than a deeply cut river channel (Miller 1992). This environment favored the development of large numbers of freshwater mollusks, particularly small gastropods. These freshwater shellfish comprise the bulk of the shell middens along the St. Johns River, which were so famously described by Jeffries Wyman (1875) in the nineteenth century. Randall (2008) documents the emergence of freshwater shellfishing along the St. Johns River as early as 7300

B.P.; however, it seems that it was primarily toward the end of the Middle Archaic that the area was substantially occupied. Radiocarbon dates from Tick Island show that between 6000 and

4000 B.P. Archaic people began to occupy the St. Johns Basin on a regular basis (Aten 1999;

Miller 1998).

25 Although freshwater shellfish exploitation is evident in the Middle Archaic period, there is very little evidence of coastal and estuarine exploitation at this time. Two notable exceptions are Spencer’s Midden in northeast Florida and the Horr’s Island site at mouth of the

Caloosahatchee River. Spencer’s Midden (8DU5626) may represent the earliest marine shell midden along the Atlantic coast, dating from 5700‑5500 B.P. (Russo and Saunders 1999). The site demonstrated continued occupation from the Middle to Late Archaic periods (Russo and

Saunders 1999). The Horr’s Island site contains a number of early marine shell middens as well as a shell ring. The site is in a distinct geological setting: sea level was at or near the present mean sea level as early as 7300‑7150 cal B.P. (Walker et al. 1995). Horr’s Island was abandoned during the mid‑Holocene high sea level stand and was reoccupied afterwards (Russo 1991;

Walker et al. 1995). These sites indicate that at least some estuarine resources were available towards the end of the Middle Archaic period. However, it is not until after 5000 B.P. that coastal and estuarine adaptations become widespread and estuarine shell middens become common in northeast Florida.

With the exceptions of Spencer’s midden and Horr’s Island, the earliest unequivocal estuarine adaptations occurred during the Late Archaic period (Russo 1991; Russo and

Saunders 1999; Reitz 2008). Isotopic analyses of human diets at the Windover site (circa 7410

B.P. [uncorrected]) near Titusville, Florida support this argument. It appears that during the

Early and Middle Archaic periods, subsistence was based on terrestrial and freshwater aquatic resources, excluding estuarine shellfish (Doran 2007:38; Tuross et al. 1994). Doran (2007:38) argued that after 6000 B.P., estuarine shellfish and other aquatic resources along the coasts became important in the diet, and virtually all people in estuarine/riverine environments took advantage of these resources.

Global sea levels rose rapidly throughout the world from 12,000 to 6000 B.P. (Colquhoun and Brooks 1986; Donoghue and White 1995; Miller 1998). However, the shoreline expression of this sea level rise was geographically variable as a result of local geography. The Georgia Bight is characterized by a broad, shallow continental shelf that is quickly inundated in times of sea level rise. During the Early and Middle Archaic periods, sea level rose considerably, which created unstable coastlines and coastal transgression that occurred until circa 6000 B.P. The

26 continually changing configuration between land and sea may have negatively impacted estuarine resources, making them less available or unstable. This dearth and/or instability of estuarine resources may have made them a less desirable target for exploitation through much of the Middle Archaic period.

Shellfish are particularly sensitive to changes in salinity that result from changes in sea level. If estuaries and bays were slowly drowned, populations of estuarine fish were probably affected by changes in shelter, breeding grounds, and food supply. The retreating shoreline may have been too unstable to allow the continued existence or development of large, sustainable populations of shellfish and fish along the coasts during the Early and Middle

Archaic periods (Miller 1998:64, emphasis added). Salt marshes, which are known to be extremely productive, would not survive in these conditions. If salt marshes formed, they would likely be destroyed or damaged by continued substantial sea level rise. The dominant marsh grass in salt marshes, such as Spartina alterniflora, can tolerate no more than a 1.2 cm sea level rise per year (Morris et al. 2002). Rising sea levels would inundate marshes with seawater, resulting in a sharp decline in productivity. This increased submergence causes reducing conditions, high sulfide concentrations, death of marsh grasses, and eventually the death of the marsh environment (Rollins and Thomas 2009:19). Miller (1998:72) stated that if coastal features

(such as estuaries) formed, they were probably quickly submerged by rising sea levels.

Sea levels along the Georgia Bight stabilized around 6000 B.P., allowing the development of barrier islands, the dune/lagoon system that characterizes northeast Florida, as well as estuaries, marshes, and inlets (Miller 1998; Rollins and Thomas 2011; Thomas 2008c).

Slowly rising sea levels favor marsh progradation, resulting in more shellfish (Rollins and

Thomas 2011:330). Behind the barrier islands, bays and low‑lying areas were flooded, creating vast salt marsh systems of tidal creeks and estuaries (Thomas 2008c). Once rising sea levels had flooded freshwater lagoons and filled in areas behind barrier islands, large shellfish beds developed (Crusoe and DePratter 1976).

At the onset of the Late Archaic, climatic conditions were similar to modern conditions and sea level in northeast Florida was near its current level (Miller 1998:65). When sea levels stabilized and sustainable estuaries, reefs, and shellfish beds formed, the coastal zone became a

27 productive area capable of supporting a continued occupation. Late Archaic (5000‑2500 B.P.) sites occur widely throughout the Southeast and moderate to extensive use of almost every area is indicated (Anderson 1996:165). Anderson argued that this is a result of much higher population levels and possibly a result of using a broader range of resources. Following a successful adaptation to relatively stable estuarine resources, particularly shellfish, population increased and sites occur more frequently in coastal areas along the Atlantic (Anderson

1996:166; Miller 1992). Florida possesses substantially large numbers of coastal sites during the

Late Archaic, with a concentration in northeastern Florida along the St. Johns River, as well as locations in the Panhandle.

It is possible that we are missing much of the evidence for earlier coastal adaptations because rising sea levels drowned early coastal sites. These sites would now be underwater or beneath marsh sediments. Given the success of identifying submerged Paleoindian sites off the

Gulf Coast (e.g., Faught 2004), however, we should be able to locate and evaluate some of these early coastal sites. Faught (2004) identified a possible Middle Archaic midden at the submerged

J&J Hunt site (8JE740) in the Gulf of Mexico. Early shell midden sites such as J&J Hunt, Horr’s

Island, and Spencer’s Midden, however, remain anomalous. It is possible that these locales represent atypical coastal conditions and adaptations, but perhaps more of these sites will be discovered along Florida’s coast. Based on the evidence at hand, it appears that large‑scale coastal adaptations may have been discouraged during the Early and Middle Archaic periods because of continually rising sea levels and inconstant estuarine habitats. Current evidence suggests that large‑scale exploitation of coastal and estuarine resources may have been a predominantly Late Archaic adaptation. Anderson (1996:157) noted a dramatic increase in the number of sites beginning in the Late Archaic period, which he argued indicates stable population levels in the Early through Middle Archaic that remained rather uniform until the

Late Archaic, when population grew markedly. Doran’s (2007) synthesis of burial sites also indicated a rise in population. He found that burial sites containing more than 50 individuals increased dramatically during the Late Archaic (Doran 2007:43), which may indicate population increase as well as a more sedentary lifestyle (although he does note an overall decrease in the number of burial sites).

28 A productive coast and, especially, a productive estuary offer a variety of resources packed into a localized area. Because most estuarine faunas are available on the coast year‑ round, Late Archaic inhabitants would have had the ability to remain in a single location throughout the year. The importance of stable estuarine resources for a sedentary lifestyle is demonstrated by archaeological evidence on the Southeastern coast during the Early Woodland period. Sometime around the end of the Late Archaic period, it appears that sea level retreated several meters, resulting in loss of estuarine environment and resources (DePratter 1977;

DePratter and Howard 1980; Thomas 2011; Thompson and Turck 2009). Thompson and Turck

(2009) found that Early Woodland period subsistence and settlement patterns in coastal Georgia were significantly impacted by these changes – namely that population decreased and mobility increased. Early Woodland people moved away from sites inhabited during the Late Archaic period and either moved their settlements to be closer to lowered marshes and coasts (that would now be buried by marsh accretion from later sea level rise) (Thomas 2011), or moved elsewhere to exploit riverine and terrestrial resources (Thompson and Turck 2009). During the

Middle and Late Woodland periods sea levels were more similar to the Late Archaic period; people moved back on to Late Archaic sites, where they followed subsistence practices similar to their Archaic ancestors.

The productivity of estuaries and the coast would have provided Late Archaic inhabitants with the opportunity to remain in one location throughout the year. A number of archaeologists have suggested sedentary lifestyles would have been possible during the Late

Archaic as a result of highly productive estuarine and marine resources (e.g., Andrus and

Crowe 2008:518; DePratter 1975; Marrinan 1975; Russo and Quitmyer 1996:227; Thompson and

Turck 2009). With the development of plentiful coastal and estuarine resources, subsistence pressure would likely lessen, especially with the ability to use mass capture techniques that provided large amounts of food relatively quickly. Lowered subsistence stress may have led to an increase in population, making mobility a less desirable, more costly alternative.

29 The European Mesolithic

The trajectory of stabilizing sea levels, extensive occupation of the coast, increasing reliance on estuarine and marine resources, and developing sedentism is not unique to the Archaic

Southeast. Similar climatic and cultural changes took place in Europe throughout the

Mesolithic (12,000‑5000 B.P.). Although the Mesolithic (12,000‑5000 B.P.) and the Archaic periods (10,000‑2500 B.P.) are not entirely contemporaneous, there is considerable merit in comparing the two. Bogucki (1999:138) argued that the two periods are analogous because they are both characterized by hunter‑gatherers adapting to and exploiting essentially modern ecosystems, including estuaries and coastal zones. Both the Mesolithic and the Archaic people utilized a broader variety of resources than is evident in previous periods (Bogucki 1999) as a result of the changes in flora and fauna in the post‑glacial environment. Specifically, marine fauna become an important dietary contribution in both periods. The Mesolithic and Archaic are similar in that they both follow drastic changes in environment and available resources after the Pleistocene, and both periods predate the agricultural proliferation of the following periods.

Both the Mesolithic and the Archaic periods saw dramatic increases in the number of coastal sites, as well as the appearance and proliferation of shell middens following the development of productive estuaries (Anderson 1996; Anderson 2007; Bjerck 2007; Colquhoun and Brooks 1986;

Jochim 2002; Miller 1992; Milner et al. 2007; Thomas 2011; van der Schriek et al. 2007).

Increased estuarine and marine exploitation culminated in sedentary societies in the latter portions of both periods. A brief summary of these changes in the European Mesolithic follows.

Although the planet as a whole was affected by eustatic sea level rise, considerable regional variability in shoreline changes was caused by local characteristics of bathymetry and tectonic actions. Worldwide, sea levels rose rapidly – as much as 100 m in some places from lowered levels during glaciation (Jochim 2002). In Europe, the interface of the land and sea was influenced by both eustatic sea level change as well as isostatic land rebound following glacial retreat. Portions of Europe’s landmass lifted significantly enough to allow for the preservation of a number Early Mesolithic sites above modern sea levels (Bjerck 2007).

30 Similar to the Archaic period, Early Mesolithic (10,000‑8000 B.P.) subsistence was based primarily on terrestrial hunting and gathering, with limited use of marine resources (Jochim

2002:122‑123). Prior to the Middle Holocene, high rates of relative sea level rise caused rapid drowning of any developing estuarine environments, which would have discouraged long‑term settlement in these areas (van der Schriek et al. 2007:177). During the Early Mesolithic, sites were small and were likely the product of small, highly mobile groups (Bjerck 2007:19; Jochim

2002). The end of this period is marked by an increase in the use of marine resources, including shellfish in some areas.

Following the end of the Pleistocene, the oceans were full of food, including mollusks, crustaceans, fish, and sea mammals (Price 1991). Water temperatures in some locations might have limited the variety of shellfish, but their range certainly expanded throughout this period

(Price 1991:215). It appears that estuarine conditions developed at substantially different times in locations throughout Europe, some of which are considerably earlier than estuaries in the

Southeastern U.S. Shellfish use throughout the Mesolithic is regionally variable: some locales show high reliance on shellfish while others show a focus on fish or sea mammals.

Detailed environmental reconstructions in Portugal show that the occurrence of shell middens is directly related to the existence of productive estuaries (van der Schriek et al. 2007).

Shell middens appear after eustatic sea level rise slowed at roughly 6000‑7000 B.P.) (Milner et al.

2007). Denmark has the highest concentration of shell middens (kokkenmoddinger, or kitchen middens) in Northern Europe – the earliest shell middens date to 7600 B.P., however the majority date to circa 6600‑6400 cal B.P. (Anderson 2007). The earliest shell middens in

Denmark correspond with rising temperature and salinity of ocean waters, as well as the presence of oysters in estuarine environments (Anderson 2007:43).

In Portugal, estuaries developed earlier than in Denmark, and consequently shell middens have a deeper history in Portugal. At approximately 8200 cal B.P. sediment output of the Tagus River was substantial enough to mitigate sea level rise, thus creating a productive estuary (van der Schriek et al. 2007). The abrupt development of estuaries corresponds with the appearance of shell middens in the area; it appears that site choice was based on proximity to tidal flats (van der Schriek et al. 2007). Interestingly, these sites were abandoned at the time

31 when the estuary began to contract and die off around 6800 B.P. (van der Schriek et al.

2007:179).

Throughout the Mesolithic period, there is a broadening of the subsistence base to include newly established resources as well as an increase in sedentism and cultural complexity. There is also a great increase in the number of coastal settlements, especially in

Northern Europe (Jochim 2002). In the Late Mesolithic we see evidence for increased sedentism, including organized settlement patterns, larger, more permanent houses, large sites located among rich and stable resources, increasing site size, larger deposits, and increasing regional variations in material culture (Bjerck 2007:17). Shellfish in particular became an important food source for many coastal Late Mesolithic groups, especially in Northern Europe where aquatic foods provided the majority of the subsistence base (Jochim 2002:133, 134). In several coastal areas, residence is “virtually sedentary” and social distinctions become apparent, hinting at increased social complexity (Jochim 2002:141).

Although sea level rise has likely inundated coastal sites in both Europe and North

America, there is little evidence to suggest an estuarine shellfish‑based adaptation prior to the

Late Archaic and the Late Mesolithic. Underwater investigations of Danish coasts have identified more than 3,000 sites, but none can be described as shell mounds and few have substantial quantities of shell (Milner et al.2007:2). Shell middens along the shallow continental slope of the southeastern U.S. may be easier to locate using technology such as side‑scan sonar; such research in the future may be able to identify estuarine adaptations prior to the Late

Archaic period. Marsh sedimentation may be a serious impediment to finding early coastal sites as well, as there is documented marsh accretion over several coastal sites (e.g. DePratter

1977; Marrinan 1975; Thomas 2011). While it is possible that we are missing the evidence for early coastal adaptation, it seems likely that during times of continually rising seas, rapidly changing and disappearing estuaries did not produce large amounts of shellfish or fish; therefore, the ability to exploit these resources would have been minimal.

Both the Mesolithic and Archaic periods were considered uninteresting by archaeologists for some time. Bogucki (1999:128) stated that neither period seemed to have big questions associated with them, and that many people considered study of the Mesolithic and

32 Archaic periods to be “an arcane, acquired taste, best left to the second‑rate students of local prehistory.” In the 1960s, foragers became interesting in their own right (Bogucki 1999),

(probably largely the result of the “Man the Hunter” symposium). In the decades that followed it became increasingly clear that forager adaptations in the Mesolithic and Archaic periods paved the way for cultural developments that occurred in later periods (Bogucki 1999:128).

Hence, hunter‑gatherer adaptations in the Mesolithic and Archaic periods became subjects of considerable archaeological importance.

In sum, the Mesolithic of Europe, particularly in Northern Europe looks rather similar to the Archaic Period in North America, albeit on a slightly different time frame. This may be explained by differences in local environmental circumstances, namely different timing in the changes along the coasts and the development of estuaries. The coastal geology and bathymetry of Northern Europe is quite different from that of the Southeastern U.S., and isostatic uplift occurs in Northern Europe but is absent in the Southeast. These factors mean that the interface of sea and shoreline in Northern Europe would have had a very different history than that of the Southeastern U.S.

Based on the current evidence, coastal sites and marine adaptations increased throughout the Late Archaic in the Southeastern U.S. and the Late Mesolithic in Europe, culminating in considerably increased sedentism during both of these periods. There are other similarities between these two periods as well, including the demonstrated presence of canoes, increased fishing technologies compared to those of the previous periods, initial production of ceramics, and a delayed adoption of agriculture in coastal areas (in some places agriculture was not adopted for centuries, despite its presence in nearby locations) (Price 1991; Wheeler et al.

2003; Widmer 1988). By examining similar developments during the Archaic Southeast and the

European Mesolithic, we may develop stronger anthropological insight into human responses to climate and resource changes.

The Guana Shell Ring

Location and Physical Characteristics

The Guana shell ring is located in Ponte Vedra Beach just north of St. Augustine, Florida.

Surrounding the Guana shell ring are several types of ecosystems that would have provided an extensive variety of potential resources for exploitation. These include brackish marshes and rivers, coastal zones, freshwater ponds, rivers, and marshes, and maritime forest. The Guana site is currently in close proximity to several bodies of water, the Tolomato River and marsh about a kilometer to the west, the Guana Lake roughly 200 meters to the east, and the Atlantic

Ocean, which is presently 1 km to the east. A small pond is present immediately west of the site. Saunders and Rolland (2006) found accumulations of calcium carbonate in the lower levels of excavation near the western arm of the ring, suggesting that at some point water levels inundated the base of the ring. It is likely that the area immediately east of the site was a freshwater marsh during its occupation in prehistory (Russo et al. 2002:5).

The Guana Lake is the result of a dam located near the mouth of a former tidal creek

(Russo et al. 2002:5). The lake flows into the Guana River and then into the Tolomato River approximately 8 km south of the site. During the Late Archaic the section of the river closest to the shell ring would have been a free‑flowing brackish river (Russo et al. 2002:5), providing a variety of aquatic resources. Prior to dam construction, the Guana Lake area was part of a long marsh/estuary system separated from the Atlantic by a barrier dune (Saunders and Rolland

2006:10). It is likely that freshwater ponds also abounded in the immediately area (Russo et al.

2002) providing freshwater aquatic resources. The shell ring is located on a relict dune formation slightly less than 180 m wide (Saunders and Rolland 2006:9). The majority of the ridge lies at approximately 10 m above mean sea level (AMSL). Saunders and Rolland (2006:9) described the soils as Tavares fine sand that is moderately well drained.

Upon first glance, the ring shape is difficult to discern because only portions of the ring surface rise above the surrounding terrain and sediment has covered much of the shell deposit.

The ring is visible on the surface as discontinuous ridges and occasional shell surrounding a

34 relatively flat interior. Surface ring height is highly variable: it is higher, equal in height, and below the ring interior in different portions of the ring. Much of the west arm of the ring is slightly higher than the ring center, while the east arm of the ring is identifiable primarily due to its height over the adjacent slough (Russo et al. 2002:10). The most visible portion of the ring is the central portion of the ring (to the north), which rises substantially higher than either the east or west arms and reaches a height of more than a meter above the ring interior (Russo et al.

2002:10). While most rings are characterized by a vacant interior, Guana has a small midden within the interior of the ring. This interior midden was identified through shovel testing at coordinates 340N, 430E and 340N, 440E (Saunders and Rolland 2006).

Subsurface probing by Russo and colleagues (2002) provided a more accurate picture of the layout, size, and shape of the ring (Figure 3.1). Probing provided information about subsurface shell thickness and depth, and revealed that the ring was significantly larger than it appeared on the surface. Probing indicated that the ring is approximately 170 meters from north to south and 150 m east to west; areas that appeared to be below the level of the ring interior in fact contained significant amounts of shell (Russo et al. 2002:10). Shell thickness revealed that the ring is horseshoe‑shaped, with the broadest section of the ring located at the central portion (or closed end) to the north (also the highest point). Russo and colleagues

(2002:11) estimated that the ring is composed of 3,970 cubic meters of shell.

Although damage to the site is minimal, it is not without historic disturbance. A dirt road crosses the shell ring on its northern end and road grading has removed some of the shell in this area. A large hole is visible on the northeast portion of the ring. The purpose for this hole and when it was dug are currently unknown (Russo et al. 2002:13). Russo et al. (2002:12) suggested that shell may have been pushed from the ring into the interior at some point in prehistory and may be the result of attempts to drain the ring and central area. However, the stratigraphy of the shell deposits in the interior of the ring suggests that these were primary deposits, and not the result of a shell push (Saunders and Rolland 2006).

Figure 3.1. Shell Density Map of the Guana Shell Ring. This map depicts the density of shell identified through probing. Yellow areas indicate high‑ density deposits of shell and typically correspond to visible rises above the ground surface. The contour interval is 10 cm. The interior midden is located at approximately 340N,430E. (Map created by Michael Russo).

36 Previous Research

The Guana shell ring first received archaeological attention in 1985 when Louis Tesar and

Henry Baker performed a surface collection at the site. They determined the site was an

Archaic period shell ring based on the presence of fiber‑tempered ceramics, such as Orange

Incised. Since this time, only two excavations have taken place at the site.

From 2001 until 2002 Mike Russo and the Northeast Florida Archaeological Society

(NEFAS) organized limited excavation and mapping of the site (Russo et al. 2002). This investigation focused on mapping the ring (including subsurface shell depths) and obtaining radiocarbon dates. NEFAS excavated one 1‑x‑2 ‑m unit in the central portion of the ring (its highest point) and placed 13 shovel tests along the ring ridge. Several shovel tests were also placed within the southwest portion of the ring interior. The excavations produced Orange period ceramics, a variety of faunal remains, and modified faunal remains. Shovel tests immediately north of the ring produced deeply buried fauna and ceramics (Saunders and

Rolland 2006:14).

In 2005, Rebecca Saunders and Vicki Rolland began excavations in the ring interior. The goals of their excavation were to evaluate human activity within the ring interior, determine its chronological relationship to the ring, and examine whether the ring might represent ceremonial or quotidian activities (Saunders and Rolland 2006:7). Saunders and Rolland (2006) bisected a large feature (identified by shovel testing and coring) with a trench of four 2‑x‑2 ‑m units (the southeast corner of the trench began at 339N, 441E). Vertebrate faunal analysis revealed a diverse and somewhat atypical assemblage for a Late Archaic shell ring. Saunders and Rolland (2006:66) described an unusually high amount of Mercenaria spp. and cartilaginous fish recovered from excavations in the ring interior. They also stated that mammal remains were high in comparison to other rings. Faunal analysis of materials obtained within the ring

(Russo et al. 2006) may reveal a similar assemblage, or it may be more typical of a Late Archaic ring deposit.

37 Cultural Affiliations and Chronology

Fiber‑tempered Orange ceramics were the most common types recovered at the site and included both Orange Incised and Orange Plain varieties. Almost half of the ceramics recovered from the ring interior were incised, and many rims were decorated as well (Saunders and Rolland 2006:66). Excavations on the ring itself revealed a similar assemblage: slightly more than half of the Orange sherds were incised (Russo et al. 2002). Orange ceramics in the area have been found in contexts dating from 4200 to 3000 B.P. (Russo et al. 2002:21), suggesting a Late Archaic occupation. A sporadic St. Johns period occupation of the site is suggested by the presence of St. Johns sherds found in the upper levels of the ring and ring interior. St. Johns sherds have been found in contexts dating from 2500 to 500 B.P. (Russo et al. 2002:21). Russo and colleagues (2002) and Saunders and Rolland (2006) posited that occupation of the ring was not continuous from the Late Archaic through the St. Johns period, but reoccupation by St.

Johns people occurred some time after its abandonment.

Russo and colleagues (2002:21) proposed that ring construction began around 3500 to

3600 cal B.P. based on four radiocarbon assays. These dates were obtained from oyster shell in contexts with Orange ceramics in the lower levels of excavation and were not subjected to a marine reservoir correction. Saunders and Rolland (2006:46) stated that Russo and colleagues’ dates indicated that accumulation of the ring (at the base) occurred between 3300‑3500 cal B.P.

(one sigma) (they recalibrated the 2002 dates using the then‑current marine reservoir correction). A terminal date for the Archaic component of the ring was not established because

St. Johns ceramics were present on the top of the ring, and transitional levels containing both St.

Johns and fiber‑tempered sherds were not subjected to radiocarbon dating. However, Saunders and Rolland (2006:65) suggested that the accumulation of the ring probably did not span longer than 200 to 300 years based on what is known from other shell rings.

Saunders and Rolland (2006) obtained six additional radiocarbon dates from the ring interior area. They stated that deposition in the ring interior occurred began ca. 3700 cal B.P.

(one sigma), and reuse of the western interior portion of the ring occurred at ca. 2500 cal B.P.

(one sigma). Saunders and Rolland (2006) calibrated their dates and Russo and colleagues’

(2002) reported dates using a marine reservoir correction of delta R of 36+14 with the Calib 5.0

38 program. However, in their graph of radiocarbon dates (Saunders and Rolland 2006:47) they listed the uncorrected age as the conventional age for two of the 2002 dates (Beta 154816 and

154817). It is unclear if they entered these uncorrected ages into the Calib program (this would have caused incorrect calibration), but the calibrated dates presented in the table are identical to those presented by Russo et al. (2002:29).

In order to better understand deposition of the ring itself, I obtained four additional radiocarbon dates from column samples in the excavation unit on the central portion of the ring

(469N, 453E) and one date from a pit that extends below the shell ring (440N, 510E) (Appendix

A.1). All dates were obtained from quahog clam shells. To make these dates comparable with other dates from the ring, I have re‑calibrated all dates using the Calib 6.0 program (Stuiver and

Reimer 1993), the Marine09 curve, which takes into account global ocean effects, and the currently available marine reservoir correction for St. Augustine: delta 265+70. This reservoir correction was obtained from a Busycon specimen collected from St. Augustine in 1900 (Thomas

2008d). Because marine reservoir corrections are a fairly new application in archaeology, I present the corrected and recalibrated radiocarbon ages as well as the conventional radiocarbon age in Table 3.1. Figure 3.2 presents the approximate locations for these radiocarbon dates.

Although the four dates that I obtained from the column sample taken from the excavation unit (469N, 453E) appear to show a reversal between the upper and lower levels, all dates overlap at the one and two sigma ranges. I interpret this not as a stratigraphic reversal or disturbance in the deposit but as a rapid deposition of shell that is not definable using standard radiometric techniques. These dates indicate that ring deposition occurred sometime around

3400 cal B.P. (corrected). I obtained a single date from the bottom level of a pit feature identified in ST 440N, 510E (on the shell ring). Russo and colleagues (2002:15) identified this feature as a pit filled with oyster, clam, and coquina shell; the pit fill appeared to be the same as the ring midden material. Russo et al. (2002:15) stated that there was no visible indication of intrusion of the pit through the shell ring, but the small size of the excavation precluded further investigation. This date also indicates deposition some time around 3400 cal B.P. (corrected), providing further evidence that ring accumulation was rapid.

39 Table 3.1. Corrected and Calibrated Radiocarbon Dates for the Guana Shell Ring.

Conventional Std One Sigma Two Provenience Age BP dev BP Sigma BP Sample # Source 469N, 453E LV 3 3830 50 3360‑3560 3260‑3670 Beta 287032 A 469N, 453E LV 6 3840 50 3370‑3570 3270‑3680 Beta 287033 A 469N, 453E LV 7 3760 40 3280‑3470 3180‑3580 Beta 287034 A 469N, 453E LV 9 3720 50 3230‑3440 3110‑3550 Beta 287035 A 440N, 510E LV 8 3840 50 3370‑3570 3270‑3680 Beta 296139 A 340N, 540E LV 4 3860 60 3380‑3600 3270‑3720 Beta 154816 B 469N, 453E LV 12 3600 50 3080‑3310 2940‑3390 Beta 154817 B 380N, 400E LV 9 3490 70 2910‑3190 2800‑3300 Beta 165598 B 410N, 520E LV 6 3590 70 3050‑3320 2900‑3400 Beta 165599 B 469N, 453E LV 2 3820 70 3340‑3570 3210‑3690 GX‑29517 C Feat 1 top* 2740 70 2020‑2270 1880‑2240 GX‑31906 C Feat 1 center* 2880 70 2150‑2420 2040‑2610 GX‑31908 C Feat 5 center* 3620 70 3080‑3340 2940‑3440 GX‑31909 C Feat 2/4 top* 3740 70 3230‑3470 3100‑3600 GX‑31907 C

340N, 440E 55‑85 cmbs* 3720 60 3220‑3440 3080‑3560 GX‑29517 C

*Indicates a date from the ring interior.

Sources: A = This dissertation; B = Russo et al. 2001; C = Saunders and Rolland 2006

Dates from the remainder of the ring are more variable. Saunders and Rolland’s date from the top level of the excavation unit at the ring interior (469N, 453E) is very similar to my dates from the excavation unit, suggesting deposition at approximately 3400 cal B.P. (corrected).

Russo and colleagues’ date from the base of the excavation unit, however, is much younger: this date suggests deposition at approximately 3200 cal B.P. (corrected). It is currently unclear why this date is roughly 200 years younger than any of the other dates in this excavation unit. The date may have been obtained from displaced shell, or it may simply be anomalous. In fact, most of the dates obtained by Russo and colleagues are somewhat younger than mine or

Saunders and Rolland’s. These dates cluster around 3000 to 3200 cal B.P. (corrected), with a single date from 340N, 540E indicating deposition at approximately 3450 cal B.P. (corrected).

This date originates from the base of the southeastern portion of the ring.

Figure 3.2. Approximate Locations of Radiocarbon Dates. This figure depicts the approximate locations for radiocarbon dates. The provided radiocarbon dates are recalibrated and corrected and are reported at 2 sigma. (Basemap created by Michael Russo.)

41 Saunders and Rolland’s dates cover a wide range of time. Based on my calibrations,

Feature 2/4 was deposited sometime around 3350 cal B.P. (corrected), and Feature 5 appears to have been deposited at approximately 3200 cal B.P. (corrected). These calibrations indicate that ages of the shell ring and the ring interior midden are similar, but that the ring interior midden may be slightly younger than the shell ring. In order to determine whether these deposits are contemporaneous or not, more radiocarbon dates from the ring interior midden are needed.

Dates from the bottom of features (rather than the center or top of features) are needed to evaluate the initiation of deposition of these features.

Based on the new radiocarbon dates, it appears that the initial deposition of the Guana shell ring occurred around 3400 cal B.P. (corrected). The similarity of date ranges from the lower levels to the upper levels of excavation unit 469N, 453E suggests that deposition of remains was quite rapid, perhaps less than 100 years. The relationship and time span between the ring deposit and the ring interior midden are currently unclear. Saunders and Rolland

(2006) suggested that the ring interior midden accumulated prior to the ring deposit. Although new calibrations of the ring interior midden dates indicate slightly younger dates, it is possible that dates from the bottoms of these features (rather than from the top and center) would look more similar to or even older than those of the ring deposit. More radiocarbon dates from the ring interior midden (particularly from the bottoms of Features 2/4 and 5) may help clarify the situation in the future. Feature 1 is clearly much younger than the remainder of the site, and is likely associated with the St. Johns reuse of the site.

42 CHAPTER 4

CONTEXTUALIZING THE GUANA SHELL RING

Characteristics of Shell Rings of the Southeast United States

Introduction

Shell rings are circular, arc, or horseshoe shaped deposits of refuse that enclose a vacant or nearly vacant interior. The matrix of these sites consists primarily of dense shell, typically oysters, hard clams, mussels, tagelus, and a variety of other shellfish. The matrix usually contains little soil (although this is variable), and a plethora of estuarine, terrestrial, and marine faunal remains.

Shell rings are found in the United States solely along the Southeastern U.S. coasts of the

Atlantic Ocean and the Gulf of Mexico. They range from South Carolina to Mississippi, although their concentrations are limited to South Carolina, Georgia, and Florida. The northern limit of shell rings is in the Charleston, South Carolina area and the southern limit is in Jupiter,

Florida (Saunders 2002:47). Archaeologists are aware of more than 60 shell rings in the

Southeast (Russo et al. 2002), with a number of possible rings awaiting further study. Shell rings are primarily found on the landward side of barrier islands of the Atlantic Ocean, although several are located on the mainland. Howard and DePratter (1980) suggested that shell rings probably existed on the ocean side of islands as well as on other portions of the island. Thomas (2008c) argued that we might be missing the evidence of these rings due to erosion and marsh accretion. This is certainly possible, as a number of rings have suffered erosion and sediment burial in the marsh (Marrinan 1975, 2010; Russo and Heide 2002). One prominent feature of shell rings is that they are typically situated on estuaries or tidal creeks and salt marshes (Hemmings 1970). This proximity to the abundant marsh resources was probably a key factor for site location.

43 Most shell rings occur as a solitary ring, but several sites include more than one ring in close proximity to another and are called ring complexes. Some examples include Fig Island, which has three rings, the Sapelo rings, numbering three, and Skull creek, which has two rings.

The shape of shell rings varies from U‑shaped, C‑shaped, to full or nearly full circles. Florida’s rings are mostly C‑ or U‑shaped and larger than their South Carolina and Georgia counterparts;

South Carolina’s rings are mostly circular; and Georgia’s rings are a mix of shapes. Russo and

Heide (2001:491) argued that Florida rings seem to be both more structurally complex than those of Georgia and South Carolina, as some Florida rings (Rollins and Horr’s Island) possess a number of attached ringlets and other associated shell constructions. Ring orientation seems to cluster by location: rings in Georgia and South Carolina typically open toward streams and tidal creeks, whereas in Florida they open toward terrestrial areas (Russo 2004b:59). The size of shell rings is also variable. Rings range from 30 to 250 meters in diameter and are less than one meter to six meters in height, although shape and height is variable within individual rings as well

(Saunders 2004).

The artifact assemblage present in Late Archaic shell rings is rather limited (Hemmings

1970). Ceramics, when present, are typically the most common artifact and are found in both decorated and undecorated forms. The ceramics found in shell rings are varieties of fiber‑ tempered wares, as well as a sponge spicule‑tempered ware at the Joseph Reed shell ring in south Florida (Russo and Heide 2002). Bone pins are often recovered, as well as shell tools, and baked clay objects; relatively few lithics are recovered.

Temporality of Shell Rings

The vast majority of shell rings date to the Late Archaic period. Some rings predate the development of ceramics, and others provide some of the earliest ceramics in the Southeast

(Russo and Heide 2001). The oldest shell ring in the Southeast is the Oxeye ring, which dates to

4960‑4545 (two sigma) cal B.P. (Saunders 2002:48). Oxeye is located in northeast Florida, not far from the Rollins shell ring (circa 3600 cal B.P.). A number of Woodland period rings have been located (Russo et al. 2006), but most Late Archaic rings appear to be abandoned at some point during the Late Archaic. The upper levels of several Late Archaic shell rings contain Woodland

44 period artifacts. These do not represent continuous occupations; rather they are the reuse of previously abandoned sites (Marrinan 1975; Russo et al. 2002; Saunders and Rolland 2006).

DePratter (1977), Thomas (2011), and Rollins and Thomas (2011) proposed that this site abandonment was due to a disruption in the availability of shellfish and other estuarine resources. At the end of the Late Archaic, it appears that sea level retreated several meters, resulting in the loss of considerable estuarine environment and its resources (DePratter 1977;

DePratter and Howard 1980; Thomas 2011). At the end of the Late Archaic people may have abandoned shell rings to move closer to remaining marshes, resulting in sites that are buried under later marsh accumulation (Thomas 2011). Perhaps when sea level returned to its previous height Woodland people moved back onto the Late Archaic shell rings.

Shell Ring Research and Ring Function

Shell rings first received attention in 1872 in a letter from William McKinley describing the three

“mound circles, having plain areas” on Sapelo Island (McKinley 1873). Following the initial discovery of these shell rings, considerable time elapsed before excavation and research commenced. Excavation and publication continued to be sporadic until 100 years after their discovery. Early work was focused on South Carolina and Georgia; shell ring research in

Florida is a more recent undertaking.

Shell rings are unusual constructions and as such, they have led to numerous speculations about their purpose and function. Most propositions of ring function can be divided into two general categories: secular and sacred/ceremonial. Proponents of the secular category classify shell rings as the result of secular activities, such as quotidian refuse disposal.

Under this model, shell rings are typically viewed as habitation sites, which may have gradually developed due to circular living arrangements. This is known as the gradual accumulation model, and is discussed further below. The lenses of crushed shell present in shell rings are viewed as habitation places (people were living on the ring) (e.g., Trinkley 1985;

Cable 1997). Proponents of the ceremonial categorization characterize shell rings as the result of monumental architecture or feasting behavior. Under this model, shell rings are considered to

45 be the result of rapid deposition following ritual feasting and/or other ceremonial activities.

The interior of the ring is viewed as a plaza that is intentionally kept clean of debris and other artifacts. Some archaeologists have argued that shell rings were used by transegalitarian societies and served as arenas where individuals of higher status could display their wealth

(e.g., Russo 2004a; Saunders 2002). Proponents of a third model, the developmental model uphold that at least some shell rings initially formed as the result of habitations, but may have later developed into locations with ceremonial and ritual events and meaning (Thompson 2007).

Speculations on ring function started with McKinley’s initial description of the shell rings at Sapelo. In his brief letter in 1872, he proposed that they were “doubtless for council or games,” or could have served as location for pow‑wows, state houses, places of torture of captives, places for dances, and/or places for athletic sports and games (McKinley 1883:423). He did not detail the composition of the mounds, but commented that the surrounding amorphous middens contained ceramics and shellfish remains, indicating places of habitation.

Moore was the first to excavate the Sapelo Rings (Moore 1897). He represented the three rings as more symmetrical than they actually were, which was common of many early archaeologists working on shell rings (Russo 2004a). Moore’s (1897) report on Sapelo argued that the deposit is certainly aboriginal in design, and is “one of those symmetrical works of the aborigines made by piling shell through a period of time to form some definite shape.” He did not specifically hypothesize about function, but did consider the rings to be an intentional design.

In 1940, Waring suggested that shell rings may have functioned as fish weirs, but it was more likely that they were the result of midden accumulation around a structure (Waring

1968b). In 1943, Flannery reported on Ritter’s 1932 and Moorehead’s 1933 excavations at

Chesterfield ring. Flannery (1943) proposed that the interior center of the ring had been occupied. Edwards excavated the Sewee Pines ring in 1965 and interpreted ring function to be both a habitation space as well as fish trap (Saunders 2002).

Waring and Larson (1968:273) stated that the Sapelo ring was “composed of occupational midden in primary position which was deposited as the result of habitation sites located on the ring.” This idea later became part of the gradual accumulation model. Despite

46 the quotidian nature of the artifacts, Waring and Larson left room for interpretation as to the function, saying, “The Sapelo shell ring then very likely represents a ceremonial or social arrangement rather unusual in this geographical location and time horizon” (Waring and

Larson 1968:273). In Waring’s paper entitled The Archaic and Some Shell Rings, he is more specific about his thoughts on shell ring function. He refers to the shell rings as the earliest monumental architecture in North America (Waring 1968a:243), and says: “What they were used for the Lord alone knows. They were certainly ceremonial enclosures of some sort and reflect a level of ritual complexity only partially suspected for so remote a time period” (Waring

1968a:246).

In 1968, Calmes published a report on his excavation at the two Ford’s Skull Creek rings and at Sea Pines. Calmes supported Waring and Larson’s (1968) gradual accumulation model.

Calmes (1968) suggested the inhabitants might have lived in a circle and disposed their refuse in a circle. He further posited that inhabitants may have walked over or lived on this refuse, causing crushing of shell and other midden material.

Marrinan’s (1975) research at the Cannons Point ring and the West ring yielded inconclusive results regarding ring function. However, she did rule out the possibility of rings functioning as fish traps. She found evidence for climatic change following occupation of the ring, which suggested the ring was not inundated by marsh tides at the time of its use. She stated, “Shell rings are purposeful accumulations of subsistence and cultural debris” (Marrinan

1975:117).

Trinkley’s (1985) publication on Lighthouse Point and Stratton Place rings focused specifically on function. He elaborated on Waring and Larson’s gradual accumulation model and argued that rings started out as a circle of house pits which filled in. He believed that people lived on the rings, causing the crushing evident in the deposits. He argued that the circular ring shape might be attributed to the egalitarian practice of living in circular arrangements (Trinkley 1985:118).

Russo’s (1991) work at Horr’s Island led him to believe that deposits characterized by loose, clean shell having little to moderate amounts of soil were indicative of feasting deposits, and not of gradual accumulation. Russo worked on several shell rings with other

47 archaeologists, and they identified deposits of this type at Joseph Reed (Russo and Heide 2002), at Rollins shell ring (Russo et al. 1993; Saunders 2004), at Guana (Russo et al. 2002), and at

Oxeye (Russo and Saunders 1999). Saunders has argued that ring construction was a planned undertaking, based on radiocarbon dates on both the East and West arms of the ring at Rollins that date to approximately the same time (Saunders 2002).

Russo (2004a:29) argued that regardless of whether rings are village sites or solely ceremonial sites, members were consuming and discarding food refuse “in great quantities suggestive of large‑scale feasting.” He argued that this feasting was ritual in nature, and suggested that the spatial patterning of shell rings indicated a transegalitarian social structure

(Russo 2004a:29). He stated that the rapid deposition of shell evident at rings is reminiscent of

“a community project requiring at least temporary leadership of a status sufficient enough to compel massive labor contributions from the populace” (Russo 2004a:67). Russo used social space theory to identify areas of high status in the center U‑shaped arrangements; he argued that these locations correspond to the highest deposits of shellfish in shell rings. He also argued that people constructed shell rings out of feasting remains to memorialize the feast (Russo

2004a).

Like Russo, Saunders (2002) also favored a more ceremonial model. She argued that the morphology of rings is not suggestive of village locations – namely the intentional mounding of shell and the presence of feasting deposits. Saunders argued that the smaller enclosures present at some rings (such as Rollins) along with higher concentrations of artifacts indicate a more hierarchical and complex society than an egalitarian one. She proposed that the gradual accumulation model with occupants living on top of rings was not a likely scenario due to the height of some rings and the difficulty of climbing to the top of them (Saunders 2002).

However, Saunders (2002) stated that rings served both habitation and ceremonial functions, and can be construed as ceremonial villages. Saunders (2002:85) further argued that habitation debris could be the result of continued habitation at a location or “found at a site in which

‘living’ occurred once a year at an annual ceremony.”

Although multiple lines of evidence have been used to identify sedentary occupations of shell rings, Claassen (2010:9) proposed that shell rings and shell middens were not habitation

48 sites but camps for the duration of rituals. She stated: “Where shell heaps have been forced into the role of mundane villages providing proof of sedentism, I highlight the evidence that they were occupied only sporadically and that sedentism was not characteristic of these [Archaic] people” (Claassen 2010:10). Nevertheless, Thompson and Turck (2009:268) argued that at least a portion of the population probably occupied shell rings throughout the year and suggested that most archaeologists view shell rings as, at least in part, places of daily habitation.

Thompson (2006) proposed that this sort of daily habitation resulted in the shell rings on

Sapelo Island, Georgia. Thompson found Ring 3 to be the result of daily activities and subsistence, not the result of intentional mound construction or large‑scale feasting (Thompson

2006:261). He argued that the seasonality assessment, analysis of activity areas, and the composition of the ring deposit were the result of daily activities and habitation; he further proposed that habitations were likely located in the interior or adjacent to shell deposits

(Thompson 2006; 2007). Further consideration of the Sapelo ring complex culminated in a blending of the sacred and secular models. Thompson (2007:92) stated that the evidence at

Sapelo supported the gradual accumulation model of midden accretion, but this did not negate the possibility of ceremonialism. As a result, Thompson (2007) proposed the developmental model, which he described as a diachronic perspective based on Binford’s (1982) Archaeology of Place – essentially the nature of occupation and function of a place changed through time.

The developmental model posited the following: 1) Shell rings initially developed from a gradual accumulation and then took on a circular shape; 2) As the circular shape became apparent, residence shifted towards the inside of the ring; 3) The ring perhaps ceased to fulfill a residential role and may have developed a more ceremonial function; at this time, the ring could become an intentional construction or monument (Thompson 2007). While Thompson viewed these changes in ring function as a possible process, he noted that the function of rings may alternate between residential and ceremonial, and that not all shell rings developed in this fashion. Thompson and Andrus’ (2011) seasonality assessment identified seasons of shellfish deposition during all four seasons of the year at the Sapelo complex. They argued that this indicated at least one function of the rings was daily habitation. Rings III and II appear to be the result of accumulations of living debris and not the result of feasting or intentional

49 mounding (Thompson and Andrus 2011). Ring I accumulated rapidly and appeared to be the result of limited seasonal collection, which they argue indicates a large processing event that could result from feasting (Thompson and Andrus 2011). If this is the case, then the Sapelo ring complex may demonstrate the shifting role of shell rings as indicated by the Developmental model.

Seasonality Studies at Late Archaic Shell Rings

A number of seasonality studies have been undertaken at Late Archaic shell rings in the

Southeast. This section will outline these studies and their results to provide an overview of the known seasons of occupation at Late Archaic shell rings. These seasonality studies are based primarily on estuarine fauna, both vertebrate and invertebrate. Table 4.1 provides a list of fauna that are commonly assessed and the methods by which they are analyzed. Methods of analysis include those based on incremental growth, stable isotopes, and modal size‑class.

Although size‑class was used to identify seasons of coquina use at one ring, this methodology has been deemed to be inaccurate (Quitmyer et al. 2005).

Seasonality studies at shell rings are a relatively recent endeavor. Marrinan’s (1975) work at the Cannons Point and West rings entailed the first analysis of subsistence remains using fine‑mesh recovery techniques (Russo 2006:45). She proposed a tenuous occupation of spring through fall based on faunal and floral remains recovered in excavation. She determined the seasonality of botanicals according to when their fruit, seeds, or nuts were available.

Quitmyer and colleagues (1985a and b) evaluated a small sample of quahog clams from

Cannons Point and determined that clam exploitation occurred during the spring season. The combined floral and faunal indicators at Cannons Point demonstrate that deposition took place in every season of the year at the Cannons Point ring.

Russo’s (1991; 1998) work at the Horr’s Island ring site included a seasonality study that revealed occupation throughout the year. He did not differentiate his seasonality estimates of the ring from those of the overall site, so I will present them together as he does. Russo utilized two different methods to determine seasonality: modal size class and incremental growth. He

50 Table 4.1. Fauna Used in Seasonality Studies Discussed in this Dissertation

Taxa Common Name Method Element Ariopsis felis Hardhead Catfish Size Class Otolith; Pectoral Spine Ariopsis felis Hardhead Catfish O2 Isotopes Otolith Lagodon rhomboides Pinfish Size Class Atlas Micropogonias undulatus Atlantic Croaker Size Class Atlas; Otolith Micropogonias undulatus Atlantic Croaker O2 Isotopes Otolith Opisthonema oglinum Threadfin Herring Size Class Atlas Brevoortia tyrannus Menhaden Size Class Atlas Crassostrea virginica* Eastern Oyster Size Class Boonea impressa Crassostrea virginica Eastern Oyster O2 Isotopes Valve Argopecten irradians Bay Scallop Size Class Valve Mercenaria spp. Quahog Clam Increment Valve Mercenaria spp. Quahog Clam O2 Isotopes Valve Donax variabilis Coquina O2 Isotopes Valve

*Crassostrea virginica as evaluated by the impressed odostome, Boonea impressa.

used modal size class to evaluate oysters by proxy of Boonea impressa, a parasite that bores into oysters and has a known rate of growth. Russo (1991) found that Horr’s Island occupants gathered oysters during the late summer. He assessed scallop seasonality using modal size class and found that scallops were gathered during the late summer or early fall (Russo 1991).

He assessed quahog clam seasonality via the incremental growth technique and determined that clams were gathered predominantly in spring and summer but were collected throughout the remainder of the year as well. Russo (1991) used modal size class studies to assess hardhead catfish, pinfish, and threadfin herring and found the following patterns. Hardhead catfish were captured primarily during fall and a few individuals were caught during the summer. Most pinfish were caught in the late summer and fall, with a few mid‑summer collections and threadfin herring were probably caught during the winter.

At the Rollins shell ring in northeast Florida, Russo and colleagues (1993) conducted seasonality studies on oysters, menhaden, croaker, and pinfish. They found that every season except fall was represented in the area around the shell ring. Seasonality of the shell ring alone yielded late winter/early spring through summer occupation. Based on Boonea impressa, Russo

51 et al. (1993) found a January period of collection for oysters. Menhaden were analyzed from two proveniences and suggested a winter/early spring collection. Croaker and pinfish were examined from the same two proveniences, and were both collected during the summer.

Coquina were also analyzed and demonstrated a late spring and summer collection. However,

Quitmyer et al. (2005) found that size class studies of coquina were inaccurate, and should not be used (because all size classes are present throughout the year and because archaeological specimens are often larger than modern specimens). Russo et al. (1993) do not report their samples sizes except to say that they are small and that this may account for the absence of the fall season in the sample (Russo et al. 1993:172).

Thompson (2006) addressed seasonality at Ring 3 on Sapelo Island, Georgia as part of his dissertation. He conducted seasonality studies of quahog clams and oysters. Thompson

(2006) analyzed the oxygen isotopic composition of ten oysters and found that they were gathered almost exclusively in the cool months of winter. He analyzed the oxygen isotopic composition of 10 clams as well, which revealed exploitation in every season of the year (six winter, 2 spring, 2 fall, 1 summer). He also analyzed the incremental growth structures of 113 clams throughout the assemblage in an attempt to identify different seasons present in the assemblage. Of these, only 42 were useful for identification to the seasonal level; these clams displayed a year‑round collection pattern (Thompson 2006). Thompson also argued for a summer or fall occupation of the site based on the seasonal abundance of hardhead catfish and star drum at this time, although they are present throughout the year in the estuary (these fish are the most common species in the assemblage). The presence of clams that were exploited in all four seasons of the year provides the strongest evidence for a year‑round occupation of

Sapelo Ring 3.

Thompson and Andrus (2011) built on this earlier work at Sapelo and assessed seasonality of oysters and quahogs clams from Rings I and II. Their recent publication in

American Antiquity details the seasonality assessments conducted at the ring. They assessed incremental growth structures for 620 quahog clams from Ring III and Ring II (280 were successfully assessed), and analyzed isotopic composition for 41 clams and 17 oysters. Their results indicated that shellfish were gathered throughout the year, indicating year round

52 deposition at the site. Thompson and Andrus (2011) found that Rings II and III were accumulated gradually, but Ring I is primarily composed of shellfish collected in a single season

(winter).

Russo and Heide’s excavation at the Joseph Reed shell ring in south Florida was small, but seasonality studies of quahog clams and oyster were undertaken. Their results demonstrated a summer collection of quahog clams (in a single unit), and a fall collection of oysters by proxy of Boonea impressa (in a single feature) (Russo and Heide 2002). They did not report on sample sizes, and stated that more work is needed to thoroughly evaluate seasonal occupation.

Saunders (2002) excavated the Fig Island complex and found limited markers of seasonality. Faunal samples were examined from Fig Island Ring 1 and Ring 3. Boonea impressa indicated that oysters were collected in the late fall and winter (Saunders 2002). Saunders hoped to analyze seasonality of fish by examining otoliths, but she did not recover sample sizes large enough for modal size class studies, and the otoliths were too small (from fish that were too young) for incremental studies. The presence of a number of fish suggests summer seasonality, but Saunders (2002) cautions that this is only a “best guess.” Quahogs could not be analyzed at the site because few were recovered and many were highly fragmentary.

Colaninno (2010) conducted isotopic analysis on hardhead catfish (Ariopsis felis) and

Atlantic croaker (Micropogonias undulatus) otoliths from four shell rings in the Georgia Bight: St.

Catherines (n=16), McQueen (n=7), Cannons Point (n=1), and West (n=7) (sample sizes are those that yielded conclusive results). Her results indicated that fishing for these species occurred during all four seasons of the year at the St. Catherines, McQueen, and West rings. She found that at the Cannons point ring, fish otoliths indicated a cold weather capture, which I have interpreted as winter. Other resources indicated year‑round deposition at the Cannons Point ring.

Quitmyer and Jones (2012) have undertaken clam seasonality studies at the St.

Catherines ring (n=117) and the McQueen ring (n=68) on St. Catherines Island. Clam seasonality assessments at these sites were based on the St. Catherines comparative collection,

53 Table 4.2. Seasonality Results at Shell Rings in the Southeast

Shell Ring Winter Spring Summer Fall Fig Island O O St. Catherines Ring CV CV V V McQueen Ring CV CV V V Sapelo Ring Complex CO C CO CO Cannons Point Ring V C F F West Ring V V V V Rollins Ring OV V V Guana Ring C C Grand Ring CV CV V V Horrʹs Island CV C COV CV Joseph Reed C O

Key: C= Clams O= Oysters V= Vertebrate resource F= Floral remains

which only permits differentiation between cool weather (winter and/or spring) and warm weather (summer and/or fall) exploitation (Quitmyer and Jones 2012). Quitmyer and Jones

(2012) identified cool weather collection (winter through spring) at both shell rings, but they were unable to determine whether exploitation occurred solely during winter or spring.

Nevertheless, these findings show that clam exploitation was a seasonal occurrence at both sites, despite continued occupation throughout the year.

Table 4.2 compiles the seasonality studies at the shell rings I have discussed above. As the sample sizes are not available for most of these studies, they cannot be considered as equally rigorous indications of seasonal occupation. Zooarchaeologists must make their sample sizes explicit in order to evaluate the comparability of these studies. Nevertheless, the data provide a basic pattern of multi‑seasonal semi‑sedentary to sedentary occupation at shell rings.

As the seasonality studies at shell rings become more common, we can make better comparisons of seasonal resource use as well as issues of sedentism. This will help form an

54 understanding of how shell ring occupants interacted with their environmental and social surroundings on a seasonal basis. Based on the data at hand, the evaluated shell rings appear to be the products of semi‑sedentary or sedentary groups who relied on estuarine resources throughout much, if not all of the year. Some resources, such as clams and oysters, appear to have been gathered on a seasonal basis despite continued deposition at the site throughout much, if not all, of the year.

55 CHAPTER 5

CLAM BIOLOGY AND SEASONALITY STUDY

Clam Biology

Mercenaria mercenaria, Mercenaria campechiensis, and their Hybrids

The genus Mercenaria, commonly known as quahogs or hard clams, includes two species and their hybrids. M. mercenaria (northern quahogs) and M. campechiensis (southern quahogs) are closely related and probably evolved from a common ancestor (Harte 2001:16). The taxonomy for M. mercenaria and M. campechiensis is as follows:

Kingdom: Animalia Phylum: Mollusca Class: Bivalvia (Linnaeus, 1758) Order: Veneroida (Adams and Adams, 1856) Family: Veneridae (Rafinesque, 1815) Genus: Mercenaria (Schumacher, 1817) Species: Mercenaria mercenaria (Linnaeus, 1758) Species: Mercenaria campechiensis (Gmelin, 1791)

M. campechiensis and M. mercenaria look similar, but morphological differences are present. The exterior shell of M. campechiensis has well‑defined concentric ridges, while M. mercenaria has less defined ridges and a smooth portion on the central area of the exterior shell

(Figure 5.1). M. campechiensis is heavier than M. mercenaria (Abbott and Morris 1995) and often reaches larger sizes than M. mercenaria (Rehder 1981). For example, Schroeder (1924) reported a bed of southern quahogs (M. campechiensis) in the Ten Thousand Islands that contained clams that averaged one pound each, and some weighed at least two pounds. Although M. campechiensis usually reaches larger sizes, M. mercenaria can grow up to 15 cm, the maximum size reported for M. campechiensis (Harte 2001:7). The shell of M. campechiensis is generally thicker and rounder than M. mercenaria (Dall 1903). The lunule in M. campechiensis is at least as

Figure 5.1. Clam Terminology This figure depicts the interior and exterior of a left valve, or side of a quahog clam.

wide as long, whereas M. mercenaria has a narrower lunule (MacKenzie et al. 2002a). The adult shell of M. campechiensis lacks the internal purple color seen on the interior margin of M. mercenaria (Harte 2001); however, some varieties of M. mercenaria also lack purple color (e.g., M. mercenaria var. notata). For a detailed description of the differences between M. mercenaria and

M. campechiensis, see Harte 2001:16‑18.

The ranges of M. mercenaria and M. campechiensis overlap considerably. M. mercenaria ranges from Florida to Canada in shallow waters (Hilbish 2001:274). M. campechiensis occurs through the Caribbean to Cape May New Jersey (Hilbish 2001), but occurs primarily from

Florida to southern Mexico (MacKenzie et al. 2002a). In Florida, M. mercenaria typically occurs in northern Florida along the Atlantic Ocean, and M. campechiensis is usually found in Southern

Florida along the Atlantic and all along the Florida Gulf Coast. Throughout their overlapping ranges in the U.S., the two species infrequently produce viable hybrid offspring. In the Indian

River Lagoon system in Florida, however, the two species frequently hybridize (Hilbish

2001:273). While M. mercenaria and M. campechiensis occupy slightly different habitats, the proximity of these habitats in the Indian River Lagoon allows interbreeding to occur regularly

57 (Hilbish 2001:275). Hybrids of M. mercenaria and M. campechiensis share qualities of both species to varying degrees. Hybrids contain a wide range of morphological variation that includes a mixture of shell sculpture, interior coloration, and valve thickness (Harte 2001:14). Hybrids are typically larger than M. mercenaria but smaller than M. campechiensis (Arnold et al. 1998).

Although morphological differences are present, the only way to identify hybrids with certainty is by testing soft tissues using techniques such as allozyme electrophoresis.

M. mercenaria has long been a commercially valuable species and is one of the most valuable aquaculture species on the East Coast of North America. Commercially farming aquatic animals, referred to as aquaculture, is a common practice in North America, and

Mercenaria aquaculture is the largest and most valuable aquaculture industry on the East Coast

(Whetstone et al. 2005:1). Most aquaculturalists grow M. mercenaria because it can be kept alive longer after its harvest than M. campechiensis, which spoils faster (Phil Cubbedge, personal communication 2010). This longer shelf life, so to speak, has led to the introduction of M. mercenaria along the Gulf Coast where they were previously absent. M. mercenaria var. notata

(Say, 1822) is a variety of M. mercenaria that has been used by many aquaculturalists, particularly to observe the rate of success of cultured juveniles in the wild (Harte 2001:13). M. mercenaria var. notata is marked by thin zigzags on the exterior of the shell (especially prominent in younger, smaller clams) and a lack of interior purple coloration (Harte 2001:13). This variety occurs throughout the natural range of M. mercenaria and appears to be caused by “a heterozygous condition of a binary genetic locus for white (no color) and a reddish brown

(color)” (Harte 2001). The reddish brown color occasionally occurs in patches along the margin where the purple coloration occurs in M. mercenaria.

Mercenaria occur in small patches to extensive beds on sand to muddy sand substrates

(Harte 2001:11). Clams can occur in very high densities (up to >500 individuals per m‑2), but are commonly found in densities ranging from 5 to 20 individuals per m‑2 (m‑2 is used to denote the average amount of clams in one square meter using data distributed over several square meters)

(Fegley 2001:383; MacKenzie et al. 2002a:3). Mercenaria are most frequently encountered on sand to muddy sand sediments, but also are found in eel grass beds, in shelly areas near oyster beds, and between and underneath oysters (Harte 2001:11‑12). I also have observed Mercenaria

Figure 5.2. Injured Clam This clam was gathered as part of the St. Augustine collection. Clams like this were occasionally encountered during collection. Archaeological collections did not include injuries that were this severe, although I did observe undulations in the margin as well as chips and other damage to the shell. The frequency and severity of injuries in the modern collection may be the result of the non‑native predator, Melongena corona (crown conch).

living in the outer edges of marsh grass (Spartina spp.). According to Grizzle et al. (2001:350), various experiments have reported that clams grow fastest on sandy substrates with little silt‑ clay or mud. Fegley (2001:387) reported that other studies have found that heterogeneous substrates (with shell or sea grass) can support higher densities than homogeneous substrates.

M. campechiensis are often found at greater depths than M. mercenaria, sometimes as deep as 12 m (Harte 2001). M. mercenaria are typically found just beneath the low tide level but can also be found above the low tide level on sand bars and marsh edges. During my yearlong collection, I observed that during the hottest months of the year Mercenaria were often buried up to 10 centimeters beneath the sand; in the coldest months of the year they were often quite shallow, and many were just beneath the surface of the sand.

Mercenaria occur in both ovate and elongate forms, which have a shortened and somewhat rounded posterior margin and an elongated and somewhat pointier posterior margin, respectively. Mercenaria develop malformations rather rarely, although they do exist.

Many animals prey on Mercenaria throughout its life including birds, stingrays, fish, sponges, shrimp, crabs, hermit crabs, starfish, worms, and other mollusks (especially whelks and conchs)

(Grizzle et al. 2001). Occasionally an animal injures the shell of a Mercenaria, but fails to kill the

59 animal, resulting in permanent damage to the shell (Figure 5.2). In my observations, shell injuries typically occurred along the posterior ventral margin. This is likely due to the orientation of the shell as it lies under the sand: with the anterior portion of the shell (and umbo) downward and the posterior upward. This orientation makes the posterior ventral margin vulnerable to predators. Injuries range from minor chips in the exterior of the shell to perforation of the shell (allowing access to the soft tissues of the clam) and large malformations in growth. Serious injuries and large malformations can lead to changes in the overall shape of the valve as well as undulations of the exterior concentric rings and in the shell margin.

Temperature is an important factor for growth of Mercenaria. In the southeast U.S., optimal growth occurs during the milder temperatures of winter and spring, whereas in the

Northeast optimal growth occurs during the warmer months of summer and fall. From

Virginia northward, clam growth ceases during the coldest months of winter, but Florida clams grow throughout the year (although not at a uniform rate) (Grizzle et al. 2001:336). The year‑ round growth of Florida clams affords them the fastest shell growth rate in the range of

Mercenaria (Ansell 1968). Hard clams reach marketable size (umbo width of one inch [2.54 cm] or a length of about 48 mm) in Florida in as little as 1.9 years, whereas in Canada it can take up to 13 years (Grizzle et al. 2001:337). Maximal shell growth occurs from 68‑75.2° F (20‑24° C) and ceases at temperatures above 87.8° F (31° C) and below 44.6° F (7° C) (Grizzle et al. 2001:341).

The pumping (feeding) rate of clams is temperature‑dependent, indicating that temperature

(and consequently, feeding rate) is the major determinant of growth (Grizzle et al. 2001:341).

Temperatures between 59‑77° F (15‑25° C) are optimal and yield rapid growth opaque shell increments; extreme high and low temperatures yield slow growth translucent shell increments

(or no shell growth if temperatures are too extreme) (Grizzle et al. 2001:342).

Salinity is also an important factor in clam growth and overall health, although to a lesser extent than temperature. While clams are tolerant of reduced salinities, growth rate typically decreases (Grizzle et al. 2001:342). Pumping rates decrease below 20 parts per thousand (ppt) and above 30 ppt (Grizzle et al. 2001:342). If such salinity extremes are sustained they can result in growth cessation and death.

60 Life Cycle and Reproduction of Mercenaria spp.

Hard clams can live to at least 46 years according to MacKenzie and colleagues (2002b:64). The oldest known archaeological clam is from St. Simons Island and was 35 years old when it was collected (Marrinan 1975; Quitmyer et al. 1985b). Mercenaria are characterized by an r‑selecting reproductive strategy: they attain sexual maturity at a young age (one to two years old) and small size (30‑35mm in length), and fecundity is high throughout life (Eversole 2001:253).

Sex of clams cannot be determined by shell morphology; sex must be determined by spawning or microscopic examination of the gonad (Eversole 2001:222). At younger ages and smaller sizes, males outnumber females in the population; however, as clams mature (by two years old) the sex ratio of the population levels out (Eversole 2011; Walker 1994 cited in

Eversole 2001). Sexual maturity occurs at shell lengths between 30 to 35 mm; Mercenaria in the southern portions of their range can reach this length in as little one year (Eversole 2001:225).

Size is not the only factor in reaching sexual maturity – age, genetic traits, and environmental conditions that affect the length of the growing season also influence age at first sexual reproduction (Eversole 2001:225).

Eversole (2001:228) describes the following reproductive cycle for Mercenaria: 1) recuperative stage after spawning, 2) active developing stage, 3) ripe stage, 4) spawning, and 5) spent stage. Spawning takes place externally when mature gametes are released into the water column (Whetstone et al. 2005). The reproductive cycle and spawning peaks vary with latitude and are influenced by internal and external factors, especially temperature (Eversole 2001). The optimal spawning temperature is 79° F (26 °C) (Whetstone et al. 2005). Research has shown that in the Indian River M. mercenaria begin spawning approximately one month earlier than M. campechiensis (March and April, respectively), but otherwise spawning length and the reproductive cycle is the same (hybrids follow the spawning patterns of the maternal parent but spawn for a shorter duration and less intensely than either of their parents) (Eversole 1997:107).

Although spawning occurs over a greater number of months in very warm climates, spawning periods are limited in duration due to high summer temperatures. Similar to other aquatic species, Mercenaria living in southern latitudes begin spawning earlier and spawn over a longer period that is often interrupted by very high temperatures; this interruption results in a second

61 spawning period (Eversole 2001). In these cases, the spring spawning peak is typically more extensive and more productive than the second spawning peak in the fall (Eversole 2001). In

Florida, the following bimodal spawning period has been observed: a major peak in the spring that continues into the summer (March – June), a period of inactivity in the late summer as a result of temperature extremes over 86° F (30° C) (July – September), and a minor peak in the fall (late September – November) (Eversole 1997; Hesselmann et al. 1989). Larger, older clams spawn for a longer period and are more productive than their smaller and younger counterparts (Eversole 2001; Grizzle et al. 2001). Mercenaria do not appear to have a stage of reproductive senescence (a decline in reproductive effort and success), even in old age.

Peterson (1986) reported that a 46 year‑old clam showed no signs of reproductive senescence.

Clams devote a considerable amount of their resources to spawning and preparation for spawning. Eversole (2001:236) stated that clams apportioned approximately 30‑50 percent of their yearly organic growth to reproduction. In England, Ansell and Lander (1967) estimated the amount of tissue devoted to spawning for two clams (each measuring 40 mm in shell length): they devoted 1.0 g and 1.4 g out of 3.75 g and 5.4 g of soft tissues, respectively. They estimated that spawning results in a loss of 20‑25 percent of the total organic production every year. It appears that as clams reach a greater size they devote less energy to shell growth and more to reproduction (Peterson 1983). Peterson and Fegley (1986) argued that clams stored resources during December and January in preparation for reproduction in the spring. They based this argument on the observation that sexually mature clams do not grow as much as juvenile clams (especially their shells) during December and January. They argued that this was likely a result of the diversion of resources in order to store energy and nutrients in preparation for reproductive activity in the spring (Peterson and Fegley 1986:608). This storage of nutrients also effects overall clam biomass. According to Peterson and Fegley, overall biomass increases in winter and especially spring, and declines in summer and especially fall.

In other words, clams store resources and gain mass prior to spawning in the spring, then lose considerable mass during and following spawning in the summer and fall. During spawning, glycogen levels decline, further reducing the caloric value of the shellfish (Quitmyer 1985:29).

62 Quahog Seasonality Studies

Shell Growth and Increment Deposition

Mercenaria grow by daily deposition of calcium carbonate (CaCO3) primarily along the margin, or outer edge, of the shell (Monks 1981). Mercenaria add new shell concentrically along the margins and across the interior surface (Fritz 2001). This means that the interior of the hinge, or umbo, and the marginal end of the clam contain the most recent growth. Significant for archaeological endeavors, Mercenaria valves, or sides, are identical in size and shape (Eble 2001); therefore only a single valve is necessary to view the organism’s ontogenetic growth history

(and seasonal markers). Because Mercenaria record environmental data about daily, seasonal, and annual cycles in their shells, they have been useful for: 1) Molluscan biology and ecology

(fisheries management, age determinations, population evaluation, and growth rates), 2)

Paleontology/Paleoecology (inferences about paleoenvironments, duration of day, seasons, and year), 3) Environmental studies (determination of natural and anthropogenic effects on individual and population growth), and 4) Archaeology/Archaeomalacology (determining seasons of occupations of prehistoric and historic sites based on seasons of death) (Fritz

2001:53).

Similar to trees, Mercenaria lay down two growth rings, or increments, per year (one light colored opaque increment and one dark colored translucent increment (Clarke [1979] termed these opaque and translucent increments based on their optical properties). By cross‑ sectioning the clam along its greatest height, one is able to view the entire ontogenetic sequence of growth (Figure 5.3). One can determine the age of a clam at the time of its death by counting the sets of opaque and translucent increments (one year = one opaque and one translucent increment). The opaque layer is deposited on the shell during times of optimal feeding for the clam – when temperatures are mild and food is plentiful; the translucent increment is deposited in times of stress – when temperatures are extreme (too hot or too cold) or during spawning

(Quitmyer et al. 1985a). Extreme stress caused by hurricanes, injuries, and other environmental activity can create false translucent increments in shells (Claassen 1998). These false increments

Figure 5.3. Ontogenetic Sequence of a Quahog Clam. This figure depicts the ontogenetic sequence of a clam beginning at the umbo and terminating at the margin. One year of growth is represented by a set of one opaque and one translucent increment. The same increments are visible through the body of the shell and in the umbo.

can be discerned by their size (compared to other increments) and their proximity to other translucent increments. The opaque increment appears white due to the increased amount of

CaCO3 in the shell structure (the translucent increment has a depleted amount of CaCO3 and an enriched amount of conchiolin available for incorporation into the growing matrix) (Claassen

1998:199).

The timing of deposition of the opaque and translucent increments is highly dependent on latitude and the onset of optimal and suboptimal temperatures. Arnold and colleagues

(1998) reported that while there are a number of factors that influence seasonal growth patterns

(e.g., salinity, reproductive activity, food availability, hydrodynamics, predation, light regime, sediment composition, and population density), temperature is the primary factor operating on seasonal growth patterns. Other studies employing isotopic analysis have also identified temperature as a primary factor in incremental and seasonal growth (Jones and Quitmyer 1996; this dissertation). In the southeastern U.S., the opaque increment occurs during rapid (optimal)

64 growth in the cooler months of winter and spring. The translucent increment occurs during slow (suboptimal) growth throughout the hotter months of summer and fall when water temperatures are high. In the North Atlantic, the growth pattern is reversed: opaque (rapid growth) occurs during the warmer months and translucent (slow growth) occurs during the colder months. Some clam populations in more northern latitudes deposit more than two annual increments. In the mid‑Atlantic region, Mercenaria annual growth consists of four increments: two opaque (spring and fall) and two translucent (summer and winter) (Fritz 2001).

This will be discussed further in Chapter 7. The reversal in season of increment deposition occurs because in the Southeast, temperatures are more extreme in the summer and fall and more moderate during the winter and spring, while the reverse is true for the Northeast.

Moderate seasonal temperatures provide optimal feeding conditions for clams, facilitating deposition of fast growth opaque increments.

It is important to note that the season of death can be identified for M. mercenaria, M. campechiensis, and their hybrids using the same methodology. Arnold and colleagues (1998) found minor genetic (species‑specific) differences in the timing of the onset of seasonal growth.

Compared to M. mercenaria, M. campechiensis and hybrids are slower to transition from translucent to opaque growth, resulting in a decreased duration of optimal (opaque) growth

(Arnold et al. 1998). This slow transition results in an increased number of M. campechiensis and hybrids exhibiting translucent growth during the winter, and to a lesser extent in the spring, essentially creating greater variability of growth phases in each season. Despite the genetic influence on seasonal growth, Arnold and colleagues (1998:106) concluded that environmental factors (namely temperature) were the dominant influence on seasonal patterns of growth. In sum, temperature is the primary determinant of seasonal growth; therefore, the incremental growth technique works for both Mercenaria species and their hybrids.

History of Quahog Seasonality Studies

Seasonality studies began to appear in archaeological publications in the 1960s, but it was not until the 1980s and 1990s that they became a significant focus of archaeological research.

Although seasonality studies are a relatively recent endeavor in archaeology, they have a long

65 history in other sciences such as biology, geology, and paleontology. Shellfish in particular have been extensively studied for their potential as seasonality indicators, as result of their ubiquity and often exceptional preservation at coastal sites. Shellfish have been used in archaeological seasonality studies since 1969 (Weide 1969), and by ecologists and biologists since the beginning of the twentieth century (Claassen 1986:21).

Quahog seasonality studies first received widespread attention in the southeast U.S. in

1985, following the seminal publication of Quitmyer, Hale and Jones (1985a), which detailed their seasonality study at Kings Bay, Georgia (the Quitmyer et al. 1985b publication reports the same seasonality study in detail but had a more limited distribution than Quitmyer et al. 1985a).

Their work outlined the methodology for determining seasonality of quahog clams based on six phases of annual shell growth (three per each increment). To determine the seasons in which quahog gathering occurred (seasons of death), one must compare the frequencies of growth phases to that of a modern comparative collection. Quitmyer and colleagues (1985a and b) provided the data for their modern comparative collection of clams gathered from Kings Bay,

Georgia. The publication of these data allowed other researchers to conduct seasonality assessments of archaeological sites in the Georgia Bight. Since this publication, five modern comparative collections have been assembled throughout Florida: Cedar Key, two collections at

Charlotte Harbor, Indian River, and St. Augustine (this dissertation).

In addition to the incremental growth technique, archaeomalacologists have evaluated the season of death of clams and other mollusks by examining oxygen isotopic ratios present in the shell. Throughout their lives, molluscan shells record information about environmental factors such as water chemistry and temperature (Lutz and Rhoads 1980). Because shells are made up of calcium carbonate (CaCO3), oxygen isotopes (that are present in water) are incorporated into shells as the organism grows. Furthermore, quahog clam shells are known to be at or nearly at equilibrium with the water in which they live (Jones and Quitmyer 1996). By following the ontogenetic sequence of growth in clams, one is able to track changes in oxygen isotopes throughout the organism’s life. These changes in the ratios of oxygen isotopes are, to some degree, dependent on water temperature. Oxygen isotopic ratios are also influenced by changes in salinity; therefore, shells from brackish water cannot provide absolute temperatures.

66 The temperatures extrapolated from the oxygen isotope ratios of brackish species are only relative values, and cannot be considered absolute temperatures.

Oxygen has two isotopes, 16O and 18O that are present in varying amounts in the Earth’s water. 18O is the heavier isotope, and is evaporated from water less readily than 16O (Dincauze

2000:171). This means that during cooler times, 16O is removed to ice caps resulting in enriched

18O values in the water and in calcium carbonate. During warmer times, 18O is evaporated out of the water to some extent, resulting in relatively depleted 18O values. By comparing the ratio of 18O to 16O, one can determine whether the water was relatively warmer or colder. Isotopic studies of mollusks use sequential samples of calcium carbonate to identify cyclical patterns of seasons in shell growth (as evidenced by fluctuating ratios of 16O and 18O). The final sample

(taken at the terminal end of growth) is then compared to these seasonal cycles and assigned a season of death.

Isotopic study has been an important tool to establish comparability between archaeological and modern comparative collections. It is important to verify that both modern and archaeological clams experienced similar cycles in temperature (as evidenced by similar oxygen isotopic ratio cycles). By identifying similar cycles that correspond to particular growth increments, one can verify that both sets of clams are comparable. This verification is essential given the considerable time span between the lives of archaeological clams and the modern comparative clams. Isotopic studies also have been used to test the validity of the incremental growth technique by comparing isotopic analysis and incremental growth and by isotopic assay of samples with known seasons of death (e.g., Andrus and Crowe 2008; Jones and Quitmyer

1996; Parsons 2008, this dissertation). This testing, along with Jones and Quitmyer’s (1996) use of isotopic analysis to evaluate the timing of the onset of seasonal growth at different latitudes, has provided a great deal of evidence supporting the validity of the incremental growth method.

The Modern Comparative Clam Collection

The application of the incremental growth technique is limited by the necessity of a modern comparative collection for which the seasons of death are known. Although quahog clams

67 would appear to be simple invertebrates, issues of seasonality and the timing of growth cycles are rather localized and demonstrate considerable differences between populations. Using both isotopic analysis and incremental growth studies, Jones and Quitmyer (1996) and Quitmyer et al. (1997) demonstrated that there are significant variations in the onset of seasonal growth increments at different locations; these variations are particularly evident at different latitudes.

For this reason, it is imperative that the modern comparative clam collection is gathered from an area in close proximity to the archaeological site of interest and is similar in climate and ecology. Collections of at least 30 clams should be taken each month over a minimum of one year to display the range of growth throughout the calendar year (Claassen 1998). By identifying the expected growth patterns (i.e., frequencies of growth phases) for each season, the researcher can use these patterns to determine the seasonality of archaeological clams.

Two modern clam collections bracket the Guana shell ring: the Kings Bay collection

(Georgia) (Quitmyer et al. 1985a and b, 1997) and the Indian River collection (Florida) (Arnold et al. 1991; Quitmyer 1995), (these collections will be discussed in further detail in Chapter 8).

The Kings Bay collection is closer to Guana; it is approximately 85 km north of the shell ring.

The Indian River collection is considerably farther; it was obtained nearly 150 km south of the

Guana shell ring. Although the Kings Bay collection might have been sufficient for comparison with the Guana clams, I decided that gathering a modern comparative collection was necessary in order to make fine‑grained seasonality assessments at Guana.

The Guana shell ring and the broader St. Augustine area are located at a mid‑point between the tide‑dominated coastline of Georgia and the wave‑action coastline of eastern

Florida. Much of the Georgia coast is characterized as meso‑tidal (tides of 2‑4 meters), whereas most of Florida is characterized as microtidal (tides less than 2 meters) (Dincauze 2000). An examination of the 2010 tidal predictions for Kings Bay, St. Augustine (at Fort Matanzas), and

Indian River (at Cape Canaveral) shows considerable differences in the tidal characterization of each area (Table 5.1). St. Augustine also lies between two different types of marshes. Kings Bay is characterized by extensive salt marshes, while the Indian River is characterized by mangrove lagoons. The St. Augustine marshes are primarily composed of marsh grass (Spartina spp.), but also contain a considerable amount of mangrove trees. The collection in St. Augustine was

68 Table 5.1. Tidal Predictions for 2010.

This table shows tidal predictions for select dates in 2010 at Kings Bay Navy Base (KB), Fort Matanzas Inlet (FM), and Cape Canaveral (CC). Each of these sites has been the location of a modern clam collection: Kings Bay (Quitmyer et al. 1985 a and b; Quitmyer et al. 1997), St. Augustine (this dissertation), and Indian River (Arnold et al. 1991; Quitmyer 1995). Tidal data was obtained from the National Oceanic and Atmospheric Administration (NOAA) tidal predictions listed at the following website: http://tidesandcurrents.noaa.gov/tides10.

Date KB Low FM Low CC Low KB High FM High CC High 1/1/10 ‑1.6 ‑1.4 ‑1.0 8.0 4.7 4.2 ‑1.2 ‑1.0 ‑0.3 6.6 4.0 4.1 2/1/10 ‑1.7 ‑1.5 ‑1.0 7.7 4.6 4.1 ‑1.6 ‑1.4 ‑1.0 7.3 4.4 4.1 3/1/10 ‑1.5 ‑1.3 ‑1.0 7.8 4.7 4.2 ‑1.5 ‑1.3 ‑1.1 8.0 4.7 4.2 4/1/10 ‑0.6 ‑0.6 ‑0.3 7.2 4.3 3.9 ‑0.5 ‑0.5 ‑0.7 8.1 4.8 4.1 5/1/10 0.2 0.2 0.1 6.6 4.0 3.6 0.3 0.3 ‑0.3 7.6 4.5 3.8 6/1/10 0.8 0.8 0.4 7.1 4.2 3.2 1.2 1.0 0.1 6.1 3.6 3.3 7/1/10 0.7 0.7 0.3 6.6 4.0 3.2 1.3 1.1 0.3 6.1 3.6 3.3 8/1/10 0.6 0.6 0.3 6.1 3.6 3.3 1.4 1.2 0.7 6.4 3.8 3.2 9/1/10 0.7 0.7 0.6 5.9 3.5 3.4 1.4 1.2 1.3 6.9 4.1 3.6 10/1/10 0.7 0.7 0.9 6.1 3.6 3.8 1.2 1.0 1.3 7.2 4.3 4.1 11/1/10 0.3 0.3 0.9 6.8 4.0 4.2 0.0 0.0 0.6 7.1 4.2 4.3 12/1/10 0.0 0.0 0.6 7.1 4.2 4.1 ‑0.6 ‑0.6 ‑0.1 6.4 3.8 3.9

69 obtained to provide the most accurate seasonal profile for comparison with the Guana archaeological collections.

I also wanted to compare the seasonal profiles of the Kings Bay and Indian River collections with a sample collected between these locations. I hypothesized that a collection between the two extant collections could shed light on the extent of variation in seasonal onset caused by changes in latitude (and changing ecosystems). I hoped that these data would be useful for determining acceptable distances between the modern comparative collection and the archaeological site in question. The following chapter will outline the methodologies I used for collecting the modern comparative collection, evaluating isotopic composition of clam shells, and determining the season of death of clams.

70 CHAPTER 6

METHODOLOGY

Modern Clam Collection

Foraging for Clams

I gathered living quahog clams every month for one year (January through December of 2010) in order to create a modern comparative collection for St. Augustine, Florida (hereafter referred to as the St. Augustine collection). Each month I collected at least forty live clams from the

Matanzas River south of St. Augustine (just south of Fort Matanzas National Monument and

Crescent Beach) (Figure 6.1). Based on shell morphology, Irv Quitmyer suggested that several specimens from the St. Augustine collection were probably M. campechiensis (southern quahog) or hybrids (the majority of specimens appeared to be M. mercenaria). Although genetic testing of living specimens is necessary to identify Mercenaria to the species level, it is plausible that M. campechiensis and hybrids are present in St. Augustine (though they are probably far less common than M. mercenaria). Arnold and colleagues (1998) found that the incremental growth technique is effective for all types of Mercenaria; therefore, the presence of M. campechiensis and hybrids in the collection does not negatively impact the study.

Within a small area of the river, I identified multiple collection locations to ensure that I could collect enough clams from this area for an entire year. These collection locations can be grouped into two general areas (Areas A and B), which I exploited over the course of 2010

(Figure 6.2). Collection Areas A and B are approximately 1,600 m apart and each consisted of multiple collection locations.

Collection Area A (Figure 6.3) consisted of three small islands around which I collected clams. This area lies approximately 520 m north of the Whitney Lab Dock, which served as my point of departure and the location for my water temperature data logger. Collection Area A is

Figure 6.1. Quad Map of the Matanzas River. This map shows the Matanzas River and Inlet, Fort Matanzas National Monument, and the Whitney Lab Dock.

accessible solely by watercraft and is part of an aquaculture lease held by local clam farmer Phil

Cubbedge. This area is off‑limits to collecting by other people and is posted as such. The remnants of old clam bags used for housing aquaculture clams as well as pipes and other staking materials surround these small islands. I found many wild clams buried underneath and around these materials, which afforded some amount of protection from stingrays and other predators.

Collection Area B consisted of a large swath of marsh surrounded by two dune ridges

(A1A Highway lies on the western ridge and houses are built on the eastern ridge) (Figure 6.4).

Figure 6.2. Ortho Map of Collection Areas. This map depicts the location of the Whitney Lab Dock, Collection Area A, and Collection Area B. Collection Areas A and B are approximately 1,600 meters apart.

This area lies approximately 2,100 m north of the Whitney Lab Dock. Most collection occurred around the island and at the edges of the marsh. This area has silted in considerably in recent years and during my collections, I observed more and more dry land with each collection.

Figure 6.5 shows the sedimentation of the marsh over time. This sedimentation is likely the result of construction along the marsh edges and Matanzas River and a lack of preventative mitigation. Collection Area B is accessible by walking into the marsh from A1A Highway, and is open to commercial fisherman. During the fall of 2010, a commercial fisherman overharvested the clam population in a particular portion of Area B (see Figure 6.4). After this

Figure 6.3. Ortho Map of Collection Area A. This map shows the location of the Whitney Lab Dock and Collection Area A. Collections were made around the three small islands highlighted in the purple circle.

occurred, it was much more difficult to find clams in this location, and many of the remaining clams were small, visibly injured, or hiding beneath oysters (which are difficult to rake through and are often overlooked by commercial fishermen).

During most months of my collection Phil Cubbedge, guided me into the marsh. Mr.

Cubbedge grew up collecting and selling wild clams in the area (before the aquaculture boom) and was an invaluable resource during collection of the samples. I relied heavily on his knowledge of the marsh and of clams in particular. He was extremely skilled at finding wild clams, and found up to 100 clams in less than 20 minutes. Phil also inspected his clam farm,

Figure 6.4. Ortho Map of Collection Area B. This map shows the location of Collection Area B, at the southern border of Summerhaven, Florida. Collection Area B is depicted in the purple rectangle. Clams were recovered around the island from tidal creeks and at the edges of the marsh (towards the outer edges of the rectangle). This map also shows the location of an area that was overharvested by commercial fishermen during 2010. The Atlantic Ocean is to the right of the collection area, and the Matanzas River is to the left of the collection.

commercially gathered wild oysters, and caught fish for his own consumption during our trips into the marsh.

During my collections, I attempted to gather a wide range of sizes and ages of clams. To accurately identify seasonal growth clams must be at least two years old (in order to make predictions about how much growth should be expected in the terminal increment). Very old clams can also be problematic because their growth increments are very close together and fast

Figure 6.5. Satellite Imagery of Collection Area B Sedimentation. These images show the gradual infilling of the marsh. Over the course of my collections in 2010 I observed considerable sedimentation in the marsh, resulting in more dry land and changing habitats for sea life. Starting in the upper left corner and moving clockwise, the images were taken in 1995, 1999, 2005, and 2010.

(opaque) growth becomes much smaller and occasionally ceases in old age. For these reasons clams that were very small or appeared senescent were discarded in the marsh. I gathered clams using two methods: raking in the sand with a short‑pronged rake and signing. Signing refers to detecting the location of a clam by identifying physical signs of its presence. These clam signs include small patches of waste that the clam has expelled, circular patches of dark sand, and keyholes. Clam waste looks like small patches of sunken sediment localized in a small area (often 5 cm2 or less). Circular patches of dark sand occur when clams burrow into deeper, darker sediments and displace this sediment on top of lighter top sediments. These circular patches were rarely observed during my trips to the marsh, but when present they were helpful in locating clams (although other animals can also create these circular patches).

Figure 6.6 Clam Keyhole and Clam. This image shows a clam keyhole and a clam shell approximately the size of the clam buried underneath the keyhole.

Figure 6.7. Clam Keyhole. This image shows a close‑up of a clam keyhole with a clam inside. A small amount of waste is also visible to the right of the keyhole.

Figure 6.8. Alternative Shape Keyhole. This image shows a keyhole that appears as two connected holes. I observed this kind of keyhole several times during the winter. The clam inside the keyhole was approximately the size of the clam shown to the right of the keyhole.

Keyholes, the most commonly identified clam sign, are small keyhole‑shaped holes in the sand through which the clam siphons water and expels waste (Figures 6.6 – 6.8). Keyholes, when present, were the easiest type of clam sign to identify.

To the novice collector it is extremely difficult to differentiate clam waste patches and keyholes from other animals’ waste and holes. I poked my finger into countless empty holes hoping find a clam (and was lucky not to find a crab or other critter). However, the more time I spent in the marsh, the better I became at identifying clam signs. By the end of the year, I was quite adept at locating clams based on signs, although I was nowhere as skilled as my mentor,

Mr. Cubbedge. In order to observe these clam signs, one must be in relatively clear, shallow water (or on temporarily dry sandbars). It is likely that Native American foragers knew these signs as well, since they were intimately familiar with their environments. Identifying clam signs (especially keyholes) allows a forager (modern or prehistoric) to gather clams much faster than by raking or treading (wading in the water and feeling for clams with the feet). Finding

78 clams by signing significantly decreases search time, making clam collection a more efficient foraging strategy. Clam signs will be discussed further in Chapter 8.

It is important to note that I was not the only predator on the clam beds. During this time, crown conchs, whelks, stingrays, fish, birds, crabs, and other people were also preying on the clams. The crown conch, Melongena corona, was the most frequently observed predator of clams. Crown conchs are not native to this particular area of Florida. Sometime in the last decade they appeared in the marsh to the south of the Fort Matanzas Inlet (Phil Cubbedge, personal communication 2010). Crown conchs have a long radula that they use to chip the margin of the clam until they are able to insert the radula into the shell and consume the clam.

Crown conchs are very numerous in the southern marshes of the Matanzas River and often congregate on clam beds to consume the clams. Many times I observed crown conchs consuming live clams, occasionally as many as three conchs were consuming a single clam.

Crown conchs cause significant problems for clam aquaculturalists because they can extend their radula into the protective nets that surround aquaculture clams. This has resulted in significant loss in aquaculture clam yields.

Water Temperature and Salinity in the Matanzas River

Two primary factors influence the growth and isotopic composition of clam shells: water temperature and salinity. In order to examine the relationship between clam growth phases and water temperature, I deployed a water temperature data logger. I used an Onset Hoboware logger along with a Hoboware shuttle to take water temperature measurements every half hour.

The logger was deployed at the Whitney Marine Lab dock, which lies approximately 520 m south of the Collection Area A, and about 2,100 m south of the Collection Area B. The logger was deployed in September of 2009 and remained in use until the conclusion of the collection on December 31, 2010. To obtain temperatures relevant to the position of clams, I attached sinkers to the logger so that it would remain near the sandy bottom of the Matanzas River. The

Whitney dock is in fairly shallow water and during low tide the logger sat on or nearly on the sandy river bottom. During high tide the logger remained in the middle of the water column, but never at the surface. This orientation of the logger provided temperature data that were

79 more similar to temperatures experienced by clams living in the substrate, rather than providing surface water temperatures. Measurements were taken every half hour throughout the entire year of collection.

To evaluate the fluctuations in salinity in the collection area, I relied upon the data collected by the National Estuarine Research Reserve System (NERRS), which is partnership between the National Oceanic and Atmospheric Administration (NOAA) and coastal states

(data is distributed by the Centralized Data Management Office, or CDMO). The NERRS has a water station just north of Fort Matanzas (upriver), located to the north of the clam collection areas (see Figure 6.1). Fort Matanzas is approximately 5.5 km north of my logger at the

Whitney Dock. The water station takes readings every 15 minutes on environmental factors including salinity levels. The NERRS conducts research under an award for the Estuarine

Reserves Division, Office of Ocean and Coastal Resource Management, National Ocean Service,

National Oceanic and Atmospheric Administration, and makes their data available on the

Internet to students, researchers, educators, and others who are studying estuaries. For more information, please consult the CDMO website: http://cdmo.baruch.sc.edu/about/overview.cfm.

Preparing the Clams for Analysis

After I collected the living clams, I stored them on ice or in the refrigerator until I was ready to process them. I processed the clams within a few days of collection. Processing involved boiling or steaming the clams until they popped open or until they were sufficiently heated

(some clams did not pop open and had to be gently pried open). Boiling and steaming the clams allowed me to easily remove the meat without damaging the shells. Once the shells were open, I removed the meat and scrubbed the shells. Once they had dried, I assigned a field specimen (FS) number (one for each month) and a clam number (unique in each FS) to each clam. I then labeled the clams with a permanent marker. After the clams were labeled, I broke the valves apart (they were still attached at the umbo by the elastic hinge ligament).

In the first monthly collection the right valves were in better condition than the left valves (fewer injuries and less chipping during storage) so I decided to use the right valves for the analysis. I measured the length and height for all shells utilized in the study with digital

80 calipers (Figures 6.9 and 6.10). For many archaeological clams the length and height could not be determined due to breakage. The umbo is the strongest part of the clam, and is often preserved relatively intact compared to other fragments. To obtain measurements of numerous clam fragments, I created a hinge length measurement, which I defined as the total length of the hinged portion of the clam (from the anterior lunule notch to the posterior end of the portion covered by the elastic hinge ligament) (Figure 6.11). This hinge length measurement permitted measurements to be made for many clams that otherwise could not have been measured. It was necessary to take all measurements before cross‑sectioning the clams because the saw blade removes a small portion of the shell (on either side of the blade), thus compromising the length and hinge length. I recorded all measurements in an Excel database. At this time I also noted any irregularities in the shell, including interior texturing and/or small bumps, predation marks, injuries to the shell, and oddities in coloration. During this examination, I observed whether clams had interior purple coloration along the margin. Extensive purple coloration can obscure growth increments near the margin of the shell and is undesirable in clam seasonality

Figure 6.9. Shell Length Measurement. This figure indicates how shell length was measured. Measurements were recorded as the longest length from the posterior to anterior margins of the shell.

Figure 6.10. Shell Height Measurement. This figure depicts how shell height was measured. Shell height was recorded as the longest height from umbo to the margin.

Figure 6.11. Hinge Length Measurement. This image shows the hinge length measurement. This measurement evaluates the total length of the hinged portion of the clam and extends from the anterior lunule notch to the end of the portion covered by the elastic hinge ligament (visible as the end of a long notch or channel towards the posterior shell.

Figure 6.12. Minimum Usable Valve Fragment. The shaded area indicates the minimum portion of the valve necessary for inclusion in the study. Using posterior margin fragments with a portion of the adductor scar ensured that a valve was sampled only once.

studies. In order to avoid this problem when possible, I selected with clams with the moderate to no purple coloration for the sample. For each month in the collection, I selected and analyzed at least 30 valves, resulting in a total sample of 411 valves.

The Archaeological Clam Sample

I prepared the archaeological clams from the Guana shell ring in a similar manner as the modern clams. To be sure that I analyzed a clam only once, I used left valves (which were more numerous than right valves). Further, I used only whole valves or valve fragments that contained the posterior margin and adductor scar (this ensured that I did not use different fragments of the same clam) (Figure 6.12). I identified and recorded all usable left valves from both archaeological collections. I defined a valve as usable if it contained at least the posterior margin and adductor scar and had an intact margin free of chipping and extensive wear. Based on extreme predation and shells that were covered in oyster spat or barnacles, it appeared that some valves were collected after their death (Figure 6.13). These clams may have been attached to living oysters (that were the target of collection), picked up by inexperienced collectors, such

Figure 6.13. Dead Clam Shell. This photograph depicts an archaeological clam from Guana that was collected after it had died. Excessive predation by boring organisms (some bore holes penetrate the shell) along with damage from oyster spat and perhaps barnacles indicate that this clam was not alive when it was collected and brought to the Guana shell ring. This clam was not used for seasonality assessment because it was not collected alive.

as children, or were gathered by mistake. Valves that were collected after their death were excluded from the study.

From the ring itself, I analyzed all usable left valves in the sample. The ring sample consisted of clams from a single test unit and ten shovel tests placed throughout the ring (three shovel tests did not provide any clams for analysis). Because the sample came from multiple locations on the ring, I felt it was necessary to take as large a sample as possible from each location. I analyzed and cross‑sectioned 241 valves from the ring (described below).

The sample from the interior of the ring originated from features identified in four test units. Saunders and Rolland (2006) identified much of the excavated matrix as features or

84 areas; as a result the majority of analyzed valves were removed from features, rather than general levels. In general, the clams from the ring interior were not preserved as well as those from the ring deposit. Many of the ring interior clams exhibited bleaching on the margin and interior umbo (probably from post‑depositional processes), and as a consequence were not usable for seasonality assessment. I analyzed and cross‑sectioned all usable valves from the ring interior; this sample consisted of 155 clams.

I labeled all usable archaeological clams with the FS number assigned during excavation, or I created an FS number if one was not previously assigned. I also assigned each valve a clam number (unique within each FS) and labeled the valves with this number.

Labeling was done with a permanent marker so that provenience information would not be lost during cross‑sectioning. Although some researchers have used pencil for labeling, I found that pencil occasionally washed off during cross‑sectioning. As with the modern clams, I measured the length, height, and hinge length for all archaeological clams prior to cross‑sectioning.

During this time I noted any irregularities of the shell, including predation marks, injuries to the shell, presence of bumps or texturing on the interior of the shell, oddities in coloration, and whether purple coloring was present in the interior of the valve. Based on shell morphology,

Irv Quitmyer suggested that several of the archaeological clams from Guana might be M. campechiensis or hybrids. The vast majority of clams, however, appeared to be M. mercenaria.

Cross‑Sectioning the Clams

In order to the view the internal growth structures of clams, one must cross‑section the clams.

The internal growth structures of clams are best observed along the axis of greatest shell height.

At the greatest shell height the growth lines are semi‑concentric and ontogenetic (beginning at the umbo and terminating at the margin) and they appear relatively straight in cross section. In other parts of the shell the growth rings slope in such a way that makes growth assessments more difficult (especially comparing the size difference between growth increments). To evaluate incremental growth structures, one can create thin sections, which must be viewed using transmitted light, or one simply saws the clams in half (thick sectioning). O’Brien and

85 Thomas (2008) evaluated the efficacy of both methods (using blind sampling of clams with known collection times). They found that although thin sections provided slightly more accurate results than thick sections (90 percent compared to 75 percent), the use of thick sections still preserved the incremental growth trends throughout the annual cycle resulting in correct interpretations of seasons of death (O’Brien and Thomas 2008:485). Thick sections require far less time to create and analyze, which allows researchers to examine larger samples in less time.

For this reason, I employed thick sectioning.

I used two types of saws to section the clams. I cross‑sectioned approximately 25 percent of the modern collection and most of the archaeological clams from the ring (excluding the ring interior) using a Ryobi seven inch wet tile saw equipped with a diamond blade. The diamond saw blade was specifically designed to cut stone and porcelain as well as other types of ceramic tiles. I used the thinnest blade available in an attempt to get the best cut. I experimented with a variety of methods to cut the clams to determine a strategy that would do the least damage to the shell. I tried cutting the shell from the umbo to the margin, from the margin to the umbo, and I also tried cutting halfway from the umbo to the margin and then turning it to complete the cut from the margin to the umbo. The saw was more likely to break up the shell at the distal end of the cut, so I sectioned most of the clams from the umbo to the margin. After cross‑sectioning, I polished some of these clams on a lapidary wheel using coarse grit and/or sanded them on 400 grade wet sandpaper. This increased the visibility of the increments and facilitated seasonal assessments.

At the suggestion of Irv Quitmyer, I began using a lapidary saw equipped with a Mark

V carbon blade (six inches long and 0.15 inches thick). I used the lapidary saw to cut 75 percent of the modern clams, nearly all of the clams from the ring interior, and several clams from the ring deposit. I sectioned clams from the margin to the umbo because the umbo is the most difficult portion to cut (it is thicker and denser than the rest of the shell). Although the carbon blade makes an extremely clean cut, there are some drawbacks that should be considered. The lapidary saw is more expensive than a simple tile saw. Also, the carbon blades are extremely thin and can break easily if the clam is turned even the slightest bit during the cut. For this reason it takes nearly twice as long to cut each shell. I broke three blades while learning how to

86 cut with the carbon blade, but once I became proficient I cut nearly 400 clams with one blade.

When using the carbon blades it is important that the saw has a shield or cover to protect the analyst from pieces of shattered blades. The carbon blades provide excellent cuts for clams in good condition, but clams that are fragile, falling apart, or are missing pieces of the umbo are another matter. Clams that brittle are likely to break the carbon blades, so I cut these clams on a lapidary saw equipped with a thin metal lapidary blade or on the tile saw. The metal lapidary blade (used on a lapidary saw) provided a cleaner, thinner cut than the tile saw and the metal blade did not break like the carbon blade. However, after around 200 cuts the blade became dull and needed to be replaced. The cut provided by the carbon blades is far superior to that of even the thinnest tile saw blade and the metal lapidary blades. The carbon blades polish while they are cutting, creating a smooth and polished cross‑section. This superior cut makes the evaluation of age and seasonality a much easier task. One should consider both the pros and cons of each technique carefully before determining which machinery to employ.

Analyzing the Clams

Macroscopic Analysis

All clams were analyzed macroscopically. Clams that were particularly difficult to assess were examined under 5‑10X magnification when necessary. All clams were examined under the same lighting in the same room to avoid changes in ambient light that might negatively impact the analysis. In some cases, a penlight was used to analyze clams under magnification.

I recorded three variables from each of the sectioned clams: 1) age, 2) growth increment, and 3) growth phase. Age was determined by counting sets of opaque and translucent increments (one opaque and one translucent equals one year of growth) from the umbo to the margin of the clam. Shell fragments that were missing part of the ontogenetic sequence were not assigned an age or age estimate. Growth increments were assigned based on the terminal increment of the clam. Growth phases were determined following the characterizations defined by Quitmyer and colleagues (1985a and b) with some slight deviation. Quitmyer and colleagues

(1985a) expected a fully‑grown increment to equal or surpass the size of the same increment of

87 Table 6.1. Growth Phase Classifications. This table shows the characteristics I used to identify each growth phase. The growth phases are depicted for both the umbo and the margin.

Growth Phase Umbo Margin Description

Opaque just O1 developing to 1/3 complete

Opaque more than O2 1/3 to 2/3 complete

Opaque more than O3 2/3 to fully complete

Translucent just T1 developing to 1/3 complete

Translucent more T2 than 1/3 to 2/3 complete

Translucent more T3 than 2/3 to fully complete

88 the previous year. For example, they defined the O3 phase as an “opaque growth phase greater than or equal to the size of the previous opaque increment” (Quitmyer et al. 1985a:31).

In the St. Augustine collection and archaeological clams, I observed a slightly different pattern: after the first three or four years of life the size of the opaque increment rarely surpassed that of the previous year (and was usually smaller than the previous year). Quitmyer and Jones

(1992:256) acknowledged this decrease in the size of increments as a result of age, but did not state how they accounted for this narrowing when assessing seasonality (they indicated that the narrow growth increments observed in older clams were the primary cause for inability to characterize seasonality).

The categorization criteria I used are the same for each increment (opaque and translucent) and are as follows: 1) increment just developing to 1/3 complete, 2) 1/3 to 2/3 complete, and 3) 2/3 to complete (Table 6.1). I consider an increment to be complete or fully developed if it has reached its full growth potential based on the growth of previous years. In my classification, the size of a fully developed growth increment can be larger than, equal to, or smaller than the same increment of the preceding year. Russo and colleagues (1993) and Russo

(1998) also categorized growth phases by thirds, which defines equal amounts of growth for each phase, rather than designating the second phase solely by one‑half of the expected growth.

Although I suspect that Quitmyer and colleagues (1985a and b) and others probably identified the second growth phases when growth was slightly less or slightly more than one half complete, this has not always been clearly indicated in the literature. Table 6.2 compares the arbitrary six‑part divisions defined and Quitmyer et al. (1985a and b) and those used for this dissertation.

While determining growth phase, I evaluated several years of growth prior to the terminal growth increment. As clams grow older, their growth increments become progressively smaller; therefore it is imperative to consider growth trends over multiple years of growth (Quitmyer and Jones 1992:251). Typically, clams in both the archaeological and St.

Augustine collections exhibited this decline in overall shell growth. I observed that after several years of growth, most clams exhibited an approximate 33 percent decline in incremental growth

89 Table 6.2. Growth Phase Characterizations. This table describes the characterizations of growth phases made by Quitmyer et al. (1985a and b) and those used for this dissertation.

Growth Phase Quitmyer et al. (1985a and b) This Dissertation O1 Opaque increment just forming to just Opaque increment just developing to under 1/2 developed 1/3 complete O2 Opaque increment approximately 1/2 Opaque increment more than 1/3 to complete 2/3 complete O3 Opaque increment greater than or Opaque increment more than 2/3 to equal to previous opaque increment complete T1 Translucent increment just forming to Translucent increment just just under 1/2 developed developing to 1/3 complete T2 Translucent increment approximately Translucent increment more than 1/3 1/2 complete to 2/3 complete T3 Translucent increment greater than or Translucent increment more than 2/3 equal to previous translucent to fully complete increment

from the previous year. This decline was evaluated macroscopically for each clam and used to determine the amount of expected growth for the terminal increment. In some cases, translucent increments became bigger and this was taken into consideration. Although most clams displayed the approximated 33 percent decline in shell growth from one year to the next, the overall growth trends for each clam were evaluated and considered when projecting the amount of expected terminal growth.

As clams grow older, growth slows and increments become smaller. Clams eventually experience more suboptimal than optimal growth, which results in shells that are characterized by more translucent (suboptimal) growth than opaque (optimal) growth. In clam seasonality literature these clams are referred to as senescent or senile (Claassen 1998; Clark 1979; O’Brien and Thomas 2008; Parsons 2008, n.d.; Quitmyer and Jones 1992), or geriatric (Andrus and

Crowe 2008; Thompson and Andrus 2011). It should be noted that in biological literature the term senescence is generally used to indicate reproductive senescence (i.e., a slowing or cessation of reproduction). For this study, I defined senescent clams as those that demonstrated more translucent than opaque growth toward the margin (i.e., smaller opaque increments than

90 translucent increments or no visible opaque increments). Determining senescence is important for seasonality studies because these clams often do not follow predictable (i.e., seasonal) patterns of growth (Andrus and Crowe 2008; Claassen 1998). For this reason senescent clams should be avoided in seasonality assessments when possible (but they may provide important age data; therefore they should be evaluated to verify senescence and to obtain age data).

In order to be sure that the assessments were accurate and precise, I examined each clam

(modern and archaeological) on at least two separate occasions. The second examination was a blind assessment. In other words, I made the second assessment without knowing the outcome of the first assessment. If the assessments were identical, the assessment was recorded as final.

If the assessments of age, increment, or growth phase were not identical, I examined the clam for a third time (without knowing the assessments of the first or second examinations). I then compared the assessments and examined the clam a final time. If I could not make a confident assessment, I indicated that it was not possible to accurately determine the specific variable(s).

I divided the traditional four seasons in the same way as Quitmyer and Jones (1992): winter (December, January, February), spring (March, April, and May), summer (June, July,

August), and fall (September, October, November). Clams from the St. Augustine collection were categorized into both monthly and seasonal samples. These samples provided distributions of growth phases that can be compared with other modern collections and can be used to determine the season of death for archaeological clams in the St. Augustine area.

Distributions of growth phases from the Guana shell ring site were compared to the distributions of growth phases for the St. Augustine collection in order to identify seasons of death for archaeological clams.

Isotopic Analysis

As discussed in Chapter 5, oxygen isotopic analysis can be used to verify the comparability of modern and archaeological clam collections and to evaluate relative temperature fluctuations throughout the life of a single clam. I chose three modern clams and four archaeological clams to sample for isotopic analysis. I conducted this analysis to verify that the incremental methodology was valid for this study and that both archaeological and modern clams were

91 comparable (despite any climate change over the last 4,000 years). In addition to the overall goal of determining comparability, each clam was chosen in order to answer a specific question.

These questions are as follows: 1) What does the overall relative temperature profile look like throughout the clam’s life? 2) Do multiple samples within a single increment follow the expected temperature curve? 3) Are translucent lines that are very close together “double lines”

(i.e., are these cases where clams experienced more than one translucent increment in a single year)? 4) Does the light gray area between these possible “double lines” indicate a true opaque increment, or is it a less defined part of a single translucent increment? I attempted to answer these questions for both modern and archaeological clams in order to confirm that the patterns

(even the unusual ones) were similar in archaeological and modern clams. Also, I hoped that a better understanding of unusual increments would increase the number of archaeological clams that I could interpret.

Each clam was cross‑sectioned and examined for suitability for isotopic testing. I selected clams that were at least 4 years old but were not senescent (primarily exhibiting translucent growth towards the margin). I selected three clams (one modern, two archaeological) that had well‑defined normal increments and normal growth patterns. These clams were selected to examine typical temperature profiles. I selected one atypical and one slightly atypical modern clam to sample, and two atypical archaeological clams. I drilled small, round holes on the cross‑sectioned surface of the middle layer of the clam to remove approximately 1 gram (g) of sample material. I used a small, flat‑tapered dental burr, which allowed me to verify that sample material was removed from a single increment (it is important that one does not accidentally drill into an adjacent increment). Samples were taken in order of ontogenetic growth over at least two years of consecutive growth. In total, 65 samples were taken: 29 from archaeological clams and 36 from modern clams. The locations of each sample will be discussed along with the results in Chapter 8.

The Stable Isotope Mass Spectrometry Laboratory, a part of the Department of

Geological Sciences at University of Florida, analyzed the samples. Samples were reacted in 100 percent orthophosphoric acid at 70° C using a Finnigan‑MAT Kiel III carbonate preparation device. Evolved CO2 gas was measured online with a Finnigan‑MAT 252 mass spectrometer.

92 Isotopic results are reported in standard delta notation relative to Vienna Pee Dee Belemnite

(VPDB). Analytical precision is estimated (1 standard deviation of standards analyzed with samples) to be +0.056 ‰ (per mil) for δ18O (n=26).

93 CHAPTER 7

RESULTS

Isotopic Analysis

I evaluated the oxygen isotopic composition of three modern clams from the St. Augustine collection and four archaeological clams from the Guana shell ring. This analysis was undertaken to verify that both archaeological and modern clams were comparable – meaning that they both experienced opaque and translucent growth during similar temperatures (or seasons). This was done to ensure that the modern St. Augustine collection is a valid baseline with which to determine the seasons of clam collection at Guana, despite any climate changes that have occurred over the past 4,000 years. The oxygen isotopic profiles of both archaeological and modern clams should display an annual temperature cycle within a year of growth. In Florida, opaque increments should be deposited in cooler temperatures (Jones and

Quitmyer 1996) and should have relatively high (often positive) ∂18O values. Translucent increments should be deposited in warmer temperatures (Jones and Quitmyer 1996) and should have relatively low (often negative) ∂18O values.

By evaluating the oxygen isotopic composition of sequential microsamples of Mercenaria spp., Jones and Quitmyer (1996) identified a cyclical pattern of isotopic values that reflected annual temperature cycles. Sequential samples of the opaque increment typically displayed a pattern of a decrease and then a rise in ∂18O values, interpreted as a decrease and then a rise in relative temperatures (Jones and Quitmyer 1996). Sequential samples of the translucent increment did not display a clear pattern of cyclical ∂18O values (i.e., relative temperatures) but instead varied by shell (Jones and Quitmyer 1996). I employed a slightly different drilling methodology, and as a result I was able to collect a maximum of three samples from a single increment (opaque and translucent) (Jones and Quitmyer 1996 were able to collect as many as

94 eight samples from a single increment). Although I was able to observe changes in ∂18O values in multiple samples from single increments, these patterns were slightly different from the observations of Jones and Quitmyer (1996). This is likely a result of the different drilling methodology I employed and my inability to evaluate the same number of samples for single increments. Below I provide interpretations of the patterns I observed in ∂18O values for single increments. The reader should be aware that these interpretations for single increments are different from conventional interpretations on cyclical ∂18O values as a result of having fewer samples per increment than were available to Jones and Quitmyer (1996). All other results of the isotopic analysis follow conventional interpretations.

Based on isotopic analysis of Mercenaria spp. conducted for a previous study (Parsons

2008) and a combination of average water temperatures (in 2010) and the growth phase distribution from the St. Augustine collection (discussed below), I expected an opaque increment to demonstrate progressively warmer relative temperatures (lower ∂18O) as the increment grew. In other words, the initiation of the opaque increment would demonstrate relatively cool temperatures, the middle of the opaque increment would demonstrate slightly warmer temperatures, and the terminal portion of the opaque increment would demonstrate the highest relative temperature. Based on the work of Jones and Quitmyer (1996) and the water temperatures and growth phase distribution in the St. Augustine collection, I expected translucent increment relative temperatures to vary by shell but exhibit predominantly negative

∂18O values. The T1 growth phase occurs almost equally in two seasons and occurs in a large range of warm temperatures. The T3 phase is the most commonly identified growth phase and occurs over a large range of temperatures, including cooling fall temperatures. Because the translucent growth phases occur over a range of warm temperatures and during cooling fall temperatures, identifying translucent growth phases through isotopic values (and relative temperatures) is a difficult endeavor.

I sampled all clams in ontogenetic order (from youngest to oldest) over multiple years of growth. Appendix B (B.1‑B.7) depicts cross‑sections of the modern and archaeological clams and the location for each sample I analyzed; it also provides the ∂18O value for each sample I assessed. I selected modern and archaeological clams for isotopic analysis to evaluate specific

95 research questions. I will provide the results for each of the modern and archaeological clams and evaluate the overall patterns in each collection. For each collection, the clam I present first is the one with the most straightforward growth (this provides a sort of baseline isotopic profile with which to compare more atypical growth profiles). In this chapter, I provide a discussion of the isotopic results because they were used to aid in the identification of growth phases for both the St. Augustine and Guana clams.

Modern Clam Number Four

Modern Clam Number Four was collected on January 28, 2010 from collection Area A. This clam was chosen to evaluate the oxygen isotopic profile of a straightforward growth profile.

Clam Number Four was five and a half years old and was sampled over four and a half years of growth in each alternating growth increment. I removed samples from every year except the first year of growth (both opaque and translucent increments). Each increment was drilled approximately in the middle of the increment, except for the sample at the terminal margin that was drilled as close to the margin as possible (to encompass the most recent growth). I determined that this clam was in the O2 growth phase at the time of its death.

The ∂18O profile (Figure 7.1) demonstrates that opaque increments are deposited in cooler temperatures (more positive values) and that translucent increments are deposited in warmer temperatures (more negative values). All samples fell within the expected range of

∂18O values. The terminal increment ∂18O value is consistent with a clam in the O2 growth phase at the time of its death.

Modern Clam Number Three

Modern Clam Number Three was collected on February 26, 2010 from Collection Area B. Clam

Number Three appeared to be five years old and I sampled years three and four (I did not sample the terminal years of growth). This clam was chosen because it did not have a straightforward seasonal growth profile. This clam appeared to have a small second translucent increment, meaning that more than one translucent increment formed during a single year. Multiple translucent and opaque increments have been observed in modern clams

96 Modern Clam 4

1.00

0.50

0.00 O

18 ‑0.50 ∂

‑1.00

‑1.50

‑2.00

Figure 7.1. Profile of ∂18O Vales for Modern Clam Number 4. This graph depicts the ∂18O profile for a clam collected in January, 2010. The gray squares indicate samples taken from opaque increments and red squares indicate samples taken from translucent increments. Higher ∂18O values indicate relatively cooler temperatures, and lower ∂18O values indicate relatively higher temperatures.

Modern Clam 3

1.00

0.50

0.00 O

18 ‑0.50 ∂

‑1.00

‑1.50

‑2.00

Figure 7.2. Profile of ∂18O Vales for Modern Clam Number 3. This graph depicts the ∂18O profile for a clam collected in February, 2010. The gray squares indicate samples taken from opaque increments and red squares indicate samples taken from translucent increments. The turquoise square indicates a sample that appeared to be translucent but is actually part of the preceding opaque increment.

97 in the Narragansett Bay, Rhode Island (Henry and Cerrato 2007), but in the southeastern U.S only single opaque and translucent increments are known to occur in one year.

The first year of sampling (samples 1 and 2) was straightforward, but samples 3 and 4 were unclear and may represent a double translucent increment. Sample 3 was taken from what appeared to be a very small opaque increment, and sample 4 was taken from a light gray area before a well‑defined translucent increment (sample 5). Sample 6 was taken from the center of an opaque increment, and sample 7 was taken from what was presumed to be the middle of a translucent increment (this increment had two dark gray lines, one at each exterior, and the central portion was a lighter gray). I did not sample the terminal year of growth for this clam.

Figure 7.2 displays the ∂18O profile for Modern Clam Number Three. The first two samples clearly fall in the expected ranges for an opaque and translucent increment. Based on the ∂18O value, sample 3 appears to be a true opaque increment, despite its diminutive size.

This means that this clam is six years old. Sample 4 (the thin light gray band), however, does not appear to be a true translucent increment. Rather, Sample 4 is actually a part of the opaque increment (from which sample 3 is derived) and is probably the result of a temporary stressor to the clam. (This also explains the small appearance of the opaque band.) These gray bands, referred to as false increments or double bands, are a result of the clam remaining closed (and not feeding) for a period of time (Claassen 1998:153), and can be caused by heavy storms, severe weather, injury or illness of the clam, drastic changes in salinity, or other temporary environmental factors. Sample 4 originates from the terminal end of the opaque increment, essentially the O3 growth phase. The relatively higher temperature of Sample 4 (compared to sample 3) is expected given its location at the termination of the opaque growth increment

(when temperatures are rising). Sample 5 is the actual translucent increment, and its ∂18O value is in the expected range. Samples 6 and 7 represent actual opaque and translucent increments, respectively, and both fall within the expected ∂18O range.

98 Modern Clam (11)

1.50

1.00

0.50

0.00 O 18 ∂ ‑0.50

‑1.00

‑1.50

‑2.00

Figure 7.3. Profile of ∂18O Vales for Modern Clam Number 11. This graph depicts the ∂18O profile for a clam collected in August, 2010 (part of the July collection). The gray squares indicate samples taken from opaque increments and red squares indicate samples taken from translucent increments. The turquoise square indicates a sample that appeared to be opaque but is actually part of the preceding and following translucent bands that make up a single translucent increment.

Modern Clam Number 11

Modern Clam Number 11 was collected on August 1, 2010 (for the July collection) from collection Area B. Clam Number 11 was a long‑lived specimen, (eleven years old) and was sampled across 9.5 years of growth to evaluate the ∂18O cycle over a long period of time in a clam with a fairly straightforward growth profile. Because the clam was quite old, the increments were close together and sampling the desired increment was somewhat difficult. I drilled a total of twenty samples from Clam Number 11, from the second year of life until the final opaque increment. Although the majority of the growth profile was straightforward, samples 3 and 4 originated from increments that appeared irregular. According to my lab notes, I drilled sample 3 from “a very tiny dark opaque band between two larger translucent bands.” I drilled sample 4 from “fairly thin translucent band.” It was possible that this

99 grouping of a small opaque band (sample 3) and a small translucent band (sample 4) represented a false translucent increment (this would make Clam Number 11 ten years old instead of eleven). Samples 10 and 11 were both drilled from what appeared to be a translucent increment that was rather large and had dark lines on both exterior edges. Two samples were removed from this band to verify that it was a continuous translucent increment. Although some of other opaque and translucent bands were small, it was clear that they were true

(complete) increments. In other words, all other samples were straightforward and appeared to come from true increments. I determined that the terminal growth phase for this clam was T1.

Because the T1 growth phase was so narrow, I was unable to remove a sample from the terminal increment.

The ∂18O profile for Modern Clam Number 11 (Figure 7.3) reveals that sample 3 is probably not a true opaque increment. It appears that sample 3 was drilled from an anomalous portion of opaque‑appearing growth and that samples 2 and 4 are part of the same translucent increment (and sample 3 is part of this translucent increment). This false opaque increment is probably similar to the false translucent increments or double translucent bands that occur in clams experiencing unusual ecological or biological occurrences. Although this opaque band is unusual, it indicates that this clam did not deposit more than two standard increments in one year, but rather experienced some unusual event that briefly initiated opaque‑appearing growth during the translucent increment. This means that this clam is ten years old (not eleven) and that samples were taken over 8.5 years of growth (not 9.5 years). The relative temperature for this single translucent increment appears to increase through time and does not show a decrease in temperature towards the terminal (T3) portion.

The ∂18O values for samples 1 and 7 (drilled from opaque increments) are somewhat lower than the other opaque values, but both samples appear to be drilled from true opaque increments and are within the minimum range of opaque ∂18O values of other modern and archaeological clams. The remaining ∂18O values for Modern Clam 11 fall within the expected ranges. The ∂18O values for samples 10 and 11 indicate that they are indeed from the same continuous translucent increment. The slightly higher (more positive) ∂18O value of sample 11

(compared to sample 10) indicates a slightly cooler relative temperature (the variation between

100 the two samples is 0.04 per mil). Sample 11 was removed approximately from the T3 growth phase near the termination of the translucent increment. The relatively cooler temperatures indicated by the ∂18O value may indicate that this T3 phase was formed when temperatures were cooling in the early fall.

In sum, sample 3 is an anomalous opaque band and not a true increment, thus samples 2 and 4 are part of the same translucent increment. This means that sample 3 is a false opaque increment, and indicates that no additional translucent increment was deposited during that year of growth (as seen in Narragansett Bay clams). This also means that this clam is ten years old, not eleven years old. As expected, samples 10 and 11 originated from the same translucent increment. With the exception of the anomalous opaque band, all other ∂18O values are in the expected range.

Summary of ∂18O Isotopic Profiles for Modern Clams from the St. Augustine Collection

The modern clams indicate that a single opaque increment and a single translucent increment form per year in Mercenaria spp. from St. Augustine. The opaque increment occurs in relatively cool temperatures (∂18O values ranging from 1.0 to ‑0.4) and the translucent increment occurs in relatively warm temperatures (∂18O values ranging from ‑0.5 to ‑2.0). Although in the

Narragansett Bay clams (Henry and Cerrato 2007) have been observed to form multiple translucent increments in a single year, this was not observed in any case from the St. Augustine collection. This means that the appearance of a second translucent band is instead a false translucent increment, or double band (as is the case for Modern Clam 3 sample 4). A false increment is an additional band within a single year of growth that is brought on by erratic environmental or biological stressors to the clam, and not by seasonal increment‑inducing temperatures. These false translucent increments are relatively common in clam assemblages and have been described by Claassen (1998:153; 158‑159; 163‑164). When mistaken for true increments, false increments can give the appearance of an additional year of growth that is not present and can confound increment and growth phase identification. These false increments can usually be identified by their irregular size when compared to other increments (they are often smaller than true increments). Indeed, the false increments identified through isotopic

101 testing were suspiciously small and I hypothesized that they were probably false increments while preparing the samples. Each clam should be assessed for false increments when assessing growth increments, terminal growth phase, and age.

A single false opaque increment was present (Modern Clam 11 sample 3). This indicates that environmental or biological factors can create the appearance of an opaque band within a translucent increment. Like false translucent increments, this can give the appearance of an additional year of growth and can confound increment and growth phase identification.

However, this false opaque increment was suspiciously small and appeared to be bounded by two small translucent increments (similar to the case for Modern Clam 11 samples 10 and 11).

In other words, like false translucent increments, false opaque increments can be successfully identified as such by their irregular size and the irregular sizes of the surrounding translucent increments. These false increments reinforce the need to consider multiple years of growth when evaluating clam cross‑sections.

Other than the anomalies identified in Modern Clam 3 (sample 4) and Modern Clam 11

(sample 3), incremental profiles were straightforward. The ∂18O values for opaque and translucent increments fell within their expected ranges: opaque increments formed during cooler temperatures, and translucent increments formed during warmer temperatures. This analysis reveals annual cycles of ∂18O values (interpreted as relative temperatures), which conform to the expected patterns of alternating opaque and translucent values. This analysis validates the use of the incremental growth technique for the modern clams in the St. Augustine collection.

Archaeological Clam Number 88

Archaeological Clam Number 88 was recovered from the Guana shell ring deposit in shovel test

440N, 410 E at level 2 (10‑20 cmbs). This shovel test was located in the northern portion of the

West arm of the ring. Archaeological Clam Number 88 was six years old and I sampled years 4‑

5. This clam was selected to analyze multiple portions of an opaque increment and to analyze somewhat irregular translucent increments. The clam was rather quite large and the opaque increments were broad. These broad opaque increments permitted me to sample three areas

102 (samples 1, 2, and 3) in the opaque increment deposited during the fourth year of life. I drilled these samples at approximately O1, O2, and O3, and anticipated that these samples should yield ∂18O values that indicate gradually increasing relative temperatures. Archaeological Clam

Number 88 had what appeared to be a translucent increment demarcated by two dark bands

(samples 4 and 5) with a light gray area in between. The light gray area between samples 4 and

5 was too small to sample, especially after drilling samples 4 and 5. The next translucent increment also had the appearance of two dark bands at the exterior with a light gray area in between. I sampled this light gray area (sample 7) to verify that it was part of a single translucent increment and not a small opaque increment between two translucent increments. I did not sample the terminal year of growth.

As predicted, the ∂18O values for samples 1‑3 (removed from a single opaque increment) indicate increasing relative temperatures through time (and growth phases) (Figure 7.4). All three ∂18O values are well within the expected range for opaque samples, verifying that archaeological clams, like modern clams, initiate opaque growth in cooler temperatures and

Archaeological Clam 88

1.50

1.00

0.50 O

18 0.00 ∂

‑0.50

‑1.00

‑1.50

Figure 7.4. Profile of ∂18O Vales for Archaeological Clam Number 88. This graph depicts the ∂18O profile for a clam recovered from the Guana ring deposit. The gray squares indicate samples taken from opaque increments and red squares indicate samples taken from translucent increments.

103 cease producing opaque increments in warmer temperatures. The ∂18O values for samples 4 and 5 are well within the expected range for translucent samples, and the value for sample 6 is also within the expected range for an opaque sample. The ∂18O value for sample 7 is clearly in the translucent increment value range, indicating that the light gray area between darker gray translucent bands is, in fact, part of a single translucent increment. Essentially, dark gray exterior bands with a lighter gray interior characterize the translucent increments in

Archaeological Clam Number 88. Samples 4 and 5 demonstrate a slight increase in relative temperature during progression of the translucent increment.

Archaeological Clam Number 222A

Archaeological Clam Number 222A was recovered from the Guana shell ring deposit in shovel test 470N, 480 E at level 5 (40‑50 cmbs). This shovel test was located in the Central portion of the ring at its northern end. This clam was chosen for its atypical growth pattern consisting of several possible false translucent increments (or double bands) and light gray areas between darker gray translucent bands (similar to Archaeological Clam 88). I estimated the age of this clam to be between five and seven years old, depending on whether two translucent bands demarcated new years or were simply false increments. I drilled sample 1 from a straightforward opaque increment, and drilled sample 2 from a fairly small translucent band.

Sample 3 was removed from a very small opaque band between a small translucent band

(sample 2) and another fairly small translucent band (sample 4). I hypothesized that samples 2 and 4 were part of the same translucent increment (and that the interior opaque band was not a true opaque increment). Sample 5 was taken from an area composed of many thin, light gray bands; this area was immediately adjacent to the previous translucent band (sample 4).

Samples 6 and 7 were removed from straightforward opaque and translucent increments, respectively. Sample 8 was removed from a narrow opaque increment that appeared to be small due to growth drop‑off as a result of old age and approaching growth senescence. I removed sample 9 from the terminal growth increment, which I determined to be in the T3 growth phase.

104 Archaeological Clam 222A

1.00

0.50

0.00 O 18 ∂ ‑0.50

‑1.00

‑1.50

Figure 7.5. Profile of ∂18O Vales for Archaeological Clam Number 222A. This graph depicts the ∂18O profile for a clam recovered from the Guana ring deposit. The gray squares indicate samples taken from opaque increments and red squares indicate samples taken from translucent increments. The turquoise square indicates a sample removed from what appeared to be a small opaque increment, but was actually a part of a single translucent increment encompassing samples 2‑4. The green square indicates a sample removed from an area of many thin gray bands that appeared to be part of the preceding translucent increment but was actually part of the following opaque increment.

The ∂18O values for samples 1 and 2 are in the expected range for an opaque and a translucent increment (Figure 7.5). Based on the ∂18O value of sample 3, this opaque band is not a true opaque increment but is part of the same translucent increment as samples 2 and 4. A false opaque increment was also present in Modern Clam Number 11 (sample 3). This result is not surprising given the very narrow size of the opaque band. The ∂18O values of three portions of the translucent increment (samples 2‑4) indicate gradually decreasing relative temperatures.

I removed sample 5 from an area composed of many thin gray bands. I hypothesized that this was part of the previous translucent increment, but the ∂18O value indicates that it is really part of the next opaque increment (it is the initial portion of the opaque deposit). As expected, the ∂18O value for sample 6 indicates an opaque increment, and the ∂18O value for sample 7 indicates a translucent increment. Although sample 8 was removed from a small opaque band, the ∂18O value suggests that it is indeed an opaque increment (as hypothesized).

105 The combination of the false increment (sample 3) and the true increment (sample 8) indicate that this clam is actually six years old. The ∂18O value for sample 9 suggests a translucent increment, as expected. When attempting to determine the season of death using oxygen isotopes, one must compare the terminal ∂18O value to those in the rest of the shell. When one compares the ∂18O value of sample 9 to the values of the three samples from the same translucent increment (samples 2‑4), sample 9 most resembles sample 4, which was taken near the termination of the translucent increment. This indicates relatively cooler temperatures at the end of the translucent increment and suggests a T3 growth phase. In other words, the ∂18O value for sample 9 replicates the T3 assessment based on the incremental growth technique.

Archaeological Clam Number 222B

Archaeological Clam Number 222B was also recovered from the Guana shell ring deposit in shovel test 470N, 480 E at level 5 (40‑50 cmbs). This shovel test was located in the Central portion of the ring at its northern end. Archaeological Clam 222B was four years old and was sampled over years two and three. This was clam was chosen to represent a typical growth pattern in an archaeological clam. I drilled sample 1 from the second year’s opaque growth increment. I had hoped to sample three portions of a translucent increment and evaluate the temperature cycle for this increment, but I was able to remove only two samples (samples 2 and

3) because the increment was not wide enough to accommodate another sample. I drilled sample 4 from an opaque increment, and sample 5 from a translucent increment. I did not remove a sample from the terminal opaque increment.

The ∂18O value for sample 1 is a little low but it is still within the expected range for an opaque increment (Figure 7.6). The ∂18O values for samples 2 and 3 both indicate a translucent increment, but are more similar than I expected (there is a ‑0.02 per mil difference between sample 2 and 3). This similarity may be the result of the following possibilities: 1) inadvertent inclusion of the overlapping middle portion of the translucent increment in samples 2 and 3, 2) inadvertent inclusion of some of the adjacent opaque increment in sample 3, or 3) it may simply be an anomaly. The ∂18O values of samples 4 and 5 fall into the expected ranges for an opaque and a translucent increment.

106 Archaeological Clam 222B

1.00

0.50

0.00

‑0.50 O 18 ∂ ‑1.00

‑1.50

‑2.00

‑2.50

Figure 7.6. Profile of ∂18O Vales for Archaeological Clam Number 222B. This graph depicts the ∂18O profile for a clam recovered from the Guana ring deposit. The gray squares indicate samples taken from opaque increments and red squares indicate samples taken from translucent increments.

Archaeological Clam 336

1.00

0.50

0.00

‑0.50 O 18 ∂ ‑1.00

‑1.50

‑2.00

‑2.50

Figure 7.7. Profile of ∂18O Vales for Archaeological Clam Number 336. This graph depicts the ∂18O profile for a clam recovered from the Guana ring interior. The gray squares indicate samples taken from opaque increments and red squares indicate samples taken from translucent increments.

107 Archaeological Clam 336

Archaeological Clam Number 336 was recovered from Guana ring interior deposit in Feature 4

(test units 2 and 3) from level 3 (170‑180 cmbd). This clam was seven years old and I sampled years five through seven. This clam was chosen for its straightforward seasonal growth and large opaque increments (to permit multiple samples in a single opaque increment). I drilled three samples in the fifth year opaque increment (samples 1‑3). I removed sample 3 from an area that appeared to be part of the opaque increment but exhibited some very light gray shading. I removed samples 4, 5, 6, and 7 from alternating standard increments. I removed sample 8 from the terminal translucent growth increment, which I determined to be in the T3 growth phase. I attempted to remove only the very marginal end of the translucent increment

(the T3 growth portion).

The ∂18O values for samples 1‑3 fall in the expected range for an opaque increment

(Figure 7.7). These samples indicate increasing relative temperatures during the growth of the opaque increment (also seen in Archaeological Clam Number 88 samples 1‑3 and Modern Clam

Number 3 samples 3‑4). The ∂18O values of samples 4‑7 fall well within their expected ranges for the appropriate increment. The ∂18O value of sample 8 indicates a very warm (relative) temperature (warmer than other translucent samples). Because I did not obtain multiple samples within a single translucent increment, it is impossible to accurately evaluate the growth phase or season of death indicated by the terminal ∂18O value. However, the ∂18O value for sample 8 is very low compared to other translucent increment ∂18O values. This could indicate that the terminal growth was deposited in extremely warm temperatures of the summer, representing either a T2 or T3 growth phase. In this case, using the incremental growth technique to determine the terminal growth phase is probably the most reliable method to determine season of death.

Summary of ∂18O Isotopic Profiles for Archaeological Clams from the Guana Shell Ring

Like the clams from the St. Augustine collection, the archaeological clams from the shell ring indicate that a single opaque and a single translucent increment occur per year. The opaque increment occurred in relatively cool temperatures (∂18O values ranging from 1.0 to ‑0.4) and the

108 translucent increment occurred in relatively warm temperatures (∂18O values ranging from ‑0.5 to ‑2.02). Archaeological Clam Number 88 was characterized by translucent increments having dark gray borders and light gray interiors. Isotopic analysis of this interior gray area revealed that this was indeed part of the translucent increment. I identified a false opaque increment in

Archaeological Clam Number 222A (sample 3), which was likely the result of erratic environmental or biological stressors to the clam. This false increment appeared as a suspiciously small opaque band. I initially interpreted an area of many thin gray lines

(Archaeological Clam Number 222A sample 5) as a possible false translucent increment, but the

∂18O value indicated that it was part of the following opaque increment.

All other archaeological clams had relatively straightforward growth profiles with no false or questionable increments. In sum, opaque increments formed during cooler temperatures and translucent increments formed during warmer temperatures, validating the use of the incremental growth technique for the archaeological clams.

Discussion of ∂18O Profiles for Modern Clams from the St. Augustine Collection and

Archaeological Clams from the Guana Shell Ring

The ∂18O profiles for both modern and archaeological collections were similar. Both collections indicated that the opaque increment is deposited exclusively during relatively cool temperatures (∂18O values ranging from 1.0 to ‑0.4) and the translucent increment is deposited exclusively during relatively warm temperatures (∂18O values ranging from ‑0.5 to ‑2.02 [‑2.00 in the modern collection]). Unlike the Narragansett Bay clams (Henry and Cerrato 2007), only a single opaque increment and a single translucent increment were identified per year of growth.

Thus far, patterns of multiple increments per year have not been identified in collections in the southeastern U.S.; instead increments are deposited during the expected annual temperature cycles (warm or cool weather), and yearly growth is comprised of a single opaque and single translucent increment. These findings indicate that the incremental growth technique works for both modern and archaeological clams.

Based on the isotopic analysis, the modern clams from the St. Augustine collection and the archaeological clams from the Guana shell ring are comparable. Opaque and translucent

109 increments for both collections formed consistently within the same temperature ranges. Table

7.1 summarizes the ∂18O values for each increment in each collection (modern or archaeological) and Figure 7.8 depicts boxplots for each category. Average ∂18O values of modern clams are slightly higher than those of the archaeological clams. These higher ∂18O values could be the result of multiple factors, including: 1) slightly lower relative temperatures experienced by the modern collection, 2) a different salinity cycle for the modern clams (salinity as well as temperature influences ∂18O values), or 3) the archaeological clams may have experienced post‑ depositional processes that slightly altered their isotopic signatures. It is possible that overall temperatures were cooler during the lives of the modern clams, resulting in slightly elevated

∂18O values when compared to archaeological clams. It is also highly likely that the anthropogenic changes to the Matanzas River have altered the salinity cycle of the river

(discussed further below). Because both absolute temperature and salinity influence ∂18O values, it is impossible to determine whether one of these factors is causing the slightly increased ∂18O values in the modern collection. While post‑depositional processes may have altered the isotopic signatures of archaeological clams, this seems somewhat unlikely given the overall similarity of ∂18O values for archaeological clams from different proveniences throughout the ring. Despite the slightly higher ∂18O values observed in the modern collection, the modern and archaeological ∂18O values are similar and show similar patterns, indicating that the two collections are comparable. This means that using the St. Augustine collection

Table 7.1. Statistics for ∂18O Values of Modern and Archaeological Clams.

Modern Archaeological Modern Archaeological Statistic Opaque Opaque Translucent Translucent Count 18 15 18 14 Mean 0.46 0.29 ‑1.22 ‑1.43 Median 0.52 0.3 ‑1.27 ‑1.36 Std Dev 0.45 0.44 0.31 0.28 Minimum ‑0.35 ‑0.46 ‑1.62 ‑2.02 Maximum 1.18 0.9 ‑0.64 ‑1.09

110

Figure 7.8. Boxplots of ∂18O Values for Modern and Archaeological Clams. This graph depicts the ∂18O values for each increment in both modern clams from the St. Augustine collection and archaeological clams from the Guana shell ring.

growth phase distributions with known seasons of death is a valid methodology to determine seasons of death for archaeological clams from the Guana shell ring.

The ∂18O value ranges for the opaque increment are well defined for both modern and archaeological collections. This analysis revealed that opaque increment ∂18O values gradually decrease throughout deposition of the opaque increment (relative temperatures steadily increase). This pattern is different than that depicted by Jones and Quitmyer (1996:343) because

111 I was unable to drill as many samples per increment as Jones and Quitmyer. In this study, the

∂18O values for the translucent increment are well defined but do not adhere to a strict pattern of gradual temperature change throughout deposition of the translucent increment. Essentially, the progression of ∂18O values in the translucent increment varied by shell. This variability was also observed in Jones and Quitmyer’s (1996) study.

Of the clams for which multiple samples were available from single translucent increments, three clams demonstrated relative temperatures that became progressively warmer

(Modern Clam 11, Archaeological Clam 88, and Archaeological Clam 222B); two clams demonstrated relative temperatures that became progressively cooler (Modern Clam 11 and

Archaeological Clam 222A). Only two clams, Modern Clam 11 and Archaeological Clam 222A, included three samples in a single translucent increment and each demonstrated a different pattern of changing relative temperatures. Although the progression of ∂18O values in single translucent increments varied by shell, the translucent increment samples all fall within the expected ranges. In other words, translucent increments still formed during specific temperature ranges, indicating that the incremental growth technique is an acceptable method to determine season of death.

The isotopic analysis of modern and archaeological clams clarified some of the unusual bands and/or increments that I observed in both collections. Isotopic analysis revealed that both collections contained false increments, which can give the appearance of an additional year of growth that is not present and can obscure increment and growth phase assessments. In both collections, these increments were characterized by their suspiciously small size and the small size of surrounding increments (e.g., Modern Clam 3, Modern Clam 11, Archaeological Clam

222A). Light gray areas and areas of thin gray lines likely belong to the adjacent opaque increment, rather than the adjacent translucent increment, as indicated by Modern Clam

Number 3 and Archaeological Clam Number 222A. Areas of light gray surrounded by small dark translucent bands, however, are likely part of a single translucent increment, as demonstrated by Archaeological Clam 88. These findings were useful in the determination of ages, growth increments, and terminal growth phases for both the St. Augustine collection and the Guana shell ring assemblage.

112 In sum, although my methodology was slightly different than that of Jones and

Quitmyer (1996), my study of ∂18O values provides valuable information about seasonal growth and verifies the methodology used for this research. The oxygen isotopic analysis revealed that patterns of incremental growth are dependent on seasonal temperatures for both the modern and archaeological clams. The overall ∂18O values for both modern and archaeological collections are very similar. This indicates that the growth phase distributions for the St.

Augustine collection are a valid baseline with which to determine the seasonality of archaeological clams from the Guana shell ring. Isotopic analysis also resolved some slightly anomalous‑appearing bands (both opaque and translucent) and aided in proper identification of age, growth increments, and growth phases.

Modern Comparative Collection

Gathering Live Clams

I successfully collected living clams from the Matanzas River for each month in 2010. Although

I attempted to collect clams at the same time every month, there was some variability in the timing of collections (due to storms, very low temperatures, availability of aquaculturalist, Mr.

Cubbedge, and a nasty case of the H1N1 flu). I typically collected clams at the end of the month and occasionally on the first day of the next month (Table 7.2). Duration of trips into the marsh varied considerably. The following reasons occasionally resulted in long trips to the marsh: difficulty finding clams, Mr. Cubbedge needed to collect oysters or other seafood to sell, Mr.

Cubbedge needed to check on his aquaculture clams, boat problems, and once the boat was inadvertently beached for the duration of an extremely low tide (December 31, 2010 collection).

I was able to gather wild clams from the marsh during every month except October. In the month of October, Mr. Cubbedge was unavailable to accompany me and I was unable to procure enough wild clams on my own. For the month of October I acquired aquaculture clams from the Devils Elbow Clam Farm. This clam farm is located in the Matanzas River approximately 5 km north of Fort Matanzas and approximately 8.5 km north of Collection Area

B. The clams from Devils Elbow were older, larger clams, referred to as chowder clams, and

113 Table 7.2. Dates and Times of Clam Collections.

Month Date Time In Time Out Type January 1/28/10 12:00 PM 1:15 PM Wild February 2/26/10 1:30 PM 3:45 PM Wild March 3/29/10 2:15 PM 3:00 PM Wild April 4/27/10 Not recorded Not recorded Wild May 5/29/10 4:50 PM 5:50 PM Wild June 6/25/10 1:15 PM 2:00 PM Wild July 8/1/10 8:30 AM 9:30 AM Wild August 9/1/10 9:00 AM 10:45 AM Wild September 9/24/10 3:30 PM 5:15 PM Wild October 10/27/10 10:15 AM 10:30 AM Aquaculture November 12/3/10 1:00 PM 4:00 PM Wild December 12/31/10 10:00 AM 3:00 PM Wild

were raised in the Matanzas River and harvested on October 27, 2010. Aquaculture clams often grow more slowly than wild clams, as a result of increased competition for food in a confined space. Aquaculture clams are usually much younger than clams found in the wild (because there is a higher demand for smaller, tender clams), and are kept in bags with clams of the same age. For these reasons, I have excluded the October collection from general discussions about size and age and I have analyzed and discussed the October aquaculture sizes and ages separately.

For every month (except October), I was able to collect clams by raking or by locating them based on signs. Clam signs were readily observable during cooler months, but were less frequent or absent during warmer months (when clams burrow into the sand to avoid high temperatures). Table 7.3 lists each collection and what signs, if any, were observed. When clam signs were present, I was able to locate and gather clams rather quickly. When clam signs were not visible, I located clams by raking the sand.

For each month in the collection, I cross‑sectioned and examined at least 30 valves. The total sample size was 411 valves. For each clam, I recorded its measurements, age, growth increment, and growth phase. The findings for each variable are discussed below.

114 Table 7.3. Occurrence of Clam Signs

Month Date Keyholes Waste Comment January 1/28/10 Yes Yes 2 wks prior very cold temps 20s and 30s February 2/26/10 Yes Yes 45 clams in 2m2; about 100 clams in 25 min

March 3/29/10 Yes No Prior rain had washed away many keyholes; collected ~100 clams in ~45 min April 4/27/10 Some No Other sign: circular spot of dark sand on top of lighter sand ‑ possibly from reburying or burying into deeper sand May 5/29/10 No No June 6/25/10 No No July 8/1/10 No No August 9/1/10 Very few No Water level stayed high during low tide; few keyholes in shaded creek September 9/24/10 Two clams No Other sign: circular dark spot; water level stayed high during low tide; few clams found by dark sand spot signs October 10/27/10 Two clams No Aquaculture from Devils Elbow; observed two clam keyholes in creek by Whitney dock November 12/3/10 Yes Yes All clams found by keyholing; found ~80 clams in 1.5 hrs December 12/31/10 Yes Yes All clams found by signing; ~70 in ~ 1 hr; extremely low tide, water very clear from cold temps killing algae

Water Temperature and Salinity Data

Water temperature data clearly exhibits the typical seasonal pattern of northeast Florida. As expected, average temperatures were highest in June, July, August, and September and lowest in December, January, and February (Figure 7.9). September, October, November had the least variation between high and low temperatures (less than 10 degrees Fahrenheit) (Figure 7.10).

December, January, and to a lesser extent, June had the highest variation between high and low temperatures (more than 20 degrees). The water temperature averages for 2010 are similar to those in 2006, 2007, 2008, and 2009 (Appendix C.1), although January, February, and December

(2010) were slightly colder than previous years. Based on the average temperatures

115 Average Temperature

100 80 60 40 20

Degrees Fahrenheit 0

Figure 7.9. Average Water Temperatures in 2010 This graph depicts the average monthly temperature for each month in 2010. The seasonal changes in temperature are subtle, but present as expected. Temperatures are reported in degrees Fahrenheit.

High and Low Temperatures

100 80 60 40 20 0

Low High Variation

Figure 7.10. High and Low Water Temperatures in 2010 This graph depicts the highest and lowest temperature for each month in 2010, as well as the variation between these high and low temperatures.

116 of the Matanzas River, I expected winter to contain primarily opaque (fast) growth, spring to contain primarily opaque growth but increasing amounts of translucent (slow) growth, summer to contain primarily translucent growth, and fall to contain predominantly translucent growth with increasing opaque growth. This pattern is exactly what I observed in the St. Augustine collection, and will be discussed further below.

Salinity levels for the Matanzas River were fairly stable throughout the year. Average salinity varied by only 4.8 parts per thousand (ppt) over the course of 2010 (Figure 7.11). Clams are tolerant of a wide range of salinity, but the salinity levels indicated for the Matanzas River are above the optimum range for Mercenaria. Grizzle and colleagues (2001:342) stated that pumping rates decrease above 30 ppt, but this is not indicated in the modern collection. It is possible that salinity is higher at the NERRS data collection center than where I gathered clams.

The NERRS data station is upriver from the Matanzas Inlet; this proximity to a large inlet may create a more saline environment at the data station than that of the southern parts of the

Matanzas River where I gathered clams. Although the salinity appears to be higher at the

NERRS data station, the general patterns of changing salinity are likely similar for both the data

Average Salinity

37 36 35 34 33 PPT 32 31 30 29

Figure 7.11. Average Salinity in 2010 This graph depicts the average salinity (in parts per thousand, or PPT) in each month of 2010. Although this graph gives the appearance of considerable variability, average monthly salinity varied by less than 5 PPT.

117 High and Low Salinity

50 40 30 20 10 0

Low High Variation

Figure 7.12. High and Low Salinity Levels in 2010 This graph depicts the high and low salinity levels for each month in 2010. Salinity is shown in PPT.

Average Salinity and Temperature

100 80 60 40 20 0

Salinty Temperature

Figure 7.13. Average Salinity and Water Temperature for 2010 This graph depicts both the average salinity and average temperatures for each month in 2010. Salinity remained fairly constant in the Matanzas River. Temperatures, however, fluctuated on a seasonal basis.

118 station and the clam gathering sites because of their proximity (both sites probably had similar amounts of freshwater input from rain and storm runoff).

Essentially, salinity averages rose from January to October, then decreased in November and December. Salinity varied the most in the month of January (34 percent variation) (Figure

7.12). Heavy rains occurred during the end of January and beginning of February, resulting in increased variability in salinity in January and a slightly lower average salinity in February.

Also, February may be slightly anomalous because approximately one week of data is missing from the dataset (from malfunctions in the data logger). When January is excluded, the highest variability in salinity occurs during March through September, when seasonal rain is frequent.

When the variability of temperature and salinity are examined, it is clear that temperature varies more throughout the year (Figure 7.13) (See Appendix C.1 and C.2 for tables of monthly lows, highs, and averages for temperature and salinity). It appears that temperature has a more prominent effect on the seasonal growth of clams in the Matanzas River. Figure 7.14

Seasonal Averages

90 80 70 60 50 Salinity 40 Temperature 30 20 10 0 Winter Spring Summer Fall

Figure 7.14. Seasonal Average Water Temperature and Salinity during 2010. This graph depicts the seasonal fluctuation in average temperature and average salinity during 2010. Temperature varies according to season, but salinity is less variable throughout the seasons.

119 displays the average temperature and average salinity for each season in 2010. Temperature varies according to typical seasonal patterns, but salinity remains fairly constant throughout the year. It should be noted that the Matanzas River is not a pristine ecosystem. Considerable anthropogenic change has occurred in the Matanzas River estuary and marsh system.

Sedimentation of the marsh continues to be a problem. Over the course of the year, I observed noticeable infilling of the marsh. This infilling caused the creation and extension of sandbars, extension of marsh grasses, and made many areas more shallow (see Figure 6.5). These changes almost certainly affected clam populations over time. Sedimentation has been a problem in portions of the Matanzas River system for some time, as evidenced by numerous docks in the river that are now sitting on dry sand. More importantly, freshwater runoff from residences and stormwater drains enters the river and marsh system on a year‑round basis. This anthropogenic freshwater influx likely masks any natural, seasonal cycles of salinity in the

Matanzas River system. The unaltered Matanzas River and marsh ecosystem may have had a very different salinity profile over the course of a year.

Measurements

Measurements were taken on all modern valves assessed for seasonality (n=376). A single clam could not be measured due to extensive injury and malformation. Measurements for the month of October are discussed separately because these are aquaculture clams and all other clams are

Table 7.4. Shell Measurements (in millimeters) for the St. Augustine Collection (excluding October).

Measurements Hinge Length Height Count 376 376 376 Missing 1 1 1 Mean 44.32 73.69 68.99 Median 43.75 72.87 68.36 Std Dev 5.95 9.41 9.22 Min 27.78 46.76 42.72 Max 65.26 108.34 102.42

120 wild. Table 7.4 provides basic measurement information for all clams except the aquaculture clams from October. Mean hinge length was 44.32 mm, mean length was 73.69 mm, and mean height was 68.99 mm. Of the three measurements taken, shell length varied the most and hinge length varied the least. I examined the relationships (ratios) between the three different measurements by calculating the following: shell height is what percent of shell length, hinge length is what percent of shell length, and hinge length is what percent of the shell height?

Table 7.5 presents the basic information derived from these relationships (excluding October).

The shape of the clam influences height and length relationships, but appears to influence hinge length to a lesser degree. Hinge and length ratios and hinge and height ratios are consistent

(standard deviations less than 3). Height and length ratios, however, varied considerably and had a lower minimum ratio, suggesting that this collection has at least some ovate or elongated clams. The ratios between height, length, and hinge width are generally linear: when one measurement increases the others usually increase as well. Figures 7.15‑17 demonstrate the relationships between each of the three measurements for all months of collection except

October.

The aquaculture clams from October were smaller than the majority of wild clams I gathered throughout the year. The mean hinge length was 43.16 mm, mean length was 72.35 mm, and mean height was 67.21 mm. All measurement averages including minimum and maximum sizes were smaller than those of the wild clams (Appendix D.1). Of the three measurements, shell length varied the most and hinge length varied the least. The aquaculture

(October) clams had less variability in size, probably because they were of similar age and had

Table 7.5. Shell Measurement Relationships for the St. Augustine Collection (excluding October).

Measure Count Average Minimum Maximum Std Dev Height Length Ratio 376 93.60% 59.53% 100.00% 2.94 Hinge Length Ratio 376 60.14% 37.76% 68.51% 2.39 Hinge Height Ratio 376 64.27% 58.41% 81.60% 2.27

121

Figure 7.15. Length and Height Scatterplot for the St. Augustine Collection. This graph depicts the length and height for all wild clams in the St. Augustine Collection (October aquaculture clams are excluded).

122

Figure 7.16. Length and Hinge Width Scatterplot for the St. Augustine Collection. This graph depicts the length and hinge width for all wild clams in the St. Augustine Collection (October aquaculture clams are excluded).

123

Figure 7.17. Height and Hinge Width Scatterplot for the St. Augustine Collection. This graph depicts the height and hinge width for all wild clams in the St. Augustine Collection (October aquaculture clams are excluded).

124 access to essentially the same amount of food and other resources. The majority of measurement relationships for the aquaculture (October) clams are very similar to those of the wild clams, although this collection does not appear to have the ovate or elongated clams apparent in the wild assemblage. All aquaculture measurement relationships were consistent

(standard deviations less than 3) and are listed in Appendix D.2).

Age

I was able to determine the age of 392 clams (95.4 percent) from the St. Augustine Collection.

The oldest clam in the collection was 19 years old; the youngest clam was two years old.

Aquaculture clams from the month of October will be discussed separately from the rest of the

St. Augustine Collection.

On average, clams from the St. Augustine collection (excluding October) were 6.26 years old (standard deviation = 3.02) (Table 7.6). Clams that were four years old were most frequent in the collection, followed closely by clams that were three years old. Survivorship declined steadily after 5 years of age, and declined rapidly after eleven years of age (Figure 7.18). Clams reaching ages greater than thirteen each contributed less than one percent of the sample. The aquaculture clams from the month of October exhibited little variation in age (because they were aquaculture). Ages observed in the October aquaculture collection include 3, 4, and 5, but the vast majority of these clams were four years old (82.4%). Clams from the October aquaculture collection averaged 3.94 years old (standard deviation 0.42) (Appendix D.3).

Table 7.6. Age Data for the St. Augustine Collection (excluding October). Age Value Count 358 Mean 6.26 Median 5.5 Mode 4 Std Dev 3.02 Min 2 Max 19

125 St. Augustine Clam Ages

70 60 50 40 30 20 10 0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Figure 7.18. St. Augustine Collection Clam Age Distribution (except October). This graph depicts the distribution of ages (counts) for the St. Augustine Collection (except October). Clams were most frequently three or four years old. Survivorship declines steadily after four years and declines rapidly after eleven years of age.

As discussed previously, I exploited wild clams from two collection areas: Area A and

Area B. I gathered clams from Area A in January and June, and from Area B in all other months except October. Area A is in a clam aquaculture lease area, meaning that people other than the lease‑holder are prohibited from gathering clams in that area. This does not ensure that people do not gather clams from this area. The clam farmer I worked with had his aquaculture clams stolen out of his bags on several occasions, indicating that people illegally gathered in this area at least occasionally (Phil Cubbedge, personal communication 2010). A boat is necessary to access this area, which discourages some illegal gathering. Area B is a non‑leased area that is open to public and commercial gathering. Area B is accessible on foot from A1A Boulevard and

I observed both commercial and recreational shellfish exploitation in this location on several occasions. Because these areas were subject to slightly different gathering pressures, I examined age at both locations. Because the sample size was small for Area A (January and June collections only), I compared Area A to a subsample of Area B (July and December collections).

This produced the samples of the exact same size. In general, clams from Area A were slightly

126 older than clams from the subsample of Area B (Table 7.7). Mean age at Area A was over one year older than at Area B. Median age was 1.5 years older at Area A and the mode of clams from Area A was one year older than clams for Area B. Standard deviations were similar for both areas. When Area A is compared to the entire Area B sample, Area A clams are still slightly older than Area B clams. Hinge length measurements for Area A are slightly larger than those for the subsample of Area B, but are nearly identical to the whole Area B sample

(Appendix E.1). The slightly increased age at Area A is likely a result of its increased protection from clam exploitation.

Although there is considerable variability in the size and age of individual clams, overall size of the clam generally increases with age (especially during the first few years of life).

However, old clams can be very small and young clams can be very large. For example, in the

St. Augustine collection a 17 year‑old clam and a four year‑old clam were approximately the same size, and a two year‑old clam and a 7 year‑old clam were nearly the same size (Figure

7.19). Nevertheless, in large enough samples, one can observe the general trend of increasing size and age.

Seasonality

The St. Augustine collection included 411 valves. Of these, I was able to determine the terminal growth increment (opaque or translucent) for 376 valves (91.5 percent of the sample).

Table 7.7. St. Augustine Collection Age by Collection Area. This table depicts the ages for Area A, which consisted of collections made in January and June, two months of Area B (July and December), and the entire Area B collection.

Age Area A Area B (Jul & Dec) Area B (All) Count 66 66 292 Mean 7.38 6.09 6 Median 7 5.5 5 Mode 4 3 3 Std Dev 2.78 2.73 3.02 Min 3 3 2 Max 13 12 19

127

Figure 7.19. Scatterplot of Hinge Width and Clam Age in the St. Augustine Collection. Although there is considerable variability in the size and age of individual clams, this scatterplot shows a general pattern of increasing size with age, particularly in the early years of life. However, this relationship is not strictly linear – very young clams can be quite large and very old clams can be rather small.

Traditionally, growth increment frequencies have been used to identify whether collections occurred predominantly in cool weather (winter – spring) or warm weather (summer – fall).

Claassen 1998:163) refers to this as the fast/slow growth technique. Based on average temperatures of the Matanzas River, I expected winter to contain primarily opaque (fast) growth, spring to contain primarily opaque growth but increasing amounts of translucent

(slow) growth, summer to contain primarily translucent growth, and fall to contain predominantly translucent growth with increasing opaque growth. This pattern is

128 Table 7.8. Increment Distributions for Winter – Spring, Summer – Fall, and Annual.

Increment Winter ‑ Spring Summer ‑ Fall Annual Opaque (N) 147 21 168 Opaque (%) 78.2% 11.2% 44.7% Translucent (N) 41 167 208 Translucent (%) 21.8% 88.8% 55.3%

Winter ‑ Spring

80 70 60 50 40

Percent 30 20 10 0 Opaque Translucent

Figure 7.20. Winter – Spring Growth Increment Distribution for the St. Augustine Collection. This graph depicts the percentage of each increment (opaque or translucent) through winter and spring.

demonstrated in the St. Augustine collection. The St. Augustine collection followed the predicted patterns of seasonal growth: opaque growth primarily occurred during cooler months of winter through spring, and translucent growth primarily occurred during the warmer months of summer through fall. Table 7.8 and Figures 7.20‑7.22 provide the frequencies for opaque and translucent increments for semi‑annual (cool and warm weather) and annual divisions. Quitmyer and colleagues’ (1997:836) work corroborated Ansell’s (1968) proposed maximum temperature for optimal shell growth. Quitmyer et al. (1997) found that the translucent growth phase initiated at approximately 77° F, and this is indicated in the St.

129 Summer ‑ Fall

90 80 70 60 50 40 Percent 30 20 10 0 Opaque Translucent

Figure 7.21. Summer – Fall Growth Increment Distribution for the St. Augustine Collection. This graph depicts the percentage of each increment (opaque or translucent) through winter and spring.

Annual

60 50 40 30

Percent 20 10 0 Opaque Translucent

Figure 7.22. Annual Growth Increment Distribution for the St. Augustine Collection. This graph depicts the percentage of each increment (opaque or translucent) over the course of one year.

130 Monthly Percentages of Increments

100.0 90.0 80.0 70.0 60.0 50.0 Translucent 40.0 30.0 Opaque 20.0 10.0 0.0

Figure 7.23. Percentage of Opaque and Translucent Growth Increments per Month in 2010.

Augustine collection as well. In the St. Augustine collection, months where the average temperature exceeded 77° F included clams that were almost exclusively in the translucent growth phase. This finding indicates the importance of influence that temperature has on the incremental growth of clams.

I calculated the frequency of each increment for every month in 2010 (Figure 7.23), but monthly distributions were not sufficiently unique to allow for the identification of single months in an assemblage. I also calculated the frequency of opaque and translucent increments for each season. Table 7.9 provides the outcomes of assessment for each season and for the entire year, and Table 7.10 provides the frequencies of each increment for individual seasons and for the entire year. Somewhat surprisingly, seasonal distributions of growth increments appear to be unique (Figure 7.24) and appear to correspond with seasonal temperature ranges

(Figure 7.25). Each season demonstrates a distinct percentage of opaque and translucent growth increments that could allow for identifications of single seasons in an assemblage.

However, the variation between the seasonal and semi‑annual percentages appears to be too great to allow for confident identification of seasons (variation between seasonal and semi‑ annual assemblages ranges from 11.2 percent to 21.9 percent). For example, the percentage of

131 Table 7.9. Outcomes for Increment Assessment by Season for the St. Augustine Collection.

Outcome Winter Spring Summer Fall Annual Assessed (N) 99 89 96 92 376 Assessed (%) 98% 79.5% 96% 93.9% 91.5% Unreadable (N) 2 23 4 6 35 Unreadable (%) 2% 20.5% 4% 6.1% 8.5% Assemblage (N) 101 112 100 98 411 Assemblage (%) 24.6% 27.3% 24.3% 23.8% 100%

Table 7.10. Growth Increment Distributions by Season for the St. Augustine Collection.

Increment Winter Spring Summer Fall Annual Opaque (N) 91 56 0 21 168 Opaque (%) 91.9% 62.9% 0% 22.8% 44.7% Translucent (N) 8 33 96 71 208 Translucent (%) 8.1% 37.1% 100% 77.2% 55.3%

Growth Increment by Season

100

60 Opaque

Percent 40 Translucent

0 Winter Spring Summer Fall

Figure 7.24. Growth Increment Distributions by Season for the St. Augustine Collection. This graph depicts the percentage of each increment (opaque or translucent) for each season.

132 Temperatures & Increments

100 90 80 70 Temperature 60 50 Opaque 40 30 Translucent 20 10 0 Winter Spring Summer Fall

Figure 7.25. Seasonal Temperature Averages and Increment Frequencies. This graph shows the average temperature along with the percentage of clams experiencing either opaque or translucent increments for each season.

clams experiencing opaque growth during the spring is 62.9 percent, while the percentage of clams experiencing opaque growth in the winter – spring is 78.2 percent. This means that relying solely on incremental growth distributions to assess the seasonality of an assemblage could result in misidentification of the season(s). Assemblages that are likely to be misidentified include those that do not precisely fit into a single season of collection, a multi‑ seasonal collection, or a multi‑seasonal collection with varying intensity of gathering. Based on the current evidence, it appears that increment frequencies are best used to determine more general cool weather or warm weather collection patterns.

I was able to successfully identify growth phases for 334 (81.3 percent) of the clams in the St. Augustine collection. Reasons for failure to identify growth phases include senescence

(71.4 percent of unreadable clams) and unclear growth patterns (28.6 percent of unreadable clams). Monthly sample sizes of successfully assessed clams ranged from 23 to 34 individuals per month (Appendix F.1 and F.2), and seasonal sample sizes ranged from 79 to 89 individuals

(Table 7.11). Although monthly growth profiles demonstrated an overall gradual progression

133 Table 7.11. Outcomes of Growth Phase Assessments for the St. Augustine Collection.

Outcome Winter Spring Summer Fall Annual Assessed (N) 82 79 84 89 334 Assessed (%) 81.2% 70.5% 84% 90.8% 81.3% Unreadable (N) 19 33 16 9 77 Unreadable (%) 18.8% 29.5% 16% 9.2% 18.7% Assemblage (N) 101 112 100 98 411 Assemblage (%) 24.6% 27.3% 24.3% 23.8% 100%

Table 7.12. Growth Phase Distributions by Count for the St. Augustine Collection.

Growth Phase Winter Spring Summer Fall Annual O1 39 6 0 21 66 O2 21 17 0 0 38 O3 14 28 0 0 42 T1 0 23 35 4 62 T2 0 3 25 5 33 T3 8 2 24 59 93

Table 7.13. Growth Phase Distributions by Percent for the St. Augustine Collection.

Growth Phase Winter Spring Summer Fall Annual O1 47.6% 7.6% 0% 23.6% 19.8% O2 25.6% 21.5% 0% 0% 11.4% O3 17.1% 35.4% 0% 0% 12.6% T1 0% 29.1% 41.7% 4.5% 18.6% T2 0% 3.8% 29.8% 5.6% 9.9% T3 9.8% 2.5% 28.6% 66.3% 27.8%

of phases, each monthly growth profile was insufficiently distinct to facilitate the identification of individual months of collection (Appendix F.3‑F.14).

Seasonal growth profiles were sufficiently distinct to identify collections during each season. Tables 7.12 and 7.13 provide the counts and percentages for each growth phase in the

St. Augustine Collection. Each season demonstrated a unique peak in a growth phase as well as a unique distribution of overall phases. The winter growth profile (Figure 7.26) is characterized

134 by a peak in O1 followed by O2, and O3, and includes a small amount of T3. The spring growth profile (Figure 7.27) is characterized by a peak in O3 followed by T1, O2, and O1. Small amounts of T2 and T3 were also present in the spring profile. All valves exhibiting the T1 and

T2 growth phases were collected in the month of May (gathered on May 27, 2010). The distribution for the May profile consists of T1 and T2 phase clams, and is nearly identical to the

June distribution (the first month of summer). It is possible that the May clams, which were gathered late in the month and were experiencing primarily T1 growth, should be excluded from the spring growth profile. For this reason, I have created an alternate spring growth profile using the data from the months of March and April (Figure 7.28). This profile resembles the original spring profile (with the month of May included) but lacks the T1 and T2 growth phases. It is important to note that in the Kings Bay collection (Quitmyer et al. 1997), spring clams peak in the T1 growth phase, indicating that the T1 clams in the St. Augustine collection may truly represent part of the spring growth profile. In the Indian River collection (Quitmyer

1995), a peak in T1 clams does not occur until the summer. It appears that St. Augustine is indeed in a transitional area for seasonal growth of clams. A second year of clam collection in

St. Augustine could clarify whether late spring clams enter the T1 and T2 growth phases or if

Winter

Percent 20

0 O1 O2 O3 T1 T2 T3

Figure 7.26. Winter Growth Profile for the St. Augustine Collection.

135 Spring

40 35 30 25 20

Percent 15 10 5 0 O1 O2 O3 T1 T2 T3

Figure 7.27. Spring Growth Profile for the St. Augustine Collection.

Alternate Spring

30 Percent 20

0 O1 O2 O3 T1 T2 T3

Figure 7.28. Alternate Spring Growth Profile for the St. Augustine Collection. This graph depicts an alternative spring growth profile for the St. Augustine Collection. This graph omits the May 2010 collection, which includes solely T1 and T2 clams and is nearly identical to the June 2010 collection (the first month of summer).

136 Summer

45 40 35 30 25 20 Percent 15 10 5 0 O1 O2 O3 T1 T2 T3

Figure 7.29. Summer Growth Profile for the St. Augustine Collection.

Fall

70 60 50 40 30 Percent 20 10 0 O1 O2 O3 T1 T2 T3

Figure 7.30. Fall Growth Profile for the St. Augustine Collection.

this transition actually occurs at the very beginning of summer. Based on the seasonal growth patterns observed in the Kings Bay and Indian River collections, however, it appears that the T1 and T2 phase clams may legitimately belong in the spring season. Regardless of the T1 and T2 growth phase issue, the spring growth profile has distinct peaks of seasonal growth that allow

137 for the correct identification of a spring collection in an assemblage. Further comparisons of the

St. Augustine, Indian River, and Kings Bay collections will be discussed in the following chapter.

The St. Augustine collection summer growth profile (Figure 7.29) demonstrates a peak in the T1 growth phase, followed by T2 and T3. The summer profile completely lacks any opaque phase clams. The fall growth profile (Figure 7.30) is distinguished by a T3 peak followed by O1; small amounts of T2 and T3 are also present. As in the Kings Bay and Indian

River collections, the T3 growth phase was the most commonly identified phase in the collection and is present in every season. In the St. Augustine collection clams in the T3 phase can be found in any season of the year, although they are most common in the fall. This highlights the necessity of having large samples in order to accurately determine the season(s) of collection.

The annual growth profile (Figure 7.31) contains clams in all phases of growth and is clearly distinct from each seasonal growth profile. Because winter and spring profiles and summer and fall profiles are similar, and because other archaeological sites have demonstrated clam collection during multiple seasons, I have created two semi‑annual graphs that depict a

Annual

15 Percent 10

0 O1 O2 O3 T1 T2 T3

Figure 7.31. Annual Growth Profile for the St. Augustine Collection.

138

Winter ‑ Spring

15 Percent 10

0 O1 O2 O3 T1 T2 T3

Figure 7.32. Winter – Spring Growth Profile for the St. Augustine Collection.

Summer ‑ Fall

50.0

40.0

30.0

Percent 20.0

10.0

0.0 O1 O2 O3 T1 T2 T3

Figure 7.33. Summer – Fall Growth Profile for the St. Augustine Collection.

139 Table 7.14. Outcomes of Growth Phase Assessments for Winter – Spring and Summer – Fall in the St. Augustine Collection.

Outcome Winter ‑ Spring Summer ‑ Fall Annual Assessed (N) 161 173 334 Assessed (%) 75.6% 87.4% 81.3% Unreadable (N) 52 25 77 Unreadable (%) 24.4% 12.6% 18.7% Assemblage (N) 213 198 411 Assemblage (%) 51.8% 48.2% 100%

Table 7.15. Growth Distributions for Winter – Spring and Summer – Fall.

Growth Winter ‑ Summer ‑ Winter ‑ Summer ‑ Phase Spring (N) Fall (N) Spring (%) Fall (%) O1 45 21 28 12.1 O2 38 0 23.6 0 O3 42 0 26.1 0 T1 23 39 14.3 22.5 T2 3 30 1.9 17.3 T3 10 83 6.2 48

winter through spring growth profile (Figure 7.32) and a summer through fall growth profile

(Figure 7.33). Table 7.14 provides the number of clams for each semi‑annual profile, and Table

7.15 provides frequencies of growth phases for each semi‑annual profile. In the following chapter I will compare the seasonal and annual growth profiles of the St. Augustine Collection to the Kings Bay and Indian River Collections.

Archaeological Clams from the Ring Interior

Clams from the Ring Interior sample were obtained from the 2005 excavation in a fairly small area from a relatively shallow deposit. I divided the Ring Deposit into upper and lower levels to examine any changes that might have occurred over time during deposition of the shell.

After its abandonment, St. Johns people briefly occupied the area of the Guana shell ring; small

140 numbers of St. Johns ceramics (n=46, 18.5 percent of the entire ceramic assemblage) were recovered on the surface and in the upper three levels of the ring interior deposit (one St. Johns sherd weighing less than 1 gram was recovered in Level 5, but as it is so small and almost certainly the result of downward migration, it is not considered here) (Saunders and Rolland

2006). Because few clams from the ring interior were in good condition, I elected to use clams from contexts that contained St. Johns sherds. However, if I removed all clams from levels containing more St. Johns sherds than Orange sherds, a total of 8 clams (5 percent of the sample) would be excluded (from Levels 1 and 2, all units). Excluding these clams from analysis does not alter the results discussed below. If I removed all clams from levels containing some St.

Johns sherds but more Orange sherds, a total of 70 clams (45.2 percent of the sample) would be excluded (from Level 3, all units). Excluding these clams from analysis does not alter the results discussed below. The Level 3 contexts are characterized by more Orange sherds than St. Johns sherds (70 clams). The presence of St. Johns sherds in these contexts is probably the result of downward migration in the midden matrix, a common problem in shell matrices. This means that these 70 clams likely represent Late Archaic period deposition and reflect Late Archaic subsistence (and should be included in the present study). Nevertheless, excluding clams that were recovered in levels with St. Johns sherds (n=78, 50.3 percent of the ring interior sample) considerably decreases the sample size, but does not change the results of the study. Therefore,

I have elected to present these clams in the results.

Clam Measurements

I was able to obtain size measurements for the majority of clams from the Ring Interior sample.

Some measurements could not be accurately obtained due to injuries, shell malformations, or breakage. Table 7.16 provides general information on each of the three measurements taken.

Mean hinge length was 50.75 mm, mean shell length was 82.39 mm, and mean shell height was

87.12 mm. Of the three measurements, height varied the most and hinge length varied the least.

As for the St. Augustine collection, I calculated the relationships (ratios), between measurements (e.g., shell height is what percent of shell length?) for the Ring Interior (Table

7.17). Ratios of shell measurements are quite consistent (standard deviations less than 3). To

141 Table 7.16. Shell Measurements (in millimeters) for the Ring Interior.

Measurements Hinge Length Height Count 115 117 110 Missing 40 38 45 Mean 50.75 82.39 87.12 Median 51.01 82.5 87.31 Std Dev 8.06 13.51 14.6 Min 35.44 58 61.78 Max 71.76 116.5 127.65

Table 7.17. Shell Measurement Relationships for the Ring Interior.

Measure Count Average Minimum Maximum Std Dev Height Length Ratio 109 94.05% 89.47% 100.11% 1.95 Hinge Length Ratio 106 58.16% 44.84% 63.78% 2.59 Hinge Height Ratio 112 61.91% 48.37% 67.56% 2.63

Table 7.18. Hinge Length Measurements (in millimeters) for Ring Interior Levels 1‑3 and 4‑6.

Hinge Length Levels 1‑3 Levels 4‑6 Count 59 56 Missing 19 21 Mean 50.12 51.41 Median 50.24 51.03 Std Dev 8.03 8.1 Min 36.87 35.44 Max 67.82 71.76

142 Ring Interior Clam Ages

0 1 2 3 4 5 6 7 8 9 10 11 16 17 20 2 3 4 5 6 7 8 9 10 11 12 13

Figure 7.34. Ring Interior Clam Ages. This graph depicts the distribution of ages (counts) for the Ring Interior sample. Clams were most frequently four or five years old. Survivorship drops considerably after five years and again after eight years of age.

identify changes in size over time I compared the measurements of levels 1‑3 to those of levels

4‑6. Table 7.18 shows the differences in hinge length for upper and lower levels (hinge length, length, and height measurement summaries are listed in Appendix G.1). In all measurements except minimum size, clams from the upper levels were slightly smaller than those from the lower levels.

Age

I was able to determine the age of 104 clams (67.1%) from the Ring Interior sample. I was unable to age clams that were missing parts of their ontogenetic sequence, were bleached, or were otherwise unreadable. The oldest clam in the collection was 20 years old; the youngest clam was two years old (Figure 7.34). On average, clams from this sample were 5.67 years old

(standard deviation = 2.79) (Table 7.19). Clams that were four and five years old were most frequent in the collection, and survivorship dropped off considerably after eight years of age.

To evaluate any change over time, I compared age data for the upper and lower levels (Table

7.20). With the exception of median age and minimum age, all values decreased considerably

143 Table 7.19. Age Data for the Ring Interior.

Age Value Count 104 Mean 5.67 Median 5 Mode 4, 5 Std Dev 2.79 Min 2 Max 20

Table 7.20. Age Data for Ring Interior Levels 1‑3 and 4‑6.

Age Levels 1‑3 Levels 4‑6 Count 54 50 Mean 5.39 5.98 Median 5 5 Std Dev 2.18 3.32 Min 2 2 Max 17 20

from the lower levels to the upper levels, (i.e., upper level clams were younger). Average age dropped from 5.98 years (lower levels) to 5.39 years (upper levels). In the lower levels seven clams lived beyond the age of eight, whereas in the upper levels only a single clam lived beyond the age of eight. Age of archaeological clams will be discussed in greater detail in

Chapter 8.

Clam Seasonality

The Ring Interior sample included 155 valves. Of these, I was able to determine the terminal growth increment for 105 valves (67.7 percent of the sample). The vast majority (86.7 percent) of clams were experiencing opaque growth at the time of death. This indicates that these clams were collected primarily during the cool weather months of winter and spring.

I was able to successfully identify growth phases for 78 valves (50.3 percent) in the Ring

Interior collection. The percentage of readable clams was quite low for this collection, primarily

144 because many of the clams were in poor condition. Many clams exhibited bleaching of the interior growth rings, primarily at the interior of the umbo and the interior margin. When present, this bleaching frequently obscured growth rings, making determinations of terminal growth phases impossible. Bleaching was observed in 36 sectioned valves (23.2 percent of the ring interior sample), and obfuscated growth phase determinations in 30 valves (19.4 percent)

(12 of these clams were also senescent). Thirty‑six valves (23.2 percent) were identified as senescent or probably senescent, and none of these could be interpreted. A total of 54 valves, or

34.8 percent of the sample was unreadable as a result of senescence or bleaching. These bleached and senescent valves comprised 70.1 percent of the total number of unreadable valves.

Twenty‑three clams (the remaining 29.9 percent) could not be read because they exhibited unclear growth patterns or were fragments that did not possess enough growth increments

(multiple increments are needed for comparison with the terminal growth phase).

Table 7.21. Growth Phases for the Ring Interior.

Growth Phase Count Percent O1 17 21.8 O2 20 25.6 O3 34 43.6 T1 4 5.1 T2 1 1.3 T3 2 2.6

The clams from the ring interior exhibited a clear pattern of seasonal exploitation. The majority of clams were in the O3 growth phase, followed by O2, and O1 (Table 7.21). Very few clams were experiencing translucent growth at their time of death. This growth profile (Figure

7.35) most closely resembles a spring collection period. To investigate any change over time in seasonal clam use, I compared the growth profiles for the upper levels (1‑3) and lower levels (4‑

6) (Table 7.22 and Figures 7.36 and 7.37). The growth profiles are extremely similar and both indicate spring collections. In the ring interior, clam collection occurred predominantly during the spring throughout the duration of the deposit.

145

Ring Interior

45 40 35 30 25 20 Percent 15 10 5 0 O1 O2 O3 T1 T2 T3

Figure 7.35. Growth Profile for Clams in the Ring Interior. This graph depicts the distribution of growth phases for the Ring Interior. This growth profile most resembles that of a spring collection.

Table 7.22. Growth Phases for Levels 1‑3 and 4‑6 in the Ring Interior.

Growth Phase Levels 1‑3 (N) Levels 4‑6 (N) Levels 1‑3 (%) Levels 4‑6 (%) O1 7 10 17.5 26.3 O2 11 9 27.5 23.7 O3 19 15 47.5 39.5 T1 2 2 5 5.3 T2 1 0 2.5 0 T3 0 2 0 5.3 Totals 40 38 100 100.1

146 Ring Interior Levels 1‑3

Percent 20

0 O1 O2 O3 T1 T2 T3

Figure 7.36. Growth Profile for the Ring Interior Levels 1‑3.

Ring Interior Levels 4‑6

40 35 30 25 20

Percent 15 10 5 0 O1 O2 O3 T1 T2 T3

Figure 7.37. Growth Profile for the Ring Interior Levels 4‑6.

147 Archaeological Clams from the Ring Deposit

Clams from the ring deposit sample were recovered from individual tests spread over a large area. For this reason, I divided the shell ring into multiple segments in order to understand any differences or changes in clam exploitation and deposition across the ring. I hypothesized that a spatial analysis of clam seasonality throughout the ring would indicate any differences in seasonal use of the shell ring. As with the Ring Interior sample, I divided the Ring Deposit into upper and lower levels to examine change over time. Although the ring deposit is deeper than the ring interior, I did not divide the ring deposit into additional levels. The ring deposit was excavated as centimeters below surface (cmbs), meaning that the same level in a different test could be at a different elevation. For this reason, I decided to divide the shell ring deposit into upper and lower levels. Additionally, I have divided the ring into three areas, the Central

Portion, the East Arm, and the West Arm (Figure 7.38). Compared to the Central Portion and

West Arm, very few clams were found in the East Arm, probably because the East Arm is the smallest and thinnest portion of the ring deposit. To examine any small‑scale differences in spatial use of the ring, I evaluated seasonality in each shovel test and test unit as a separate entity.

As in the ring interior, small numbers of St. Johns ceramics (n=60, or 5 percent of the total ceramic assemblage) were recovered on the surface and in the upper four levels of the ring deposit (Russo et al. 2002). In order to achieve large samples of clams, I elected to use clams from contexts that contained St. Johns sherds. However, if I removed all clams from contexts containing more St. Johns sherds than Orange sherds, a total of 5 clams (2.1 percent of the ring sample) would be excluded. If I removed all clams from contexts containing some St. Johns sherds but containing more Orange sherds, a total of 13 clams (5.4 percent of the sample) would be excluded. Excluding any of these clams from the analysis does not alter the results discussed below. The presence of St. Johns sherds in these contexts is probably the result of downward migration in the midden matrix, a common problem in shell matrices. This means that these 13 clams probably represent Late Archaic period deposition and reflect Late Archaic subsistence strategies (and should be included in the present study). Nevertheless, excluding any clams

148

Figure 7.38. Distribution of Ring Deposit Areas, Shovel Tests, and Test Unit. Site map adapted from Russo et al. 2002. The black rectangle represents the test unit and the black circles represent shovel tests that yielded clams for analysis. The Central Portion includes the test unit and shovel tests on either side. The East Arm includes all shovel tests to the right of the Central Portion and the West Arm includes all the shovel tests to the left of the Central Portion.

149 that were recovered in contexts with St. Johns sherds (n=18, 7.5 percent) slightly decreases the sample size but does not alter the results of the study. Therefore, I have elected to present these clams in the results.

Clam Measurements

I was able to obtain size measurements for the majority of clams from the Ring Deposit sample.

Some measurements could not be accurately obtained due to injuries, shell malformations, or breakage. Table 7.23 provides general information on each of the three measurements taken.

Mean hinge length was 47.92 mm, mean shell length was 82.88 mm, and mean shell height was

77.15 mm. Of the three measurements, length varied the most and hinge length varied the least.

I calculated the relationships (ratios) between measurements (e.g., shell height is what percent of shell length?) for the Ring Interior (Table 7.24). Ratios of shell measurements are very consistent (standard deviations of 2 or less). In order to identify changes in size over time I compared the measurements of levels 1‑5 to those of levels 6‑12. Table 7.25 shows the differences in hinge length for upper and lower levels (hinge length, length, and height measurement summaries are listed in Appendix G.2). In all measurements except maximum size, clams from the upper levels were slightly smaller than those from the lower levels.

Table 7.23. Shell Measurements (in millimeters) for the Ring Deposit.

Measurements Hinge Length Height Count 170 167 186 Missing 71 74 55 Mean 47.92 82.88 77.15 Median 47.91 82.9 76.98 Std Dev 7.79 14.06 13.34 Min 27.46 47.87 46.48 Max 72.24 130 137.73

150 Table 7.24. Shell Measurement Relationships for the Ring Deposit.

Measure Count Average Minimum Maximum Std Dev Height Length Ratio 159 93.32% 88.40% 99.72% 2 Hinge Length Ratio 155 58.48% 52.36% 64.13% 1.93 Hinge Height Ratio 169 62.72% 57.72% 68.14% 1.95

Table 7.25. Hinge Length Measurements (in millimeters) for Ring Deposit Levels 1‑5 and 6‑ 12.

Hinge Length Levels 1‑5 Levels 6‑12 Count 83 87 Missing 41 30 Mean 47.68 48.16 Median 47.82 48.30 Std Dev 8.41 7.19 Min 27.46 31.88 Max 72.24 64.76

Age

I was able to determine the age of 186 clams (77.2 percent) from the Ring Deposit sample. I was unable to age clams that were missing parts of their ontogenetic sequence, were bleached, or were otherwise unreadable. The oldest clam in the collection was 11 years old; the youngest clam was two years old (Figure 7.39). On average, clams from this sample were 4.2 years old

(standard deviation = 1.87) (Table 7.26). Clams that were three and four years old were most frequent in the collection, and survivorship dropped off considerably after five years of age, with a secondary drop after eight years of age. To evaluate any change over time, I compared age data for the upper and lower levels (Table 7.27). Mean age and mode decreased in the upper levels (i.e., upper level clams were younger). Average age decreased from 4.28 years

(lower levels) to 4.11 years (upper levels), and the mode decreased from four to three. The median age, minimum age, and maximum age remained the same. Age of archaeological clams will be discussed in greater detail in Chapter 8.

151 Ring Deposit Clam Ages

30 Percent 20

0 2 3 4 5 6 7 8 9 10 11 1 2 3 4 5 6 7 8 9 10

Figure 7.39. Ring Deposit Clam Ages. This graph depicts the distribution of ages (counts) for the Ring Deposit sample. Clams were most frequently three or four years old. Survivorship drops considerably after five years of age and drops off again after eight years of age.

Table 7.26. Age Data for the Ring Deposit.

Age Value Count 186 Mean 4.2 Median 4 Mode 3 Std Dev 1.87 Min 2 Max 11

Clam Seasonality

The Ring Deposit sample included 241 valves. Of these, I was able to determine the terminal growth increment for 199 valves (82.6 percent of the sample). The vast majority (87.4 percent) of clams were experiencing opaque growth at the time of death. This indicates that these clams were collected primarily during the cool weather months of winter and spring.

152 Table 7.27. Age Data for Ring Deposit Levels 1‑5 and 6‑12.

Age Levels 1‑5 Levels 6‑12 Count 90 96 Mean 4.11 4.28 Median 4 4 Mode 3 4 Std Dev 1.94 1.82 Min 2 2 Max 11 11

Table 7.28. Growth Phases for the Ring Deposit.

Growth Phase Count Percent O1 71 39 O2 41 22.5 O3 47 25.8 T1 8 4.4 T2 4 2.2 T3 11 6

I was able to successfully identify growth phases for 182 valves (75.5 percent) in the Ring

Deposit collection. The percentage of readable clams was similar to that of the modern collection and was much higher than the Ring Interior sample. The clams from the Ring

Deposit were better preserved than those in the Ring Interior, resulting in a higher percentage of clams that could be interpreted. Bleaching, which was fairly common in the Ring Interior sample, was absent from the Ring Deposit sample. A total of 59 valves, or 24.5 percent of the sample, was unreadable. Thirty valves (12.4 percent) were identified as senescent or probably senescent, and none of these could be interpreted. An additional twenty‑nine clams (12.0 percent) could not be read because they exhibited unclear growth patterns or were fragments that did not possess enough growth increments (multiple increments are needed for comparison with the terminal growth phase).

153 Ring Deposit

40 35 30 25 20

Percent 15 10 5 0 O1 O2 O3 T1 T2 T3

Figure 7.40. Growth Profile for Clams in the Ring Deposit. This graph depicts the distribution of growth phases for the Ring Deposit. This growth profile most resembles that of a winter collection with limited gathering in the spring.

The clams from the ring deposit exhibited a clear pattern of seasonal exploitation. The majority of clams were in the O1 growth phase, followed by O3, and O2 (Table 7.28). Very few clams were experiencing translucent growth at their time of death. This growth profile (Figure

7.40) most closely resembles a winter collection period with some additional gathering in the spring. The growth profile does not indicate that clams were gathered equally throughout the winter and spring. Clams from the ring deposit were gathered primarily in the winter with some clams gathered during the spring. In order to investigate any change over time in seasonal clam use, I compared the growth profiles for the upper levels (and 1‑5) and lower levels (6‑12) (Table 7.29 and Figures 7.41 and 7.42). The growth profiles are quite similar and both indicate predominantly winter collections. In the lower levels, clam collection appears to be slightly more constricted, occurring primarily in the winter with very limited collection during the spring. In the upper levels, collection occurred throughout the winter with some collection during the spring. The upper levels evidenced slightly more spring gathering than the lower levels, indicated by a higher percentage of O3 and T1 growth phases. Although there

154 Table 7.29. Growth Phases for Levels 1‑5 and 6‑12 in the Ring Deposit.

Growth Phase Levels 1‑5 (N) Levels 6‑12 (N) Levels 1‑5 (%) Levels 6‑12 (%) O1 31 40 34.8 43.0 O2 19 22 21.3 23.7 O3 23 24 25.8 25.8 T1 5 3 5.6 3.2 T2 3 1 3.4 1.1 T3 8 3 9.0 3.2 Totals 89 93 99.9 100

Ring Deposit Levels 1‑5

35 30 25 20 15 Percent 10 5 0 O1 O2 O3 T1 T2 T3

Figure 7.41. Growth Profile for the Ring Deposit Levels 1‑5.

155 Ring Deposit Levels 6‑12

45 40 35 30 25 20 Percent 15 10 5 0 O1 O2 O3 T1 T2 T3

Figure 7.42. Growth Profile for the Ring Deposit Levels 6‑12.

Table 7.30. Age Distribution for the West Arm, Ring Interior, and East Arm of the Ring.

Age West Central East Count 86 78 21 Mean 4.5 3.99 3.57 Median 4 4 3 Mode 3 4 3 Std Dev 2.14 1.66 0.87 Minimum 2 2 2 Maximum 11 9 5

was slightly more spring gathering in the upper levels, the difference is very minor and the overall gathering pattern is essentially unchanged throughout the deposit.

Spatial Comparisons of the Ring Deposit

Three areas of the ring had the following distribution of clams selected for analysis: West Arm =

111 (46.1 percent); Central Portion = 90 (37.3 percent); East Arm = 39 (16.2 percent). Age distributions for the three areas were not straightforward (Table 7.30). The mean age was highest in the West Arm (4.5 years) and declined moving eastward around the ring (Central

156 Table 7.31. Hinge Length Measurements (in millimeters) for Ring Deposit Areas.

Hinge West Central East Count 79 71 20 Missing 32 19 19 Mean 47.54 48.27 48.22 Median 47.05 48.77 47.6 Std Dev 7.32 8.65 6.55 Minimum 31.5 27.46 36.35 Maximum 72.24 66.24 59.37

Table 7.32. Growth Phase Distributions for Areas of the Guana Shell Ring.

Phase West (n) West (%) Central (n) Central (%) East (n) East (%) O1 71 39 71 39 71 39 O2 41 22.5 41 22.5 41 22.5 O3 47 25.8 47 25.8 47 25.8 T1 8 4.4 8 4.4 8 4.4 T2 4 2.2 4 2.2 4 2.2 T3 11 6 11 6 11 6

Portion mean = 3.99 years and East Arm mean = 3.57 years). The median and mode did not follow this pattern. The size of clams in each area was very similar (Table 7.31). Mean hinge length for each area is as follows: West Arm = 47.54 mm; Central Portion = 48.27 mm; East Arm

= 48.22 mm.

Table 7.32 summarizes the frequency of growth phases for each area of the ring. The

West Arm indicates that gathering occurred in the winter and, to a lesser extent, during the spring. The growth profiles for the upper and lower levels of the West Arm were quite similar.

Lower levels demonstrated that clam collection occurred primarily during the winter and in the spring to a much lesser extent. Upper levels indicate collection during the winter and to a lesser extent in the spring, but also exhibit an unusually high percentage of the T3 phase. The T3 phase is most common during the summer and fall, but the lack of other translucent phases indicates that these clams were probably not gathered during the summer or fall. During the cooler months the T3 growth phase is more common in the winter than the spring, and it is

157 likely that these clams were gathered then. Clams in the Central Portion were gathered during the winter. Seasonality was essentially identical in the upper and lower levels of the Central

Portion. Clams in the East Arm were gathered during the winter and spring with an emphasis on spring collection. The East Arm deposit was too small to divide into upper and lower levels, and the majority of usable clams were in level 5.

In order to explore any fine‑grained differences in seasonal collection strategies throughout the ring deposit, I also examined each shovel test and test unit as a separate deposit.

Unfortunately, the number of readable clams for each shovel test rarely reached more than 30 clams (the minimum sample size needed to be statistically significant). Although the sample sizes are small, I have summarized the season of collection for the test unit and four shovel tests that contained at least 20 clams (Figure 7.43). Unfortunately, the small sample sizes preclude the identification of fine‑grained spatial seasonal use of the ring (beyond the three areas of the ring). The seasons of collection identified in each shovel test and test unit are identical to the overall pattern for the area in which they are located with one exception. The shovel test located at 340N, 410E (in the West Arm) demonstrated exclusively winter collection, rather than predominantly winter collection with some additional spring gathering.

I identified high percentages of the T3 growth phase in shovel test 440N, 410E. This shovel test contained 33 clams for which growth phase could be determined, six (18.2 percent) of which were in the T3 growth phase (Table 7.33). T3 was the only translucent growth phase identified in this sample, thus ruling out the possibility of a summer or annual collection. In the

St. Augustine collection, T3 phase clams are most common in the fall but can be found during any season of the year. It is possible that the T3 clams were collected at the end of fall, but given the distribution of opaque growth phases (and overall lack of other translucent phases) it seems more likely that these T3 clams were collected during the winter (perhaps the early winter).

Summary of Results

Based on the isotopic analysis, the modern clams from the St. Augustine collection and the archaeological clams from the Guana shell ring are comparable. Opaque and translucent

158

Figure 7.43. Distribution of Seasonality throughout the Ring Deposit. Site map adapted from Russo et al. 2002. This image depicts the horizontal variability in seasons of clam collection throughout the ring. Only samples of at least 20 clams are included. Blackened areas of the circles represent seasons that are represented.

159 Table 7.33. Distribution of Growth Phases for Shovel Test 440N, 410E.

440N, 410E Count Percent O1 13 39.4 O2 5 15.2 O3 9 27.3 T1 0 0 T2 0 0 T3 6 18.2 Total 33 100.1

increments for both collections formed consistently within similar ∂18O value ranges: opaque increments formed primarily during the cooler months and translucent increments formed primarily during the warmer months. Increments were deposited on a seasonal basis and occurred only once per year. Although no additional yearly increments were identified, false increments occurred in both archaeological and modern clams. Isotopic analysis helped to clarify these and other unclear growth patterns, and facilitated the accurate analysis of a larger number of clams.

Water data revealed that temperature has a more prominent effect than salinity on the seasonal growth of clams. Water temperatures varied seasonally, but salinity remained relatively constant throughout the year. Seasonal peaks of increments followed the expected patterns based on temperature data. Growth phase distributions also tracked seasonal changes in water temperatures.

The St. Augustine collection was composed of 411 clams. On average, clams had a hinge length of 44.32 mm and were 6.26 years old. The oldest clam in the collection was 19 years old, and the youngest clam was 2 years old. Clams from Area A were on average one year older than clams from Area B. Area A is a protected aquaculture lease site, although illegal gathering still occurs as well as occasional gathering of wild clams by the leaseholder. Area B is an area that is accessible by foot and is open to commercial and recreational gathering.

I was able to identify the terminal growth increment for 91.5 percent of clams from the

St. Augustine collection. Based on average temperatures of the Matanzas River, I expected

160 winter to contain primarily opaque (fast) growth, spring to contain primarily opaque growth but increasing amounts of translucent (slow) growth, summer to contain primarily translucent growth, and fall to contain predominantly translucent growth with increasing opaque growth.

Growth increment frequencies for the St. Augustine collection demonstrate this pattern exactly as hypothesized. Although seasonal peaks of increments are somewhat distinct, they are not sufficiently distinct to allow for the identification of seasons (winter, spring, summer, and fall) based solely on the terminal increment. Instead, the St. Augustine collection supports the use of terminal increments to identify more general patterns of cool weather collection (in the winter and spring) and warm weather collections (in the summer and fall).

I was able to identify the terminal growth phase for 81.3 percent of the St. Augustine collection. I was not able to identify the terminal growth phase for the remainder of the collection because clams were senescent or had unclear growth profiles. Monthly growth phase profiles were not sufficiently distinct to facilitate the identification of individual months of collection. Seasonal growth phase profiles were sufficiently distinct to allow for the identification of seasonal collections. Each season demonstrated a unique peak and distribution of growth phases that can be used to identify seasons of collections for archaeological clams.

The annual growth profile contains clams in all phases of growth and is sufficiently distinct to identify annual collection in an archaeological assemblage.

The assemblage of analyzed clams from the Ring Interior consisted of 155 valves. Clams from this collection demonstrated a mean hinge length of 50.75 mm and an average age of 5.67 years. The oldest clam in the collection was 20 years old and the youngest clam was 2 years old.

From the lower levels of the deposit to upper levels average age decreased by just over half a year.

Clams from the Ring Interior were in poor condition compared to the Ring Deposit and

St. Augustine collection. I was able to identify the terminal growth increment for 67.7 percent of the Ring Interior. The vast majority of these clams were experiencing the opaque growth increment at the time of their death, indicating that collection took place primarily during the cooler months of winter and spring. I was able to identify the terminal growth phase for 50.3 percent of the Ring Interior assemblage. The growth phase distribution indicates that clam

161 collection occurred primarily during the spring. Upper and lower level deposits exhibited the same pattern of collection.

The assemblage of analyzed clams from the Ring Deposit consisted of 241 valves.

Clams from this collection demonstrated a mean hinge length of 47.92 mm and an average age of 4.2 years. The oldest clam in the Ring Deposit was 11 years old and the youngest clam was 2 years old. From the lower levels of the deposit to upper levels average age decreased slightly

(less than one quarter of a year).

I was able to identify the terminal growth increment for 82.6 percent of the Ring Deposit.

Like the Ring Interior, the vast majority of these clams were experiencing the opaque growth increment at the time of their death, indicating that collection took place primarily during the cooler months of winter and spring. I was able to identify the terminal growth phase for 75.5 percent of the Ring Deposit. The growth phase distribution indicates that clam collection occurred predominantly during the winter with some additional gathering during the spring.

Growth profiles for upper and lower levels were very similar. I observed some horizontal variability in seasonal collection, which will be discussed further in Chapter 8. The following chapter provides comparisons of these findings to other modern comparative collections and archaeological assemblages.

162 CHAPTER 8

DISCUSSION

St. Augustine Collection

Seasonal Profiles

The St. Augustine collection is comprised of clams collected from every month in 2010. The seasonal profiles I established are sufficiently distinct to permit identification of seasons of death for archaeological clams in the St. Augustine area. The St. Augustine collection lies between two modern clam collections – Kings Bay, Georgia, which is approximately 125 km to the north and Indian River, Florida, which is approximately 150 km to the south. The St.

Augustine collection is approximately 40 km south of the Guana shell ring (Kings Bay is approximately 85 km north of Guana and Indian River is approximately 190 km to the south of

Guana). The Kings Bay collection was gathered over the course of two years and included 451 assessed valves (Quitmyer et al. 1997). The Indian River collection was gathered over the course of several years and included 1100 assessed valves (Quitmyer 1995; Quitmyer et al.

1997).

Based on tidal charts and ecology, I characterized the St. Augustine area as a transitional area between the tide‑dominated salt marshes of coastal Georgia and the wave‑dominated coasts and mangrove swamps of Central and South Florida. As such, I hypothesized that the St.

Augustine collection would demonstrate some similarity with modern clam collections to the north and south, but would have a distinct group of seasonal growth profiles. A comparison of the three growth profiles shows this hypothesis to be correct (Figure 8.1). The St. Augustine collection shares characteristics with both the Kings Bay and Indian River collections, although overall it is slightly more similar to the Indian River collection. The fact that the St. Augustine

163 Kings&Bay& ! !!!Winter!! !!!!!!!!!Spring!!! !!Summer!! !!!!!!!!!!!!Fall!!! !!Annual! 92.62 85.22 662 512 56.52 25.82 16.32 14.82 19.12 6.22 11.52 10.62 6.72 7.42 11.82 2.12 02 02 3.82 02 02 02 02 02 02 02 02 5.122.92 4.72

O12O22O32 T12 T22 T32 O12O22O32T12 T22 T32 O12O22O32T12 T22 T32 O12O22O32T12 T22 T32 O12O22O32T12 T22 T32 !!!!! ! ! St.&Augustine& ! !!!Winter!! !!!!!!!!!Spring!!! !!Summer!! !!!!!!!!!!!!Fall!!! !!Annual!

66.32 47.62 35.42 41.72 25.62 29.12 29.8228.62 27.82 17.12 21.52 23.62 19.82 18.62 9.82 7.62 11.4212.62 9.92 02 02 3.822.52 02 02 02 02 02 4.525.62

O12O22O32T12 T22 T32 O12O22O32T12 T22 T32 O12O22O32T12 T22 T32 O12O22O32T12 T22 T32 O12O22O32T12 T22 T32 !!!!! ! ! Indian&River& ! !!!Winter!! !!!!!!!!!Spring!!! !!Summer!! !!!!!!!!!!!!Fall!!! !!Annual!

39.92 39.32 35.32 33.32 24.12 24.12 302 282 212 222 16.92 182 17.82 18.5217.62 21.92 13.5210.22 7.62 9.52 112 13.12 4.62 4.723.42 1.424.42 4.32 12 3.42

O12O22O32T12 T22 T32 O12O22O32T12 T22 T32 O12O22O32T12 T22 T32 O12O22O32T12 T22 T32 O12O22O32T12 T22 T32 !!!!!

Figure 8.1. Seasonal Growth Profiles for the Kings Bay, St. Augustine, and Indian River Modern Comparative Collections. This image depicts the seasonal growth profiles for the three modern comparative collections near the Guana shell ring. Data for the Kings Bay collection is from Quitmyer et al. 1997, data for the St. Augustine collection is from this dissertation, and data for the Indian River collection is from Quitmyer et al. 1997.

164 collection resembles the Indian River collection more than the Kings Bay collection is especially interesting given the greater distance between the Indian River and St. Augustine collections.

The similarities between the St. Augustine collection and the Indian River collection may be a result of the presence of the southern quahog, Mercenaria campechiensis, and hybrids of the southern quahog and the northern quahog (Mercenaria mercenaria). Arnold and colleagues’

(1998) study revealed minor species‑specific differences in the timing of the onset of seasonal growth. This variation results in greater variability of growth phases in each season (as seen in the Indian River and, to a lesser extent, in the St. Augustine collection). Nevertheless, Arnold and colleagues (1998:106) concluded that environmental factors (namely temperature) were the dominant influence on seasonal patterns of growth. In other words, incremental growth studies examining M. mercenaria, M. campechiensis, and their hybrids are acceptable because seasonal temperatures are the primary influence on seasonal growth, regardless of the species. In sum, the St. Augustine collection shares similarities with both the Kings Bay and Indian River collections as a result of a gradually changing habitat, latitude, and seasonal influences.

The St. Augustine collection seems to split many of the differences between the Kings

Bay and Indian River collections. Essentially, variability of growth phases within a season increases as one moves from north to south (i.e., the most southern collection, Indian River has the greatest number of growth phases per season with the smallest peaks in a particular growth phase). The Kings Bay collection is characterized by high percentages of one growth phase in each season (the smallest peak percentage is 51 percent), while the Indian River collection is characterized by the presence of many growth phases with more modest peaks in each season.

The St. Augustine collection includes clams in numerous growth phases for each season, and seasonal peaks in phases that are slightly larger than the Indian River peaks but smaller than the Kings Bay peaks. Winter and spring profiles resemble the Indian River collection, but the summer and fall profiles more closely resemble the Kings Bay collection. Like the Kings Bay collection, the summer St. Augustine profile contains clams exclusively in the translucent increment, whereas the Indian River summer collection contains clams in all phases except O1.

The St. Augustine fall collection more closely resembles the Kings Bay collections, with

165 concentrations in the T3 and O1 growth phases, rather than a nearly even distribution of translucent growth phases (Indian River).

Although the modern collections are similar, the use of either the Kings Bay or Indian

River collections to identify the seasons of clam collection at Guana would result in incorrect interpretations. For example, if I had used the Kings Bay collection to interpret a collection of clams with a peak in the T1 growth phase, I would have assigned this to a spring, rather than a summer collection. If I had used the Indian River collection to interpret a collection of clams with a very high percentage in the T3 phase, I might have classified it as a very brief late fall through early winter collection, rather than a fall collection. The St. Augustine collection demonstrates the necessity of using a modern comparative collection that is in close proximity to the archaeological site in question.

I also hypothesized that a collection between the two extant collections could shed light on the extent of variation in seasonal onset caused by changes in latitude (and changing ecosystems). I hoped that these data would be useful for determining acceptable distances between the modern comparative collection and the archaeological site in question. Based on the differences and distance between the Kings Bay and St. Augustine collections, it is clear that

125 km is too far to accurately employ a modern comparative collection. This research has indicated that using the Kings Bay collection to identify the clam seasonality at Guana would provide incorrect results (although not drastically incorrect results). This means that 85 km is slightly too far to accurately employ a modern comparative collection. It appears that the maximum acceptable distance between the modern collection and archaeological site in question is somewhere between 45‑80 km (Quitmyer and Jones [1992] found that collections in

Charlotte Harbor that were approximately 40 km apart were similar). Based on the present results, the maximum acceptable distance may be somewhere around 70 km, but this remains to be tested. The fact that the Indian River collection is more similar to the St. Augustine collection than the Kings Bay collection (despite the closer proximity of St. Augustine and Kings Bay) also exemplifies that clam growth and initiation of growth phases is not simply a product of latitude

– ecosystem comparability is an important factor in seasonality studies that should not be ignored. In sum, distances around 40 km are acceptable for accurate identification of seasonal

166 growth profiles, but distances of more than 70 km are likely too great for accurate seasonality assessment. Furthermore, careful attention should be paid to the overall climate and habitat as well as the makeup of quahog species at archaeological sites and at the location of the modern comparative collection.

Age Data

Age data of clam assemblages can provide an indicator of the amount of pressure on local clam beds (Quitmyer et al. 1985a; Quitmyer 1995). Pressure on clam beds can come from natural factors as well as from human predation. Age data are currently available for two collections near the St. Augustine collection: Kings Bay (Quitmyer et al. 1985a and b, 1997) and Indian

River (Quitmyer 1995; Quitmyer et al. 1997). The Kings Bay, St. Augustine, and Indian River collections are subjected to very different amounts of human predation pressure. Commercial and recreational clam fishing has been prohibited at Kings Bay at least since the 1970s when the

Naval Base was built (though prohibitions probably date back to the 1950s when the U.S. Army began acquiring the land). Human predation in Kings Bay is nonexistent. The St. Augustine collection was gathered from two areas with slightly different human predation pressures, but the vast majority of clams were gathered from an area where commercial and recreational shellfish gathering is permitted and fairly frequent. Human predation pressure in the St.

Augustine collection area is considered moderate. The Indian River has been a hotspot of commercial clam fishing since the 1980s and provides the majority of Florida’s quahog landings

(MacKenzie et al. 2002b). Human predation on Indian River clams is considered heavy.

The three collections show rather different age profiles, which are probably the result of the varying degrees of human predation at each area (Table 8.1). The mean age for each collection decreases with increased human predation: Kings Bay has the highest mean age, St.

Augustine has a lower mean age, and Indian River has the lowest mean age. On a smaller scale this pattern is further demonstrated in the two areas of exploitation for the St. Augustine collection. Collection Area A was in a protected lease area where gathering of wild clams was infrequent. Human predation pressure in Area A is considered light. Collection Area B was in

167 Table 8.1. Mean Ages for Modern Comparative Clam Collections

Modern Collection Human Predation Mean Age Std Dev Sample Size Kings Bay Nonexistent 10.5 6.4 65 St. Augustine Moderate 6.3 3 358 Indian River Heavy 5.3 2.9 687

a location that was open to both commercial and recreational fishing and where harvesting wild shellfish was a fairly common activity. Human predation pressure in Area B is considered moderate. As discussed in Chapter 7, on average Area A clams were over a year older than

Area B clams. The Kings Bay, Indian River, and St. Augustine modern collections demonstrate that human exploitation can significantly alter the age structure of clam populations.

Clam Collection at the Guana Shell Ring

Population Structure of Clams from the Guana Shell Ring

Archaeological clams from the Guana shell ring show increasing predation pressure through time. Over time, clams at Guana became progressively younger (Table 8.2). Saunders and

Rolland (2006:49) considered the basal portion of the ring interior midden to be older than the shell ring, but it appears that this deposit may be younger than the ring deposit. Clams from the lower levels (4‑6) of the Ring Interior assemblage had a mean age at death of 5.98 years

(standard deviation= 3.32). This assemblage also provided the oldest clam in the collection – a twenty year‑old clam. The upper levels (1‑3) of the ring interior are assumed to be more recent than the lower levels based on stratigraphy. Clams in the upper levels of the Ring Interior assemblage had a decreased mean age at death of 5.39 years (standard deviation= 2.18).

Saunders and Rolland (2006:46) argued that the ring deposit began ca. 3300‑3500 1 cal B.P., although Russo and colleagues (2001:21) argued that ring deposition began ca. 3500‑3600 based on uncalibrated radiocarbon dates. Clams from lower levels (6‑12) of the ring deposit had a mean age at death of 4.28 years (standard deviation= 1.82), showing a considerable decline in average age from the ring interior deposit. Clams from the upper levels (1‑5) of the ring are

168 Table 8.2. Age Data for Archaeological Clams from the Guana Shell Ring. This table shows the decline in age of archaeological clams from Guana. Compare with table 8.1, which shows the age data for the nearest modern clam collections.

Minimum Maximum Sample Guana Context Mean Age Std Dev Age Age Size Ring Interior (Lower) 6.0 3.3 2 20 50 Ring Interior (Upper) 5.4 2.2 2 17 54 Ring (Lower) 4.3 1.8 2 11 96 Ring (Upper) 4.1 1.9 2 11 90

assumed to be more recent than lower levels of the ring. These clams have a slightly decreased mean age at death of 4.11 years (standard deviation=1.94).

If the ring interior midden is older than the ring deposit, the overall decline in age may be a gradual decline over time. If the ring interior midden is younger than the ring deposit, the decline in age may be the result of heavy exploitation during ring deposition, followed by a time of less intense clam collection or brief site abandonment. The clams in the ring interior deposit may indicate a slight rebound in the age structure of the clam population, followed be a gradual decline throughout the deposit of the ring interior midden. More radiocarbon dates from the ring interior midden (especially basal deposits) are needed to determine the exact chronology of the ring interior midden and the ring deposit. Regardless, a gradual decline in age is seen in both deposits.

The overall decline in age through time indicates increasing pressure on nearby clam beds. While this increased pressure may be the result of natural factors, the gradual decline over time suggests that the decline is likely the result of increased human predation (see

Quitmyer et al. 1985a:37‑38). Quitmyer and Jones (2000) identified overexploitation of clams at five sites in coastal Georgia and Florida based on clam ages. They argued that this overexploitation seems to occur in places where human populations are dense and sedentary and where clams are an important part of the diet (Quitmyer and Jones 2000:165). The heavy predation pressure indicated by the decline in age at Guana might be the result of increased sedentism, increased population, increased use of clams rather than other resources, or may be

169 the result of continued heavy exploitation of clams over the course of occupation.

Quantification of both vertebrate and invertebrate use at the Guana site is needed to tease apart these possibilities. The depressed mean age at death of clams from the shell ring suggests heavy exploitation of clams and rather severe pressure on local clam beds. Guana clams from both deposits (ring interior and ring deposit) yield a mean age at death of 4.73 (standard deviation=2.35). The mean age at death declined over time at Guana, eventually reaching a low of 4.28 years old. The Indian River modern collection was subject to heavy human predation pressure from commercial and recreational clam fishing and yielded a mean age at death of 5.3 years. Based on mean age at death, it appears that Guana residents put more pressure on their clam beds than modern clam fishers in the Indian River.

The size of clams at Guana also steadily declined over time (Table 8.3). Mean hinge length of clams from the lower levels of the ring interior was 51.41 mm (standard deviation=

8.1). Mean hinge length of clams from the upper levels of the ring interior declined slightly to

50.12 mm (standard deviation= 8.03). Like age, the mean hinge length of clams from the ring deposit was considerably lower than the ring interior: mean hinge length in the lower levels of the ring deposit was 48.16 mm (standard deviation= 7.19). Mean hinge length declined from the lower ring levels to the upper ring levels. Mean hinge length in the upper levels of the ring deposit was at its lowest at 47.68 mm (standard deviation= 8.41).

Within a clam population, clam size varies greatly (Fegley 2001). At Guana, a pattern of decreasing size through time is apparent, indicating increased pressure on clam beds. The

Table 8.3. Size Data for Archaeological Clams from the Guana Shell Ring. This table shows the decline in mean hinge width (in mm) over time.

Mean Standard Minimum Maximum Sample Guana Context Hinge Width Deviation Hinge Hinge Size Ring Interior (Lower) 51.41 8.1 35.44 71.76 56 Ring Interior (Upper) 50.12 8.03 36.87 67.82 59 Ring (Lower) 48.16 7.19 31.88 64.76 87 Ring (Upper) 47.82 8.41 27.46 72.24 83

170 declining sizes at Guana indicate that human predation was extensive enough to impact overall clam size as well as age. This type of decline in age and size of clams has also been identified in the Mosquito Lagoon at Canaveral National Seashore, Florida (Parsons n.d.). The combination of reduced age and size of clams at Guana indicates that the steady decline is not the result of occupants selecting a specific age or size class. The decrease in clam size means that meat yields from clams also decreased over time (although probably not dramatically). Based on the evidence of decreased age and size, it appears that Guana residents gradually overexploited clams resulting in decreased yields.

Patterns of Seasonal Clam Exploitation at the Guana Shell Ring

Guana residents gathered clams exclusively during the cool months of winter and spring.

Clams from the ring interior were gathered predominantly during the spring with little, if any, gathering during the winter. Seasonality of clams in the upper and lower deposits was essentially the same. Clams from the ring deposit, however, were gathered primarily during the winter with some limited gathering during the spring. Clams in the upper and lower levels of the ring deposit demonstrated similar seasonality, although the collection time in the lower levels appears to be slightly more restricted (less spring collection). It is currently unclear why gathering practices are different in the ring interior and the ring itself. Perhaps Guana occupants began focusing on other resources during the spring once accumulation of the ring began. In order to test this hypothesis, a detailed analysis of vertebrate and invertebrate fauna at Guana must be undertaken. Next year, I intend to conduct an in‑depth analysis of the vertebrate fauna collected from the Guana excavations. This analysis should provide evidence with which to evaluate my hypothesis of changing resource exploitation.

To evaluate whether the entire circumference of the Guana shell ring was deposited simultaneously throughout the year or if portions of the ring were used during different times of the year, I examined samples from multiple areas of the ring. Each area had a sample of more than 30 readable clams, although density of the deposit varied greatly in areas (the East

Arm deposit is much thinner and more sparse than the other ring areas). I observed some slight variability in the season of clam deposition into the ring. The Central Portion clams were

171 deposited during the winter, the East Arm clams were deposited primarily during the spring but also in the winter to a lesser extent, and the West Arm clams were deposited predominantly in the winter with some deposition during the spring as well (spring was more prevalent in the upper levels than the lower levels of the West Arm). Examination of individual shovel tests and test units revealed essentially the same patterns of seasonal use as those demonstrated by the ring areas. I had hoped that horizontal variability in the seasonality of clam deposits might provide evidence on how the ring functioned. Unfortunately, seasonality of clam deposits is not markedly different in various areas of the ring, providing little evidence with which to evaluate ring function.

It is difficult to decipher the meaning of the variability in seasonal deposits at Guana without an understanding of how this ring was used. The function of shell rings has remained a debated topic in archaeology. Three models have been proposed: 1) a gradual accumulation model that argues rings are comprised of occupation refuse that is the result of primarily quotidian affairs (e.g., Cable 1997; Thompson 2006; Trinkley 1985; Waring and Larson 1968); 2) a ceremonial model that argues rings are the result of ceremonial feasting and are composed primarily of feasting refuse (Russo 2004a; Russo and Saunders 1999); 3) a developmental model that argues that the ring shape initially formed as a result of habitation, but gradually developed into a site with a ceremonial meaning that may have functioned less as a residence and more as a place of ceremony and ritual (Thompson 2007).

If, as some have suggested, rings are the result of habitation near, on, or within the ring, variable seasonal deposition could be the result of differences in resource use by households or other groups. If this is the case, it is possible that occupants of the East Arm of the ring were fewer in number or exploited less shellfish, resulting in an overall sparser deposit in the East

Arm. If the ring is the result of ceremonial activity at the site, the distribution of remains and variability in seasonal clam use is more difficult to understand. Russo (2004a) has used social space theory to identify the central portion of the ring as a place of high status during social events. However, the central portion of the ring produced the most constricted seasonal consumption of clams (exclusively during the winter), while both arms of the ring displayed winter and spring consumption of clams (with concentration in opposite seasons). Biologically,

172 clams experience similar processes during the winter and spring and are not thought to have different consistencies or taste until the summer and fall when clams are spent or recovering from spawning (see Chapter 5). Therefore biological or taste preferences do not seem to indicate a reason for different preferences or status of clams between the winter and spring.

Furthermore, the density of the East Arm deposit seems to indicate that it was used to a lesser extent than other areas of the ring. This means that the central portion of the ring may not have been a center‑point of focus and status used during ceremonies or special events. The developmental model is difficult to evaluate with regard to variability in seasonal deposits.

However, the use of a small midden and the transition to the deposition of refuse in a large ring shape indicates changing patterns of refuse disposal that could be related to use of the site. This change could have been a shift in the arrangement of residences or a change in the way in which refused was disposed, whether from habitation or ceremonial activities.

Highly seasonal deposits indicative of feasting are lacking from Guana. The seasonal use of clams is probably a result of continued use of the site and repeating yearly cycles of seasonal exploitation. Large‑scale excavation of the ring could provide sizeable samples with which to evaluate cyclical seasonal use of clams through levels of deposition, but current samples do not permit this kind of evaluation. Based on the evidence at hand, it seems likely that variation in seasonal clam use may have been the result of household differences in clam consumption, rather than seasonal differences of serving practices during ceremonies. Analysis of vertebrate fauna and the addition of seasonality studies of other resources (especially fish and oysters) could provide additional evidence with which to evaluate use of the ring and its overall function.

Contextualizing the Seasonal Exploitation of Clams at the Guana Shell Ring

To contextualize the seasonal use of clams at the Guana shell ring site, I have examined published clam seasonality data for fourteen sites in the Georgia Bight. I have compared these sites to evaluate patterns in exploitation across different areas and time periods. Seasonality studies are very sensitive to sample size (Monks 1981) because the genus Mercenaria exhibits variation in the onset of seasonal indicators (Arnold et al. 1998; Jones and Quitmyer 1996).

173 Quitmyer and colleagues (1997) argued that samples larger than 30 are generally acceptable to accurately identify seasons of death in an assemblage. Therefore, I have limited the selection of sites to those with samples of at least 30 clams. Table 8.4 lists all the sites for which seasonality

(both vertebrate and invertebrate) is discussed in this chapter. Table 8.5 lists the outcomes of clam seasonality studies at selected sites in the Georgia Bight. The examined sites range from the Late Archaic to the Mississippi period. At sites where multiple samples were assessed, I have combined them into one overall collection pattern (unless different time periods were represented). This was done to bolster the size of the samples and to examine broad patterns of seasonality. Seasonality assessments at nearly all of the sites examined make use of the Kings

Bay comparative collection. Clam seasonality assessments at two sites, the St. Catherines shell ring and the McQueen ring were based on the St. Catherines comparative collection, which only permits differentiation between cool weather (winter and/or spring) and warm weather

(summer and/or fall) exploitation (i.e., Quitmyer and Jones [2012] identified cool weather collection at these two sites, but were unable distinguish winter from spring collection).

Clam exploitation at the Guana shell ring fits into a general pattern of winter‑spring clam exploitation seen at most Late Archaic shell rings examined. Clam seasonality has been assessed at four Late Archaic shell ring sites in the Georgia Bight: St. Catherines ring (Quitmyer and Jones 2012), McQueen ring (Quitmyer and Jones 2012), Sapelo ring complex (Thompson and Andrus 2011), and Cannons Point Ring (Quitmyer et al. 1985). Studies at these sites indicated that clams were exploited exclusively during the winter and spring with the exception of the Sapelo ring complex. Clams from the three Sapelo rings were evaluated through both incremental growth and isotopic analysis, which indicated that collection occurred throughout the year (Thompson and Andrus 2011). Thompson and Andrus (2011) also evaluated oyster seasonality through oxygen isotopic analysis, and found that oysters were gathered during all seasons except spring. Additionally, the only known Mississippi period shell ring (Grand) evidenced clam gathering exclusively during the winter and spring (Parsons 2008).

Clam seasonality data were available for only a single Woodland period site: Kings Bay

(Quitmyer et al. 1985a and b). Kings Bay provided evidence for year‑round clam collection,

174 Table 8.4. Seasonality Assessments of Selected Archaeological Sites in the Georgia Bight.

Site Site No. Time Period Sources St. Catherines Ring 9LI231 Late Archaic Colaninno 2010; Quitmyer and Jones 2012 McQueen Ring 9LI648 Late Archaic Colaninno 2010; Quitmyer and Jones 2012 Sapelo Ring Complex 9MC23 Late Archaic Thompson and Andrus 2011 Cannons Point Ring 9GN57 Late Archaic Colaninno 2010; Quitmyer et al. 1985a and b; Marrinan 1975 West Ring 9GN76 Late Archaic Colaninno 2010; Marrinan 1975 Guana Ring 8SJ2554 Late Archaic This dissertation Kings Bay 9CM171 Woodland Quitmyer et al. 1985a and b Meetinghouse Field 9LI21 Mississippi Saunders and Russo [1986] 2010 Devils Walkingstick 9CM177 Mississippi Quitmyer et al. 1985a and b Harris Neck 9MC141 Mississippi Braley et al. 1986 Harrison Homestead 8NA41d Mississippi Russo et al. 1993 Crane Island 8NA705 Mississippi Quitmyer et al. 1990 Planted Pine Midden 8DU7499 Mississippi Russo et al. 1993 Grand Ring 8DU1 Mississippi Ashley et al. 2007; Parsons 2008 Cedar Point North 8DU64 Mississippi Russo et al. 1993 Jones 8DU7498 Mississippi Russo et al. 1993

Table 8.5. Clam Seasonality Assessments at Selected Sites in the Georgia Bight

Site Winter Spring Summer Fall St. Catherines Ring X X McQueen Ring X X Sapelo Ring Complex X X X X Cannons Point Ring X Guana Ring X X Kings Bay X X X X Meetinghouse Field X X X X Devils Walkingstick X X X X Harris Neck X X X X Harrison Homestead X X Crane Island X Planted Pine Midden X Grand Ring X X Cedar Point North X X X x Jones X X X

175 with particularly heavy exploitation during the spring (Quitmyer et al. 1985b). More Woodland period samples must be analyzed to understand whether Kings Bay represents a typical

Woodland clam gathering strategy. These samples may be difficult to locate, however. Some time around the end of the Late Archaic, it appears that sea level retreated several meters, resulting in changes in the estuarine environment and resource availability (DePratter 1977;

DePratter and Howard 1980; Thomas 2008c). People may have moved their settlements to be closer to remaining or newly developing marshes. These areas of low‑stand marsh are now buried beneath two meters of marsh deposition that followed later sea level rise (Thomas

2011:38).

Nine Mississippi period sites demonstrated a general pattern of exploitation during the spring and often the winter; summer and fall collections are represented to a lesser extent.

Three sites were located in Georgia: Meetinghouse Field (Saunders and Russo [1986] 2010),

Devils Walkingstick (Quitmyer et al. 1985a and b), and Harris Neck (Braley et al. 1986). All of these sites demonstrated annual collection with a focus on clam gathering during the spring.

Six Florida sites met the criteria for inclusion: Harrison Homestead (Russo et al. 1993), Crane

Island (Quitmyer et al. 1990), Planted Pine Midden (Russo et al. 1993), Grand shell ring (Parsons

2008), Cedar Point North (Russo et al. 1993), and Jones (Russo et al. 1993). These sites evidenced a minimum of spring collection, often included winter collection, occasionally included summer collection, and one collection may include a fall component. The Cedar Point North site is the only northeast Florida site that may represent year‑round gathering, although the fall season is represented in only a single provenience with a sample of less than 30 clams. However, the majority of proveniences indicated a cool weather collection strategy with a focus on clam gathering during the spring. The Jones site is the only other site in northeast Florida to indicate summer collection, but it too demonstrated a focus on clam gathering during the spring. In sum, the northeast Florida sites show a strong regional pattern of spring and often winter collection with infrequent representation of other seasons.

Although the sample of sites is small, the patterns of seasonal exploitation are compelling. Late Archaic shell rings demonstrate exclusively winter and spring exploitation with the exception of the Sapelo ring complex, which has an overall annual pattern of collection.

176 At Woodland and Mississippi period sites in Georgia, an annual collection strategy with an emphasis on spring gathering is evident. In northeast Florida, a regional pattern of cool weather collection has emerged. One Late Archaic shell ring and four out of six Mississippi period sites indicate exclusively cool weather collection during the spring and often the winter

(the other two sites demonstrated collection during spring and winter but with some additional collection in other seasons). Studies of clam seasonality at Woodland period sites in northeast

Florida are necessary to determine whether this pattern was continuous over several thousand years. Essentially, the seasonality of clam gathering observed at Guana fits in with the general pattern for Late Archaic shell rings and the northeast Florida Mississippi sites – cool weather collection during the winter and spring.

Incremental growth studies for 4 sites and site complexes in eastern Central Florida have been published. These sites include Tomoka Stone (Russo and Ste. Claire 1992), Edgewater mound complex (Russo et al. 1989), Seminole Rest (Quitmyer 1995), and Honeymoon Hill

(Quitmyer et al. 1990). Data compiled from a survey of islands and other small sites in

Mosquito Lagoon near Seminole Rest is forthcoming (Parsons n.d.). Tomoka Stone is a Late

Archaic site which indicated year‑round collection of clams. The Edgewater sites (Mound B and D) are both Woodland period sites and each indicated year‑round gathering (although gathering occurred primarily during the fall and winter). The Seminole Rest sites date primarily to the Woodland period but extend into the Mississippi period. Each site indicated a different seasonal harvesting pattern: Synders Mound (St. Johns I and II) indicated annual collecting, Fiddle Crab Mound (St. Johns I and II) indicated winter‑spring gathering, and

Midden II (St. Johns I) indicated summer‑fall exploitation. The Honeymoon Hill site evidenced exclusively winter collection. The sites in the Mosquito Lagoon survey date to the Woodland and Mississippi period (most date to the Mississippi period, including the largest sample). Data from this survey indicate that clams were gathered primarily during the fall and during the winter to a much lesser extent.

The sites in eastern Central Florida indicate that fall and winter were the peak seasons to exploit clams at most sites. Annual collection was also a common exploitation strategy in this

177 region. This pattern is clearly different from the majority of sites in the Georgia Bight, and may be a result of the overall different ecology of this region.

Explanatory Model for the Seasonal Use of Clams in Northeast Florida

Residents of the Georgia Bight clearly preferred to exploit clams during the cool weather seasons of spring and winter. I have identified a regional pattern of preference for collecting clams in the cooler seasons in northeast Florida based on a sample of seven sites. Several coastal Georgia sites also indicated that clam exploitation occurred predominantly during the cooler parts of the year; even sites that represented year‑round gathering displayed an emphasis or increased amount of exploitation during the spring. Spring collection is always present at sites in the Georgia Bight, and winter collection is nearly always present. Summer and fall collections occur on a much more limited basis. Summer and fall collections were evident at a single Late Archaic shell ring complex, at all Mississippi period Georgia sites, and at a single northeast Florida site (with an additional Florida site evidencing winter through summer collection). The majority of sites in the Georgia Bight reveal a preference for gathering clams during the cooler months of winter and especially the spring. This exploitation strategy was practiced by people with considerably different ceramic types, different languages, and different ethnic identities and it persisted over thousands of years.

There are several possible explanations for the overall seasonal nature of clam gathering in the Georgia Bight (especially northeast Florida). It is unlikely that clams were sought primarily when other vertebrate resources were unavailable. During previous research

(Parsons 2008), I examined the seasonal availability of the most common fish in the assemblage at the Grand shell ring. All of these fish were available in the estuary during the spring and several were available year‑round (Ashley et al. 2007; Parsons 2008). Commonly eaten fish were readily available during the spring, thus obviating a need to seek clams as a resource of last resort. Wild plant foods were certainly an important part of the diet, yet their role in the diet of Georgia Bight occupants is less understood than the role of faunal constituents. Claassen

(1986:34) argued that shellfish might represent an important carbohydrate, particularly in

178 seasons when plant foods were less available. Floral analysis has not been undertaken at the

Guana shell ring and we do not currently know which plants were an important source of food and nutrients. Based on the availability of fish, which often make up the majority of the vertebrate diet in coastal sites (Parsons and Marrinan n.d.), it does not appear that clams were needed because other food items were scarce.

In a previous study, I suggested that the emphasis on cool weather clam exploitation was likely the result of the increased overall biomass (or meat weight) of clams during the winter and spring (Parsons 2008). Clams store nutrients during the winter in preparation for spawning, which results in an overall increase in biomass during the winter and spring until spawning has occurs, at which time clams enter the spent phase and biomass is greatly reduced

(Peterson and Fegley 1986). The increase in biomass prior to spawning occurs in most shellfish species (Waselkov 1987). Peterson and Fegley (1986) found that clam biomass increases during the winter and peaks during the spring, and then declines in the summer and bottoms out in the fall. Once spawning commences (most commonly in the spring‑summer), clams rapidly lose biomass. Spawning also results in a loss of glycogen, further reducing the caloric value of clams

(Quitmyer 1985:29). Furthermore, some have argued that clams are less palatable during and immediately after spawning (Waselkov 1987) and that they have a watery consistency at this time. Native Americans who relied upon clams as a food source almost certainly observed the seasonal changes that occurred in the biomass and consistency of clams. Previously, I proposed that coastal occupants probably planned their yearly gathering strategy to exploit clams when they were most nutritious and had the most meat‑weight during the winter and spring (Parsons

2008).

I still believe that the high biomass of clams during the winter and spring was an important consideration in the seasonal scheduling of clam exploitation. However, during the course of gathering the modern collection, I identified another factor that I believe was more influential in the decision to gather clams during the winter and spring. During the cooler months of winter and spring clams create signs that provide visible indications of a clam’s location. Keyholes, the most commonly identified clam sign, are small keyhole‑shaped holes in the sand through which the clam siphons water and expels waste (see Figures 6.6‑6.8). I

179 observed many keyholes from December through March and I identified other clam signs during April. It is likely that clams begin to make keyholes and other signs sometime in

November, and cease producing them in April. I very rarely observed keyholes or other signs during the remainder of the year.

By identifying clam signs (especially keyholes), one is able to drastically reduce the search time for clams. Signing, the use of signs to locate clams, has been used as a location tool by commercial foragers for some time and is commonly practiced in Virginia and North

Carolina (MacKenzie et al. 2001:653). Signing can be used to identify clam locations on intertidal flats and in clear, shallow water where the signs can be readily observed. A skilled forager can locate more than enough clams for one day in less than half an hour of searching by identifying clam signs. In February, Mr. Cubbedge found approximately 100 clams in about 25 minutes. After a year of observing Mr. Cubbedge, I was quite adept at locating clams based on signs, and I was able to locate large quantities of clams in a very short period of time. As discussed in Chapter 6, Native American foragers were intimately familiar with their environment and likely knew how to identify keyholes and other clam signs. The increased visibility of clam locations during the winter and spring drastically reduced search time, making clams a more optimal resource during this time. I believe that this visibility of clams during cooler months is the primary cause for the predominance of winter and spring collecting in northeast Florida and much of the Georgia Bight.

Sedentism at the Guana Shell Ring

Clam exploitation at the Guana shell ring was clearly a seasonal endeavor. Clam gathering alone does not indicate a year‑round presence at the site, but the lack of seasons in the collection strategy of a single resource cannot be cited as evidence that people were absent from the site.

Often, coastal residents exploited food on a seasonal schedule despite the extended availability of a resource. Analysis of vertebrate remains may shed further light on the nature of occupation of Guana. Seasonality assessment of other invertebrate resources such as oysters and the assessment of seasons of death for fish could be used to complement clam seasonality data.

180 Clam exploitation at the Guana shell ring was very similar to patterns observed at other shell rings in the Georgia Bight and Mississippi period middens in northeast Florida. Although

Guana occupants gathered clams solely during the winter and spring, residents probably remained at the site for a longer duration. Shell rings subjected to both clam seasonality study and vertebrate faunal analysis typically demonstrate deposition throughout all four seasons of the year. Based on a combination of faunal and floral indicators, six shell rings (five Late

Archaic and one Mississippi) exhibited year‑round occupation by some part of the population

(Table 8.6). These are as follows: St. Catherines ring (Colaninno 2010; Quitmyer and Jones

2012), McQueen ring (Colaninno 2010; Quitmyer and Jones 2012), Cannons Point ring (Marrinan

1975; Quitmyer et al. 1985), West Ring (Colaninno 2010) and Grand shell ring (Ashley et al.

2007; Parsons 2008). Of these sites, only one (West Ring) has not had a clam seasonality assessment. Four out of five of these shell rings demonstrated clam gathering solely during the winter and spring, yet other evidence indicates that these sites were occupied by at least a part of the population for the duration of the year. This suggests that clam gathering was a seasonal strategy and that shell ring residents targeted other resources for the remainder of the year.

Table 8.6. Seasonality of Clams, Vertebrate Resources, and Floral Remains at Selected Shell Rings in the Georgia Bight.

Shell Ring Winter Spring Summer Fall St. Catherines Ring CV CV V V McQueen Ring CV CV V V Sapelo Ring Complex C C C C Cannons Point Ring V C F F West Ring V V V V Guana Ring C C Grand Ring CV CV V V

Key: C= Clams V=Vertebrate resource F= Floral remains

181 The shell rings discussed above provide strong evidence for sedentary populations at shell ring sites. Based on this evidence, I believe it is reasonable to suggest a sedentary occupation at the Guana shell ring as well. Although clam gathering at Guana occurred on seasonal schedule, this appears to be a typical pattern of exploitation at shell rings and does not indicate a limited seasonal presence at the site. Late Archaic coastal groups were clearly capable of living sedentary lives and chose to do so at several shell ring sites.

Summary

In this chapter I have compared the St. Augustine collection to two other nearby collections:

Kings Bay, Georgia (Quitmyer et al. 1997) and Indian River, Florida (Quitmyer 1995; Quitmyer et al. 1997). The St. Augustine collection is different from each collection and verifies that its use is warranted in the St. Augustine area. Although this study does not provide a minimum distance of acceptability between modern comparative collection and archaeological site, it appears that the maximum acceptable distance between the modern collection and archaeological site in question is somewhere between 45‑80 km. Based on the present results, the maximum acceptable distance may be somewhere around 70 km, but this remains to be tested.

At the Guana shell ring site, I have identified a gradual decline in the mean age and size of clams, indicating that Guana residents exploited clams to the point of decreasing yields.

Based on data from modern collections, it appears that Guana residents placed more predation pressure on local clam beds than modern commercial clam fishers in the Indian River.

Seasonality at Guana fits into a general pattern of cool weather collection during the spring and winter that is seen at most sites in the Georgia Bight. Elsewhere, I have suggested that this pattern was the result of targeting clams when their biomass and nutritional content were highest (Parsons 2008). Although this was likely an important consideration in the exploitation strategy, I have posited that the primary cause for cool weather clam collection is that clams are more visible during this time because of the presence of clam signs (primarily keyholes). These clam signs drastically reduce search time, resulting in decreased cost of procurement. This

182 means that clams are a more optimal resource during the winter and spring when clam signs are apparent.

Although Guana occupants gathered clams exclusively during the winter and spring, this does not mean that occupants were absent for the remainder of the year. Other shell rings in the Georgia Bight have demonstrated that clam collection occurred on a seasonal schedule despite occupants’ presence at the site throughout the year. Late Archaic coastal residents were clearly capable of living sedentary lifestyles. I have proposed that it is reasonable to expect sedentary occupation at Guana based on evidence at other Late Archaic sites. Future analysis of seasonality of other invertebrate and vertebrate resources at Guana will likely confirm this type of sedentary occupation.

183 CHAPTER 9

CONCLUSIONS

Summary

The primary goal of this dissertation was to examine the seasonal consumption of quahog clams at the Guana shell ring site in order to evaluate whether Guana occupants were semi‑sedentary or sedentary. A secondary goal of the dissertation was to amass a modern comparative clam collection from the St. Augustine area that would provide large amounts of biological data and permit accurate assessment of clam seasonality at the Guana site and other shell bearing sites in the area. The discussion below indicates areas of future research and summarizes the major findings of this research.

Future Research

The St. Augustine Collection

Although one year of gathering for a modern comparative clam collection is sufficient to identify seasons of death in an archaeological assemblage, I would like to add a second year of clam collection and water data to the St. Augustine collection. Claassen (1998:154) has argued that two years of collection should be undertaken for modern comparative clam collections, although the vast majority of sites in the Southeast have been assessed using collections gathered over the course of one year. Quitmyer and colleagues undertook a second year of collection at Kings Bay, which revealed only slight differences in the overall growth phase distributions and did not affect confidence in assessment of archaeological clams (Quitmyer et al. 1997:836). Although an additional year of data is not required for the St. Augustine collection, a second year of collection would be useful. A second year of collecting clams and

184 water data could demonstrate some inter‑annual variability in seasonal temperatures and seasonal growth, although I strongly suspect that such differences would be minimal. It also would be helpful to gather larger samples from collection Area A to verify the pattern of increased age in this area. Larger samples in general would provide more data on age structure and the overall size of clams.

Increasing the number of successfully assessed clams (i.e., growth phase has been identified) might provide slightly better monthly growth profiles. Unfortunately, the monthly profiles for 2010 were not sufficiently distinct to evaluate monthly growth distributions (this was also true of the Kings Bay collection). If any of the monthly St. Augustine profiles became distinct after a second year of collection, they could be used to identify seasonality on a monthly scale, rather than just on a seasonal scale. However, even with a second year of collection, it seems unlikely that these monthly profiles will be distinct. A second year of collection would primarily be useful to verify the overall patterns of growth observed in the 2010 collection.

The Guana Shell Ring Site

To date, only a very small portion of the Guana ring deposit has been excavated and evaluated.

Russo and colleagues estimated that the shell ring was comprised of 3,970 cubic meters (m3) of shell (Russo et al. 2002:11). Their excavations included a single 1‑x‑2 ‑m test unit in the central portion of the ring and 13 shovel tests placed along the ring arms, totaling approximately 6.2 m3. This amounts to roughly 1.6 percent of the estimated ring deposit, which is spread out over various areas of the ring. A block excavation on the ring using 1/8 in and 1/16 in screen would provide a larger dataset of faunal, ceramic, and other artifact classes for analysis. Extending the excavation slightly beyond the ring deposit (both interiorly and exteriorly) might also provide evidence for structures and may clarify the cultural stratigraphy of the ring (by examining its edges). This type of excavation would provide samples in close proximity to one another, and may also provide information on how deposition proceeded. This kind of exposure could be important given the slight seasonal variability of clam exploitation I observed in portions of the ring. Perhaps this kind of excavation would provide evidence indicating that the ring accumulated through the refuse disposal of discrete households, as I have hypothesized.

185 A block excavation on the Guana ring would likely provide a large clam sample for seasonality study. A large sample recovered from a dense area of the ring might provide a sufficient number of clams to permit the identification of yearly cycles of clam exploitation during the winter and spring. This would provide valuable information on how the Guana deposit accumulated and may provide some understanding of its stratigraphy. The identification of yearly cycles may also provide some indication as to the rapidity of accumulation and the duration of occupation. These insights would be helpful since our understanding of the duration of occupation is currently based solely on radiocarbon dates that overlap at one sigma. These dates suggest relatively fast accumulation and an occupation spanning perhaps less than a few centuries.

Detailed analyses of the extant vertebrate faunal collections from Guana would be useful to provide a better understanding of the general subsistence at Guana. Saunders and Rolland

(2006:66) found a high percentage of cartilaginous fish and mammalian remains in the ring interior. Detailed analysis of extant samples could elucidate whether the high mammalian frequency is a result of dietary preferences or taphonomic processes (preservation of shell in the ring interior was rather poor). Analysis of the faunal assemblage from the ring deposit could provide information with which to compare the ring interior assemblage, which could shed further light on the relationship between these two deposits. Analysis on the size of other resources, such as fish, and evaluations of whether these resources declined in size over time would provide interesting data for comparison with clam overexploitation. Finally, a detailed analysis of vertebrate remains might provide evidence for occupation in seasons other than winter and spring, which would be a valuable contribution to our understanding of the site.

Quantification of invertebrate remains at Guana also could enhance our overall understanding of subsistence and resource scheduling. Saunders and Rolland (2006:66) argued that quahog clams were more common in the ring interior midden than at other rings.

Quantification of invertebrate remains may provide information about specific habitats that were targeted and salinity levels, and would indicate the importance of invertebrates in the overall diet (i.e., percent of overall biomass).

186 Further seasonality assessments at Guana would likely provide evidence for year‑round occupation at Guana. Incremental growth assessments and isotopic assay of fish otoliths, isotopic assay of oysters and other shellfish, and assessments of size class of Boonea impressa (to examine oyster seasonality) are examples of faunal seasonality studies that could be conducted at Guana. Floral remains could also provide important insight on seasonal use of the site, but analysis of these remains has not been undertaken. These seasonality studies could shed light on overall subsistence patterns, resource use, and general sedentism at Guana.

In sum, much work remains to be done at Guana. Future research on both extant collections and new excavations could provide us with a substantial amount of new information about the site as a whole. Through this dissertation research, I have identified direct evidence for occupation of the site during the winter and spring, and I have proposed that the site was occupied throughout the year based on studies of other shell rings. Nevertheless, summer and fall occupations have yet to be identified at Guana; analysis of extant faunal collections could provide the data needed to evaluate whether summer and fall occupations occurred. Although

I have hypothesized that the ring may have accumulated from the refuse disposal of individual households, this hypothesis needs to be vigorously tested with new excavations of the ring deposit.

Summary of Findings

The St. Augustine Comparative Collection

In order to accurately identify seasons of death for archaeological clams, one must employ a modern comparative collection that was gathered in close proximity to the archaeological site in question. The Guana shell ring is bracketed by the Kings Bay, Georgia collection, gathered by

Quitmyer and colleagues (Quitmyer et al. 1985a and b, 1997) and the Indian River collection, published by Quitmyer (1995; Quitmyer et al. 1997). The Kings Bay collection was gathered approximately 85 km to the north of the Guana shell ring, and the Indian River collection was approximately 190 km to the south of the ring. I believed that the distance between the available collections and the archaeological site was probably too great to accurately identify

187 seasons of clam collection at Guana. Furthermore, I identified St. Augustine as a transitional area between the tide‑dominated salt marshes of the Georgia Bight and the wave‑dominated coasts and mangrove swamps of coastal Central Florida. The Kings Bay and Indian River collections originate from somewhat dissimilar ecosystems, and the St. Augustine area possesses characteristics of both of these ecosystems. For these reasons I decided that it was necessary to compile a modern comparative collection for the St. Augustine area.

I compiled the St. Augustine collection by making monthly trips into the marsh to collect living quahog clams. I gathered wild clams from the Matanzas River during every month in

2010 except October, when I had to purchase aquaculture clams from a nearby clam farm. In order to evaluate the influence of water temperature and salinity on seasonal growth, I deployed a water temperature data logger near the collection areas and utilized salinity data collected from a data logger maintained by the NERRS. These loggers tracked water temperature and salinity every 15‑30 minutes throughout 2010. These data revealed that, as expected, water temperature has a greater influence on seasonal growth patterns of clams.

Water temperature fluctuated on a seasonal scale as predicted. Salinity, however, remained fairly constant throughout the year, probably a result of the anthropogenically altered estuarine system. Transitions to growth phases occurred in conjunction with seasonal temperature shifts, indicating that seasonal temperature is a primary factor in the onset of seasonal growth.

Examination of age data for the St. Augustine collection supports the assertion that human predation can significantly alter age in a population (Quitmyer 1995; Quitmyer and

Jones 2000; Quitmyer et al. 1985b; this dissertation). I acquired wild clams for the St. Augustine collection from two areas in the Matanzas River. Area A was located on an island that is part of a clam lease; gathering of clams by anyone other than the leaseholder is prohibited, although illegal collection occurs with unknown regularity. Area B is accessible by foot from the A1A

Highway and is open to both recreational and commercial shellfish gathering. A comparison of ages from these two areas shows that Area A, which is more protected from gathering, had a higher mean age than clams from Area B. Clams from the Indian River, where human predation is high due to heavy commercial clam exploitation, had a further depressed mean age. These data contrast with the clams from Kings Bay, where human predation is

188 nonexistent. At Kings Bay, clams had a much higher mean age than those of either the St.

Augustine or Indian River collection. These modern collections demonstrate that human predation exerts considerable pressure on clam populations, resulting in decreased age statistics

Figure 9.1. Map of the Guana Shell Ring and Modern Comparative Clam Collections Discussed in Chapter 9. The Kings Bay collection area is approximately 85 km north of the Guana shell ring and approximately 125 km north of the St. Augustine comparative collection area. The St. Augustine collection area is approximately 40 km south of the Guana shell ring, approximately 125 km south of the Kings Bay collection area, and approximately 150 km north of the Indian River collection area.

189 within a population. Essentially, clams populations with little to no human predation exhibit older ages than populations with moderate to heavy human predation.

The St. Augustine collection area is approximately 125 km south of the Kings Bay collection area and approximately 150 km north of the Indian River collection area. Growth phase distributions from the St. Augustine collection reveal that the St. Augustine area is indeed a transitional area between the different ecosystems of the Georgia Bight and the Indian River area of Central Florida. The St. Augustine collection shares characteristics of seasonal growth with both the Kings Bay collection and the Indian River collection, but is sufficiently distinct to warrant its use in the St. Augustine area. The use of either the Kings Bay or Indian River collections to interpret the Guana archaeological collection would result in incorrect identification of seasonal exploitation. The St. Augustine collection shows growth phase distributions roughly in between those of the two surrounding modern collections. Based on the differences among modern comparative collections, I have proposed that the maximum acceptable distance between the archaeological site in question and the modern comparative collection lies somewhere between 45‑80 km (and is probably somewhere around 70 km), but ecology must be considered. Based on distance, it seems reasonable to suggest that seasonality studies of sites to the north of the St. Johns River should employ the Kings Bay comparative collection, and sites to south of the St. Johns River to New Smyrna Beach should employ the St.

Augustine collection (Figure 9.1). Sites south of New Smyrna Beach should employ the Indian

River collection. The establishment of the St. Augustine collection is an important contribution because it allows archaeologists to accurately assess clam seasonality at the numerous prehistoric and historic middens in the St. Augustine area. This collection will be curated at the

Florida State University Department of Anthropology.

The Guana Shell Ring Site

I hypothesized that Late Archaic shell rings in the Southeast were occupied by sedentary or semi‑sedentary fisher‑hunter‑gatherers. I based this hypothesis on recent zooarchaeological analyses of sites in the Georgia Bight that have shown that Late Archaic people were capable of living sedentary lifestyles in the coastal zone and often did so at the coastal middens and shell

190 rings that dot the Georgia Bight. I also based this hypothesis on the availability of highly productive and predictable estuarine and coastal resources, which can be acquired with mass capture techniques using relatively simple technology. The majority of estuarine and coastal resources can be obtained throughout the year, affording Late Archaic people the opportunity to remain in a single location throughout the year. I believe that the sedentism observed at Late

Archaic sites is a direct result of newly formed and/or newly stabilized estuaries following sea level stabilization around 6000 B.P. The stabilization of estuaries, bays, barrier islands, and coastlines would have allowed shellfish and fish to flourish, permitting dependence on these resources throughout the year. This stabilization would also have created patches of resources in predictable areas that would limit search time and could be exploited year after year.

For my dissertation research, I have evaluated the degree of sedentism at the Guana shell ring. I have examined sedentism by identifying the seasons of exploitation of quahog clams. Clam seasonality data indicate that Guana occupants were semi‑sedentary, meaning that they remained at the site for a minimum of two seasons. Guana occupants gathered clams exclusively during the cooler months of winter and spring. Collection times differed slightly between the ring interior and the ring deposit at Guana. Clams in the ring interior were gathered primarily during the spring with little, if any gathering during the winter. Clams from the ring deposit were gathered mainly during the winter with some additional gathering during the spring. Currently, there is no clear explanation for why clam collection varied between the two deposits. However, the general profile of cool weather collection during the winter and spring fits into a regional pattern of cool weather collecting in northeast Florida. This pattern of cool weather collection is observed throughout the Georgia Bight, although there is more variability in additional seasons of collection in Georgia sites.

Seasonality assessments of portions of the Guana shell ring indicate some slight differences in clam exploitation, but do not indicate strict use of ring portions during specific seasons. Based on the evidence at hand, it appears that the minor differences in clam seasonality in different portions of the ring may have been the result of slightly different seasonal use of clams in different households, rather than seasonal differences in serving practices for different portions of the ring during feasts. The current evidence tentatively

191 supports a model of gradual accumulation of domestic refuse into a horseshoe‑shaped ring.

Larger samples of clams, that would allow identification of cyclical seasonal exploitation through time in portions of the ring, could provide strong evidence with which to evaluate ring function. The detailed analysis of the vertebrate faunal assemblage could also provide valuable data about the assemblage and whether it represents feasting or quotidian refuse.

Guana residents exploited clams on a seasonal schedule during the cooler months of winter and spring. This fits into a pattern of cool weather collection observed at other shell rings and middens in the Georgia Bight. Most Late Archaic shell rings in the Georgia Bight evidence clam exploitation exclusively during the cool months of spring and often winter.

Midden sites from the Woodland to Mississippi period often revealed that clam collection occurred predominantly during the spring, although several sites included year‑round collections. I have identified the trend for cool weather clam exploitation in northeast Florida as a regional pattern. This exploitation pattern of exclusively cool weather collection is evident at nearly every site examined thus far in northeast Florida.

Elsewhere I have suggested that this pattern is a result of targeting clams when their biomass and nutrition are highest (Parsons 2008). Although this was likely a factor in the timing of clam exploitation, I have identified another explanation for the pattern of cool weather collection in the Georgia Bight. While I was assembling the St. Augustine collection, I observed that during the cool months of the year clams create signs (the most common of which are keyholes) that indicate their location. These signs make clams that are otherwise invisible under the sand readily apparent. This increased visibility of clams during cool months significantly reduces search time, making clams a more optimal resource during the winter and spring. I believe that the presence of clam signs, which increase visibility and result in decreased costs of procurement, is the primary reason for the pattern of winter and spring quahog clam collection seen at most sites in the Georgia Bight.

Although clam seasonality indicates only two seasons of occupation at Guana, this does not exclude the possibility of year‑round occupation. Many archaeologists have reiterated the fact that one cannot use negative evidence to infer the absence of people at a site during a specific time (because we may simply be missing the indicators for this time in the

192 archaeological record). Furthermore, other shell rings have demonstrated the same limited seasonal exploitation strategy of quahog clams yet demonstrated year‑round occupation at the site (through vertebrate and floral indicators). This is probably a result of decreased costs of procuring clams during the winter and spring. Once clams cease producing signs, costs of procurement would rise, and Late Archaic people probably focused their gathering efforts on other, less costly resources. The Guana occupants probably resided at the site throughout the year but employed a seasonal collection strategy for clams in order to optimize their foraging efforts. This assertion remains to be verified through the analysis of other types of data (e.g., flora, vertebrate fauna, and oyster seasonality).

Human overexploitation or deleterious effects on ecosystems are often associated with

European colonists and modern societies. However, research has shown that indigenous people made substantial impacts on their local environments (Erlandson and Rick 2008; Jackson et al. 2001; Quitmyer and Jones 2000). These impacts include diminished populations or extinction of species, altered ecosystems due to overexploitation of keystone species, and decreased size and age in a population. I have determined that Guana occupants heavily exploited clams, perhaps to the point of overexploitation. This heavy reliance on clams resulted in reduced yields, decreased mean age, and decreased size in the local clam population.

Modern collections of quahog clams have demonstrated that human predation can exert substantial pressure on clam populations, evidenced by decreased age and size. At Guana, clam age declines gradually over time, indicating increasing human predation pressure on local clam beds. The heavy predation pressure indicated by this decline might be the result of increased sedentism, increased population, increased use of clams rather than other resources, or could be the result of continued heavy exploitation of clams over the course of occupation at

Guana. Mean age for clams at Guana reached a low of 4.28 years, which is lower than the mean age for the Indian River modern collection (Indian River clams are subject to heavy human predation by commercial fishers). Based on mean age at death, it appears that Guana residents put more pressure on their clam beds than commercial fishers in the Indian River.

The size of archaeological clams at Guana also declines over time. This decline in mean size tracks declines in mean age at death, indicating that the decline in these variables is

193 probably caused by the same factor – human predation. The decrease in clam size means that meat yields from clams also decreased over time, although probably not dramatically. Based on the evidence of decreased age and size, it appears that during occupation of the site Guana residents overexploited clams to the point of decreased yields.

Sedentism at Late Archaic Sites in the Southeast

Faunal analysis of both vertebrate and invertebrate remains has been a key factor in identifying sedentary and semi‑sedentary societies in coastal Georgia and Florida. Previously, archaeologists viewed prehistoric societies as mobile groups that seasonally moved between the coast and interior (e.g., Milanich and Fairbanks 1980). These models of seasonal transhumance were based on notions of the coast as unsuitable or inhospitable during particular seasons of the year and modern preferences for seasonal consumption of bivalves. Until detailed faunal analyses of coastal sites were conducted, much of our knowledge about group mobility was based on marginalized extant foragers and ethnohistoric accounts of coastal groups under duress from European demands for food and encroachment upon their lands and resources

(e.g., Bennet 2001). During this time, most archaeologists conceptualized Archaic people as non‑sedentary, falling somewhere on the spectrum between highly mobile Paleoindians and sedentary agriculturalists of the Woodland and Mississippi periods (Russo 1998). Traditional markers of sedentism, such as storage pits, ceramics, deep midden deposits, large site size, formal cemeteries, cultivated or domesticated plants, and a dependable subsistence economy

(Binford 1980; Kelly 1998; Price and Brown 1985; Rafferty 1985; Rocek 1998; Russo 1998) were often considered insufficient evidence for sites of such antiquity.

Beginning in the 1970s, several archaeologists proposed that Late Archaic coastal sites were occupied on a more sedentary basis (e.g., Colquhoun and Brooks 1986; DePratter 1975,

1979; Marrinan 1975; Milanich and Fairbanks 1980:150; Price and Brown 1985:11‑12). Over the last fifty years, methodologies have been developed to identify the season of death of both vertebrate and invertebrate remains (e.g., Clark 1979; Killingly 1981; Quitmyer et al 1985a and b;

Russo 1991, 1998; Shackleton 1973; Simons 1986; Wheeler and Jones 1989). These types of

194 studies have become more common in archaeological research since the 1970s and 1980s. The ability to directly identify seasons of occupation based on the death of exploited resources has provided strong evidence for sedentary societies living along the coast since at least the Late

Archaic period.

Although clam seasonality studies at the Guana shell ring revealed that quahog clam collection occurred exclusively during the winter and spring, people likely remained at the site throughout the year and focused on other resources. Detailed faunal analyses examining both vertebrate and invertebrate remains have identified sedentary occupations at several

Southeastern shell rings (Ashley et al. 2007; Colaninno 2010; Marrinan 1975; Parsons 2008;

Quitmyer et al. 1985a and b; Quitmyer and Jones 2012; Russo 1998; Russo et al. 1993; Thompson and Andrus 2011). It is reasonable to suspect that as at other shell rings, Guana occupants gathered clams during the cooler months and remained at the site while focusing on other resources during the summer and fall.

In sum, over the last forty years zooarchaeological research in the coastal zone has demonstrated that during the Late Archaic period, people were well equipped to reside in the coastal zone throughout the year. Seasonality studies of vertebrate and invertebrate remains have been particularly useful to identify semi‑sedentary and sedentary occupations in the coastal strand. These analyses present a strikingly different view than that of highly mobile hunter‑gatherers migrating between inland areas and the coast. Rather, these studies show that coastal cultures were highly adapted to their environments and were capable of living a sedentary lifestyle as early as the Late Archaic period.

195 APPENDIX A

RADIOCARBON DATES OBTAINED FOR THIS DISSERTATION

Table A.1. Radiocarbon Dates from the Guana Shell Ring Obtained for this Dissertation. This table presents the radiocarbon dates and calibrations provided by Beta Analytic, Inc. All dates were obtained from Quahog clam shell.

Conventional With Measured 13C/12C Radiocarbon Reservoir Intercept 1 Sigma 2 Sigma Provenience Beta # Age Ratio Age BP Correction Cal BP Cal BP Cal BP 469N, 453E 287032 3430+40 ‑0.6 3830+50 3570+90 3450 3360‑3560 3260‑3680 LV 3 469N, 453E 287033 3440+50 ‑1.0 3840+50 3580+90 3460 3370‑3580 3290‑3680 LV 6 469N, 453E 287034 3350+40 ‑0.1 3760+40 3500+80 3380 3320‑3460 3210‑3560 LV 7 469N, 453E 287035 3310+50 ‑0.4 3720+50 3460+90 3340 3240‑3430 3130‑3550 LV 9 440N, 510E 296139 3440+40 ‑0.7 3840+50 3580+90 3460 3370‑3580 3290‑3680 LV 8

196 APPENDIX B

LOCATION OF SAMPLES FOR OXYGEN ISOTOPIC TESTING

Figure B.1. Modern Clam Number 4: Isotopic Samples. This image depicts the cross‑section of modern clam number 4 and the location of isotopic samples. The O and T indicate the apparent type of increment from which the sample was taken. Samples were drilled in ontogenetic order, and are indicated by the numbers.

Figure B.2. Modern Clam Number 3: Isotopic Samples. This image depicts the cross‑section of modern clam number 3 and the location of isotopic samples. The O and T indicate the apparent type of increment from which the sample was taken. Samples were drilled in ontogenetic order, and are indicated by the numbers.

197

Figure B.3 Modern Clam Number 11. This image depicts the cross‑section of modern clam number 11 and the location of isotopic samples. The O and T indicate the apparent type of increment from which the sample was taken. Samples were drilled in ontogenetic order, and are indicated by the numbers.

Figure B.4. Archaeological Clam Number 88. This image depicts the cross‑section of archaeological clam number 88 and the location of isotopic samples. The O and T indicate the apparent type of increment from which the sample was taken. Samples were drilled in ontogenetic order, and are indicated by the numbers.

198

Figure B.5. Archaeological Clam Number 222A. This image depicts the cross‑section of archaeological clam number 222A and the location of isotopic samples. The O and T indicate the apparent type of increment from which the sample was taken. Samples were drilled in ontogenetic order, and are indicated by the numbers.

Figure B.6. Archaeological Clam Number 222B. This image depicts the cross‑section of archaeological clam number 222B and the location of isotopic samples. The O and T indicate the apparent type of increment from which the sample was taken. Samples were drilled in ontogenetic order, and are indicated by the numbers.

199

Figure B.7. Archaeological Clam Number 336. This image depicts the cross‑section of archaeological clam number 336 and the location of isotopic samples. The O and T indicate the apparent type of increment from which the sample was taken. Samples were drilled in ontogenetic order, and are indicated by the numbers.

200 APPENDIX C

WATER TEMPERATURE DATA FOR THE ST. AUGUSTINE

COLLECTION

Water Temperatures for Fort Matanzas

90 85 80 75 70 65

Temperature (F) 60 55 50 Jul 06 Jul 07 Jul 08 Jul 09 Jul 10 Jan 06 Jan 07 Jan 08 Jan 09 Jan 10

Sep 06 Sep 07 Sep 08 Sep 09 Sep 10

Mar 06 Mar 07 Mar 08 Mar 09 Mar 10

Nov 06 Nov 07 Nov 08 Nov 09 Nov 10 May 06 May 07 May 08 May 09 May 10

Figure C.1. Water Temperature Data for Fort Matanzas From 2006 to 2010. This graph shows mean monthly water temperature data for five years from the NERRS water data collection station at Fort Matanzas. These temperatures were taken at a water depth ranging between 1 to 3 m below the surface. The average monthly temperatures for 2010 were quite similar to those of other years. January, February, and December of 2010 were slightly cooler than averages for other years; however, the averages are not so different as to negatively impact the usefulness of the St. Augustine comparative collection. During these months, rainfall was quite high, and probably resulted in slightly cooler temperatures than in previous years. Data was provided by NERRS through the CDMO website.

201 Table C.2. Monthly Low, High, and Average Water Temperature for 2010 (St. Augustine). These data are from the water temperature data logger I deployed in St. Augustine.

Low High Average Month Temperature Temperature Temperature January 43.03 65.4 54.82 February 48.36 63.09 55.02 March 52.16 70.5 60.3 April 63.43 77.99 71.15 May 70.76 86.53 79.24 June 70.76 91.08 82.14 July 73 91.64 83.22 August 74.73 90.24 83.28 September 78.73 87.21 82.88 October 71.62 81.33 76.81 November 66.3 72.57 70.19 December 47.11 70.97 56.24

Table C.3. Monthly Low, High, and Average Salinity for 2010 (St. Augustine). These data are from the NERRS water monitoring station at Fort Matanzas.

Low High Average Month Salinity Salinity Salinity January 23.1 35 32.77 February 27.6 34.6 32.07 March 29.4 35.7 33.62 April 30.6 36.1 34.49 May 28.8 37.9 34.94 June 30.7 37.5 35.25 July 33 37.3 35.97 August 31.9 37.7 36.45 September 32.4 38.2 36.62 October 35.1 37.9 36.83 November 33.3 37.3 36.64 December 36.2 37.5 36.73

202 APPENDIX D

DATA FOR THE MONTH OF OCTOBER FOR THE ST. AUGUSTINE

COLLECTION (AQUACULTURE CLAMS)

Table D.1. Shell Measurements (in millimeters) for the October St. Augustine Collection.

Measurements Hinge Length Height Count 34 34 34 Missing 0 0 0 Mean 43.16 72.35 67.21 Median 43.19 72.94 67.51 Std Dev 2.97 5.45 4.85 Min 37.93 62.02 58.76 Max 49.24 86.1 77.26

Table D.2. Shell Measurement Relationships for the October St. Augustine Collection.

Measure Count Average Minimum Maximum Std Dev Height of Length 34 92.97% 89.73% 97.98% 1.88 Hinge of Length 34 59.70% 55.82% 64.61% 1.86 Hinge of Height 34 64.22% 60.07% 67.72% 1.54

203 Table D.3. Age Data for the October St. Augustine Collection.

Age Value Count 34 Mean 3.94 Median 4 Mode 4 Std Dev 0.42 Min 3 Max 5

204 APPENDIX E

HINGE LENGTH DATA FOR THE ST. AUGUSTINE COLLECTION

AREAS

Table E.1. Hinge Length for Area A, the Subsample of Area B, and all of Area B.

Area B Area B Hinge Area A (Jul & Dec) (All) Count 67 67 309 Missing 0 0 1 Mean 44.75 43 44.23 Median 43.72 42.28 43.76 Std Dev 5.01 4.88 6.14 Minimum 37.38 33.87 27.78 Maximum 63.02 57.79 65.26

205 APPENDIX F

GROWTH PHASE DATA BY MONTH FOR THE ST. AUGUSTINE

COLLECTION

Table F.1. Growth Phase Distributions for the St. Augustine Collection by Percent.

PHASE JAN FEB MAR APR MAY JUN JUL AUG SEPT OCT NOV DEC O1 26.1 43.3 7.4 15.4 0 0 0 0 3.7 35.3 28.6 69 O2 34.8 26.7 22.2 42.3 0 0 0 0 0 0 0 17.2 O3 34.8 20 66.7 38.5 0 0 0 0 0 0 0 0 T1 0 0 0 0 88.5 76.9 40 4.4 14.8 0 0 0 T2 0 0 0 0 11.5 23.1 34.3 30.4 11.1 0 7.1 0 T3 4.3 10 3.7 3.8 0 0 25.7 65.2 70.4 64.7 64.3 13.8

Table F.2. Growth Phase Distributions for the St. Augustine Collection by Count.

Phase JAN FEB MAR APR MAY JUN JUL AUG SEPT OCT NOV DEC O1 6 13 2 4 0 0 0 0 1 12 8 20 O2 8 8 6 11 0 0 0 0 0 0 0 5 O3 8 6 18 10 0 0 0 0 0 0 0 0 T1 0 0 0 0 23 20 14 1 4 0 0 0 T2 0 0 0 0 3 6 12 7 3 0 2 0 T3 1 3 1 1 0 0 9 15 19 22 18 4 Total 23 30 27 26 26 26 35 23 27 34 28 29

206 January 40 35 30 25 20

Percent 15 10 5 0 O1 O2 O3 T1 T2 T3

Figure F.3. Growth Phase Distribution for the Month of January.

February 50 45 40 35 30 25

Percent 20 15 10 5 0 O1 O2 O3 T1 T2 T3

Figure F.4. Growth Phase Distribution for the Month of February.

207 March 80 70 60 50 40

Percent 30 20 10 0 O1 O2 O3 T1 T2 T3

Figure F.5. Growth Phase Distribution for the Month of March.

April 45 40 35 30 25 20 Percent 15 10 5 0 O1 O2 O3 T1 T2 T3

Figure F.6. Growth Phase Distribution for the Month of April.

208 May 100 90 80 70 60 50

Percent 40 30 20 10 0 O1 O2 O3 T1 T2 T3

Figure F.7. Growth Phase Distribution for the Month of May.

June 90 80 70 60 50 40 Percent 30 20 10 0 O1 O2 O3 T1 T2 T3

Figure F.8. Growth Phase Distribution for the Month of June.

209 July 45 40 35 30 25 20 Percent 15 10 5 0 O1 O2 O3 T1 T2 T3

Figure F.9. Growth Phase Distribution for the Month of July.

August 70 60 50 40 30 Percent 20 10 0 O1 O2 O3 T1 T2 T3

Figure F.10. Growth Phase Distribution for the Month of August.

210 September 80 70 60 50 40

Percent 30 20 10 0 O1 O2 O3 T1 T2 T3

Figure F.11. Growth Phase Distribution for the Month of September.

October 70 60 50 40 30 Percent 20 10 0 O1 O2 O3 T1 T2 T3

Figure F.12. Growth Phase Distribution for the Month of October.

211 November 70 60 50 40 30 Percent 20 10 0 O1 O2 O3 T1 T2 T3

Figure F.13. Growth Phase Distribution for the Month of November.

December 80 70 60 50 40

Percent 30 20 10 0 O1 O2 O3 T1 T2 T3

Figure F.14. Growth Phase Distribution for the Month of December.

212 APPENDIX G

DATA FOR THE ARCHAEOLOGICAL CLAMS IN THE RING

INTERIOR AND RING DEPOSIT

Table G.1. Shell Measurements (in millimeters) for the Ring Interior for Upper and Lower Levels.

Hinge Hinge Length Length Height Height Statistic Levels 1‑3 Levels 4‑6 Levels 1‑3 Levels 4‑6 Levels 1‑3 Levels 4‑6 Count 59 56 54 56 59 58 Missing 19 21 24 21 19 19 Mean 50.12 51.41 85.61 88.57 81.48 83.32 Median 50.24 51.03 86.25 87.98 82.5 82.5 Std Dev 8.03 8.1 14.2 14.95 13.03 14.02 Min 36.87 35.44 63.7 61.78 59.57 58 Max 67.82 71.76 118 127.65 110.3 116.5

Table G.2. Shell Measurements (in millimeters) for the Ring Deposit for Upper and Lower Levels.

Hinge Hinge Length Length Height Height Statistic Levels 1‑5 Levels 6‑12 Levels 1‑5 Levels 6‑12 Levels 1‑5 Levels 6‑12 Count 83 87 87 80 56 94 Missing 41 30 37 37 21 23 Mean 47.68 48.16 83.46 82.24 88.57 77.82 Median 47.82 48.3 84.15 82.67 87.98 78.13 Std Dev 8.41 7.19 15.98 11.68 14.95 13.01 Min 27.46 31.88 47.87 54.39 61.78 50.9 Max 72.24 64.76 130 108.44 127.65 137.73

213 REFERENCES

Abbot, R. T., and P. A. Morris 1995 Shells of the Atlantic and Gulf Coasts and the West Indies. Fourth Edition. Houghton Mifflin Company, Boston.

Adams, H., and A. Adams 1857 The Genera of Recent Mollusca. II. John van Voost, London.

Anderson, D. G. 1996 Approaches to Modeling Regional Settlement in the Archaic Period Southeast. In Archaeology of the Mid‑Holocene Southeast, edited by K. E. Sassaman and D. G. Anderson, pp. 151‑176. University Press of Florida, Gainesville.

Anderson, S. H. 2007 Shell Middens (ʺKokkenmoddingerʺ) in Danish Prehistory as a Reflection of the Marine Environment. In Shell Middens in Atlantic Europe, edited by N. Milner, O. E. Craig and G. Bailey, pp. 31‑45. Oxbow Books, Oxford.

Andrus, C. F. T., and D. E. Crowe 2008 Isotope Analysis as a Means for Determining Season of Capture for Mercenaria. In Native American Landscapes of St. Catherines Island, Georgia, edited by D. H. Thomas, pp. 498‑518. American Museum of Natural History, Anthropological Papers Number 88, Part II, New York.

Ansell, A. D. 1968 The Rate of Growth of the Hard Clam Mercenaria mercenaria (L.) throughout the Geographical Range. Journal du Conseil, Permanent International pour lʹExploration de ler Mer 31:364‑409.

Ansell, A. D., and K. F. Lander 1967 Studies on the Hard‑Shell Clam, Venus mercenaria, in British Waters. III. Further Observations on the Seasonal Biochemical Cycle and on Spawning. The Journal of Applied Ecology 4(2):425‑435.

Arnold, W. S., T. M. Bert, I. R. Quitmyer, and D. S. Jones 1998 Contemporaneous Deposition of Annual Growth Bands in Mercenaria mercenaria (Linnaeus), Mercenaria campechiensis (Gmelin), and the Natural Hybrid Forms. Journal of Experimental Marine Biology and Ecology 223:93‑109.

214 Arnold, W. S., D. C. Marelli, T. M. Bert, D. S. Jones, and I. R. Quitmyer 1991 Habitat‑Specific Growth of Hard Clams Mercenaria mercenaria (L.) from the Indian River, Florida. Journal of Experimental Marine Biology and Ecology 147:245‑265.

Ashley, K., V. Rolland, and R. A. Marrinan 2007 A Grand Site: Archaeological Testing of the Grand Shell Ring (8DU1). Report submitted to Talbot Islands State Park, Duval County, Florida.

Aten, L. E. 1981 Determining the Seasonality of Rangia cuneata from Gulf Coast Shell Middens. Bulletin of the Texas Archeological Society 52:179‑200.

1999 Middle Archaic Ceremonialism at Tick Island, Florida: Ripley P. Bullenʹs 1961 Excavation at the Harris Creek Site. The Florida Anthropologist 52(3):131‑187.

Bailey, G., M. R. Deith, and N. J. Shackleton 1983 Oxygen Isotope Analysis and Seasonality Determinations: Limits and Potential of a New Technique. American Antiquity 48(2):390‑398.

Bailey, G., and J. Parkington 1988 The Archaeology of Prehistoric Coastlines: An Introduction. In The Archaeology of Prehistoric Coastlines, edited by G. Bailey and J. Parkington, pp. 1‑10. Cambridge University Press, Cambridge.

Balsillie, J. H., and J. F. Donoghue 2004 High Resolution Sea‑Level History for the Gulf of Mexico since the Last Glacial Maximum. Florida Geological Survey Report of Investigations, Number 103.

Bar‑Yosef Mayer, D. E. 2005 An Introduction to Archaeomalacology. In Archaeomalacology: Molluscs in Former Environments, edited by D. E. Bar‑Yosef Mayer, pp. 1‑5. Oxbow Books, Oxford.

Bar‑Yosef, O., and T. R. Rocek 1998 Introduction. In Seasonality and Sedentism: Archaeological Perspectives from Old and New World Sites, edited by T. R. Rocek and O. Bar‑Yosef, pp. 1‑8. Peabody Museum Bulletin 6. Peabody Museum of Archaeology and Ethnology, Harvard University, Cambridge, Massachusetts.

Bennett, C. E. 2001 Three Voyages. University of Alabama Press, Tuscaloosa.

215 Bettinger, R. L. 1980 Explanatory/Predictive Models of Hunter‑Gatherer Adaptation. Advances in Archaeological Method and Theory, edited by M. B. Schiffer, pp. 189‑255. vol. 3. Academic Press, New York.

1991 Hunter‑Gatherers: Archaeological and Evolutionary Theory. Plenum Press, New York.

Binford, L. R. 1980 Willow Smoke and Dogsʹ Tails: Hunter‑Gatherer Settlement Systems and Archaeological Site Formation. American Antiquity 45(1):4‑20.

1982 The Archaeology of Place. Journal of Anthropological Archaeology 1:5‑31.

2006 Bands as Characteristic of ʺMobile Hunter‑Gatherersʺ May Exist Only in the History of Anthropology. In Archaeology and Ethnoarchaeology of Mobility, edited by F. Sellet, R. Greaves, and P. Yu, pp. 3‑22. University Press of Florida, Gainesville.

Bjerck, H. B. 2007 Mesolithic Coastal Settlements and Shell Middens (?) in Norway. In Shell Middens in Atlantic Europe, edited by N. Milner, O. E. Craig, and G. Bailey, pp. 5‑30. Oxbow Books, Oxford.

Bogucki, P. 1999 The Origins of Human Society. Blackwell Publishers, Malden, Massachusetts.

Braley, C. O., L. D. OʹSteen, and I. R. Quitmyer 1986 Archaeological Investigations at 9McI41, Harris Neck National Wildlife Refuge, McIntosh County, Georgia. Report Submitted to the U.S. Department of the Interior Fish and Wildlife Service, Contract No. 14‑16‑004‑84‑019. Southeastern Archaeological Services, Inc., Athens, Georgia.

Cable, J. S. 1997 The Ceremonial Mound Theory: New Evidence for Possible Ceremonial Function of Shell Rings. South Carolina Archaeology Week Poster: Shell Rings of the Late Archaic. South Carolina Institute of Archaeology and Anthropology, University of South Carolina, Columbia.

Calmes, A. 1967 Test Excavations at Three Late Archaic Shell‑Ring Mounds on Hilton Head Island, Beaufort County, South Carolina. Southeastern Archaeological Conference Bulletin 8:1‑22.

216 Claassen, C. 1986 Shellfishing Seasons in the Prehistoric Southeastern United States. American Antiquity 51(1):21‑37.

1998 Shells. Cambridge University Press, Cambridge.

2010 Feasting with Shellfish in the Southern Ohio Valley: Archaic Sacred Sites and Rituals. University Press of Tennessee, Knoxville, Tennessee.

Clark, G. R., II 1979 Seasonal Growth Line Variation in the Shells of Recent and Prehistoric Specimens of Mercenaria mercenaria from St. Catherines Island, Georgia. In The Anthropology of St. Catherines Island: 2. The Refuge‑Deptford Mortuary Complex, edited by D. H. Thomas and C. S. Larsen, pp. 161‑172. American Museum of Natural History Anthropological Papers Number 56, New York.

Colaninno, C. E. 2010 Zooarchaeological Analysis of Vertebrate Remains from Five Late Archaic Shell Rings on the Georgia Coast, USA. Ph.D. Dissertation, Department of Anthropology, University of Georgia, Athens, Georgia.

Colquhoun, D. J., and M. J. Brooks 1986 New Evidence from the Southeastern U.S. for Eustatic Components in the Late Holocene Sea Levels. Geoarchaeology 1(3):275‑291.

Coutts, P. 1970 Bivalve Growth Patterning as a Method for Seasonal Dating in Archaeology. Nature 226:874.

Coutts, P., and C. Higham 1971 The Seasonal Factor in Prehistoric New Zealand. World Archaeology 2(3):266‑277.

Crook, M. R. 1992 Oyster Sources and their Prehistoric Use on the Georgia Coast. Journal of Archaeological Science 19:483‑496.

Crusoe, D., and C. B. DePratter 1976 A New Look at the Georgia Coastal Shellmound Archaic. The Florida Anthropologist 29(1):1‑23.

Dall, W. H. 1903 Contributions to the Tertiary Fauna of Florida. III, Part VI. Transactions of the Wagner Free Institute of Science, Philadelphia.

217 Deith, M. R. 1983 Molluscan Calendars: The Use of Growth‑Line Analysis to Establish Seasonality of Shellfish Collection at the Mesolithic Site of Morton, Fife. Journal of Archaeological Science 10:423‑440.

1988 Oxygen Isotope Analysis of Marine Molluscs from Franchthi Cave. In Marine Molluscan Remains from Franchthi Cave, edited by N. J. Shackleton, pp. 133‑155. Indiana University Press, Indiana.

DePratter, C. B. 1975 The Archaic in Georgia. Early Georgia 3(1):1‑16.

1976 Shellmound Archaic on the Georgia Coast. Masterʹs Thesis, Department of Anthropology, University of Georgia, Athens.

1977 Environmental Changes on the Georgia Coast During the Prehistoric Period. Early Georgia 5:1‑14.

1979 Shellmound Archaic on the Georgia Coast. South Carolina Antiquities 11(2):1‑69.

DePratter, C. B., and J. D. Howard 1980 Indian Occupation and Geologic History of the Georgia Coast: A 5,000 Year Summary. In Excursions in Southeastern Geology: The Archaeology‑Geology of the Georgia Coast, edited by J. D. Howard, C. B. DePratter, and R. W. Frey. Guidebook 20. Geological Society of America.

Dincauze, D. F. 2000 Environmental Archaeology: Principles and Practice. Cambridge University Press, Cambridge, Great Britain.

Donoghue, J. F., and N. M. White 1995 Late‑Holocene Sea Level Change and Delta Migration, Apalachicola River Region, Northwest Florida, U.S.A. Journal of Coastal Research 11(3):651‑663.

Doran, G. 2007 A Brief Continental View from Windover. In Ancient Health, edited by M. N. Cohen and G. M. M. Crane‑Kramer, pp. 35‑51. University Press of Florida, Gainesville.

Eble, A. F. 2001 Anatomy and Histology of Mercenaria mercenaria. In Biology of the Hard Clam, edited by J. N. Kraeuter and M. Castagna, pp. 117‑220. Elsevier, Developments in Aquaculture and Fisheries Science ‑ 31, Amsterdam.

218 Edwards, W. E. 1965 A Preliminary Report on the Sewee Mound Shell Ring, Charleston County, South Carolina. Report submitted to the U.S. Forest Service. MS on file, South Carolina Institute of Archaeology and Anthropology, University of South Carolina, Columbia.

Erlandson, J. M. 2001 The Archaeology of Aquatic Adaptations: Paradigms for a New Millennium. Journal of Archaeological Research 9(4):287‑350.

Erlandson, J. M., and M. L. Moss 2001 Shellfish Feeders, Carrion Eaters, and the Archaeology of Aquatic Adaptations. American Antiquity 66(3):413‑432.

Erlandson, J. M., and T. C. Rick 2008 Archaeology, Marine Ecology, and Human Impacts on Marine Environments. In Human Impacts on Ancient Marine Ecosystems: A Global Perspective, edited by T. C. Rick and J. M. Erlandson, pp. 1‑19. University of California Press, Berkeley.

Evans, J. G. 1969 The Exploitation of Molluscs. In The Domestication and Exploitation of Plants and Animals, edited by P. Ucko and G. W. Dimbleby, pp. 478‑484. Gerald Duckworth & Co. Ltd. and Aldine Publishing Company, Chicago, Illinois.

Eversole, A. G. 1997 Gametogenesis of Mercenaria mercenaria, M. campechiensis and Their Hybrids. The Nautilus 110(4):107‑110.

2001 Reproduction in Mercenaria mercenaria. In Biology of the Hard Clam, edited by J. N. Kraeuter and M. Castagna, pp. 221‑260. Elsevier, Developments in Aquaculture and Fisheries Science ‑ 31, Amsterdam.

Fairbanks, R. G. 1989 A 17,000‑Year Glacio‑Eustatic Sea Level Record: Influence of Glacial Melting Rates on the Younger Dryas Event and Deep‑Ocean Circulation. Nature 342:637‑642.

Faught, M. K. 2004 The Underwater Archaeology of Paleolandscapes, Apalachee Bay, Florida. American Antiquity 69(2):275‑289.

Fegley, S. R. 2001 Demography and Dynamics of Hard Clam Populations. In Biology of the Hard Clam, edited by J. N. Kraeuter and M. Castagna, pp. 383‑422. Elsevier, Developments in Aquaculture and Fisheries Science ‑ 31, Amsterdam.

219 Flannery, R. 1943 Some Notes on a Few Sites in Beaufort County, South Carolina. Bureau of American Ethnology, Bulletin 123:143‑155.

Fritz, L. W. 2001 Shell Structure and Age Determination. In Biology of the Hard Clam, edited by J. N. Kraeuter and M. Castagna, pp. 53‑76. Elsevier, Developments in Aquaculture and Fisheries Science ‑ 31, Amsterdam.

Gibson, J. L., and P. J. Carr 1998 Big Mounds, Big Rings, Big Power. In Signs of Power: The Rise of Cultural Complexity in the Southeast, edited by J. L. Gibson and P. J. Carr, pp. 1‑9. University of Alabama Press, Tuscaloosa.

Gmelin, J. F. 1791 Caroli a Linne. Systema Naturae per Regna Tria Naturae: Secundum Classes, Ordines, Genera, Species cum Characteribus, Differentils, Synonymis, Locis. Lipsiae: impensis. George Emanuel Beer.

Greaves, R. 2006 Forager Landscape Use and Residential Organization. In Archaeology and Ethnoarchaeology of Mobility, edited by F. Sellet, R. Greaves, and P.‑L. Yu, pp. 127‑152. University Press of Florida, Gainesville.

Grizzle, R. E., V. M. Bricelj, and S. E. Shumway 2001 Physiological Ecology of Mercenaria mercenaria. In Biology of the Hard Clam, edited by J. N. Kraeuter and M. Castagna, pp. 305‑382. Elsevier, Developments in Aquaculture and Fisheries Science ‑ 31, Amsterdam.

Harte, M. E. 2001 Systematics and Taxonomy. In Biology of the Hard Clam, edited by J. N. Kraeuter and M. Castagna, pp. 3‑52. Elsevier, Developments in Aquaculture and Fisheries Science ‑ 31, Amsterdam.

Hemmings, E. T. 1970 Emergence of Formative Life on the Atlantic Coast of the Southeast. Report prepared by the Institute of Archaeology and Anthropology, University of South Carolina.

Henry, K. M., and R. M. Cerrato 2007 The Annual Macroscopic Growth Pattern of the Northern Quahog [=Hard Clam, Mercenaria mercenaria (L.)] in Narragansett Bay, Rhode Island. Journal of Shellfish Research 26(4):985‑993.

220 Herbert, J. M., and L. C. Steponaitis 1998 Estimating the Season of Harvest of Eastern Oyster (Crassostrea virginica) with Shells from the Chesapeake Bay. Southeastern Archaeology 17(1):53‑71.

Hesselmann, D. M., B. J. Barber, and N. J. Blake 1989 The Reproductive Cycle of Adult Hard Clams, Mercenaria spp. in the Indian River Lagoon, Florida. Journal of Shellfish Research 8(1):43‑49.

Higham, T. F. G., and P. L. Horn 2000 Seasonal Dating Using Fish Otoliths: Results from the Shag River Mouth Site, New Zealand. Journal of Archaeological Science 27:439‑448.

Hilbish, T. J. 2001 Genetics of Hard Clams, Mercenaria mercenaria. In Biology of the Hard Clam, edited by J. N. Kraeuter and M. Castagna, pp. 261‑280. Elsevier, Developments in Aquaculture and Fisheries Science ‑ 31, Amsterdam.

Hubbard, D. K., G. Oertel, and D. Nummedal 1979 The Role of Waves and Tidal Currents in the Development of Tidal‑Inlet Sedimentary Structures and Sand Body Geometry: Examples from North Carolina, South Carolina, and Georgia. Journal of Sedimentary Petrology 49(4):1073‑1092.

Jackson, J. B. C., M. X. Kirby, W. H. Berger, K. A. Bjorndal, L. W. Botsford, B. J. Bourque, R. H. Bradbury, R. Cooke, J. M. Erlandson, J. A. Estes, T. P. Hughes, S. Kidwell, C. B. Lange, H. S. Lenihan, J. M. Pandolfi, C. H. Peterson, R. S. Steneck, M. J. Tegner, and R. R. Warner 2001 Historical Overfishing and the Recent Collapse of Coastal Ecosystems. Science 27:629‑637.

Jelgersma, S., and M. J. Tooley 1995 Sea‑Level Changes During the Recent Geological Past. Journal of Coastal Research 17:123‑139.

Jochim, M. 1998 A Hunter‑Gatherer Landscape. Southwest Germany in the Late Paleolithic and Mesolithic. Plenum Press, New York.

2002 The Mesolithic. In European Prehistory: A Survey, edited by S. Milisauskas, pp. 115‑142. Springer Science, New York.

Jones, D. S., and I. R. Quitmyer 1996 Marking Time with Bivalve Shells: Oxygen Isotopes and Season of Annual Increment Formation. Palaois 11(4):340‑346.

221 Kelly, R. R. 1992 Mobility/Sedentism: Concepts, Archaeological Measures, and Effects. Annual Review of Anthropology 21:43‑66.

1998 Foraging and Sedentism. In Seasonality and Sedentism: Archaeological Perspectives from Old and New World Sites, edited by T. R. Rocek and O. Bar‑Yosef, pp. 9‑24. Peabody Museum Bulletin 6. Peabody Museum of Archaeology and Ethnology, Cambridge, Massachusetts.

Killingley, J. S. 1981 Seasonality of Mollusk Collecting Determined from O‑18 Profiles of Midden Shells. American Antiquity 46:152‑158.

Linnaeus, C. 1758 Systema Naturae. 10th Edition I., Laurentius Salvius, Stockholm.

Lutz, R. A. and D. C. Rhoads 1980 Growth Patterns within the Molluscan Shell: An Overview. In Skeletal Growth of Aquatic Organisms: Biological Records of Environmental Change, edited by D. C. Rhoads and R. A. Lutz, pp. 203‑254. Plenum Press, New York.

MacKenzie, C. L., Jr., A. Morrison, D. L. Taylor, V. G. Burrell, Jr., W. S. Arnold, and A. T. Wakida‑Kunsonoki 2002a Quahogs in Eastern North America: Part I, Biology, Ecology, and Historical Uses. Marine Fisheries Review 64(2):1‑55.

2002b Quahogs in Eastern North America: Part II, History by Province and State. Marine Fisheries Review 64(3):1‑64.

MacKenzie, C. L., Jr., D. L. Taylor, and W. S. Arnold 2001 A History of Hard Clamming. In Biology of the Hard Clam, edited by J. N. Kraeuter and M. Castagna, pp. 651‑674. Elsevier, Developments in Aquaculture and Fisheries Science ‑ 31, Amsterdam.

MacMahon, D. A., and W. H. Marquardt 2004 The Calusa and their Legacy: South Florida People and their Environments. University Press of Florida, Gainesville.

Marrinan, R. A. 1975 Molluscs, Ceramics, and Sedentism: The Late Archaic Period on the Georgia Coast. Ph.D. Dissertation, Department of Anthropology, University of Florida, Gainesville.

222 2010 Two Late Archaic Period Shell Rings, St. Simonʹs Island, Georgia. In Trend, Tradition, and Turmoil: What Happened to the Southeastern Archaic?, edited by D. H. Thomas and M. C. Sanger, pp. 71‑102. American Museum of Natural History, Anthropological Papers Number 93, New York.

McKinley, W. 1873 Mounds in Georgia. In Annual Report of the Board of Regents of the Smithsonian Institution for the year 1872, pp. 422‑428. Government Printing Institution, Washington, D.C.

Meehan, B. 1982 Shell Bed to Shell Midden. Australian National University, Canberra, Australia.

Milanich, J. T. 1994 Archaeology of Precolumbian Florida. University Press of Florida, Gainesville.

Milanich, J. T., and C. H. Fairbanks 1980 Florida Archaeology. Academic Press, New York.

Miller, J. J. 1992 Effects of Environmental Change on Late Archaic People of Northeast Florida. The Florida Anthropologist 45(2):100‑106.

1998 An Environmental History of Northeast Florida. University Press of Florida, Gainesville.

Milner, N., O. E. Craig, and G. Bailey 2007 Shell Middens in Atlantic Europe. In Shell Middens in Atlantic Europe, edited by N. Milner, O. E. Craig, and G. Bailey, pp. 1‑4. Oxbow Books, Oxford.

Monks, G. G. 1981 Seasonality Studies. Advances in Archaeological Method and Theory, edited by M. B. Schiffer, pp. 177‑240. vol. 4. Academic Press, New York.

Moore, C. B. 1897 Certain Aboriginal Mounds of the Georgia Coast. In The Georgia and South Carolina Coastal Expeditions of Clarence Bloomfield Moore, edited by L. H. Larson, Jr., pp. 91‑240. University of Alabama Press, Tuscaloosa.

Morris, J. T., P. V. Sundareshwar, C. T. Nietch, B. Kjerfve, and D. R. Cahoon 2002 Responses of Coastal Wetlands to Rising Sea Level. Ecology 83(10):2869‑2877.

223 Murdock, G. P. 1967 Ethnographic Atlas: A Summary. Ethnology 6(2):109‑236.

National Oceanic and Atmospheric Administration, Office of Ocean and Coastal Resource Management, National Estuarine Research Reserve System‑Wide Monitoring Program. 2011 Centralized Data Management Office, Baruch Marine Field Lab. University of South Carolina, Columbia. http://cdmo.baruch.sc.edu

OʹBrien, D. M., and D. H. Thomas 2008 The Molluscan Incremental Sequence. In Native American Landscapes of St. Catherines Island, Georgia, edited by D. H. Thomas, pp. 475‑497. American Museum of Natural History, Anthropological Papers Number 88, Part II, New York.

OʹConnell, J. F., and K. Hawkes 1981 Alyawara Plant Use and Optimal Foraging Theory. In Hunter‑Gatherer Foraging Strategies, edited by B. Winterhalder and E. A. Smith, pp. 99‑125. University of Chicago Press, Chicago.

Orians, G. H., and N. E. Pearson 1979 On the Theory of Central Place Foraging. In Analysis of Ecological Systems, edited by D. J. Horn, G. R. Stairs, and R. D. Mitchell, pp. 155‑177. Ohio State University, Columbus, Ohio.

Parsons, A.L. 2007 Optimal Foraging Theory and Food Taboos: An Exploration of Shellfish and Small Aquatic Fauna. MS in possession of author.

2008 The Grand Shell Ring: A Study of Site Seasonality, Hard Clam Exploitation, and Resource Scheduling. Masterʹs Thesis, Department of Anthropology, Florida State University, Tallahassee.

n.d. An Overview of Seasonal Clam Exploitation in the Mosquito Lagoon. In A Phase II Archeological Survey of Mosquito Lagoon, Canaveral National Seashore, Volusia County, Florida, edited by R. Hellman. Report submitted to Southeast Archeological Center, Tallahassee, Florida.

Parsons, A. L., and R. A. Marrinan n.d. An Assessment of Coastal Faunal Data from Coastal Georgia and Northeast Florida. In Life Among the Tides: Recent Archaeology on the Georgia Bight, edited by V. D. Thompson and D. H. Thomas. American Museum of Natural History, Anthropological Papers, New York.

224 Peterson, C. H. 1983 A Concept of Quantitative Reproductive Senility: Application to the Hard Clam, Mercenaria mercenaria (L.)? Oecologia 58:164‑168.

1986 Quantitative Allometry of Gamete Production by Mercenaria mercenaria into Old Age. Marine Ecology Progress Series 29:93‑97.

Peterson, C. H., and S. R. Fegley 1986 Seasonal Allocation of Resources to Growth of Shell, Soma and Gonads in Mercenaria mercenaria. Biological Bulletin 171(3):597‑610.

Piatek, B. J. 1992 Tomoka State Park Survey and Preliminary Test Excavation Results. The Florida Anthropologist 45(4):326‑335.

Price, D. T. 1991 The Mesolithic of Northern Europe. Annual Review of Anthropology 20:211‑233.

Price, D. T., and J. A. Brown 1985 Aspects of Hunter‑Gatherer Complexity. In Prehistoric Hunter‑Gatherers: The Emergence of Cultural Complexity, edited by D. T. Price and J. A. Brown, pp. 3‑20. Academic Press, Orlando, Florida.

Quitmyer, I. R. 1985 The Environment of the Kings Bay Locality. In Aboriginal Subsistence and Settlement Archaeology of the Kings Bay Locality, edited by W. H. Adams, pp. 1‑32. vol. 2. Report Submitted to OICC Trident, Kings Bay Naval Submarine Base, U.S. Department of the Navy, Kings Bay, Georgia.

1995 Seasons of Harvest and Age Structure of Mercenaria spp. Excavated from St. Johns II (A.D. 590‑1400) Archaeological Contexts of the Mosquito Lagoon, Florida. In Final Report on the Archaeological Investigations at the Seminole Rest Site (CANA‑ 063/8VO124), Canaveral National Seashore, Volusia County, edited by E. Horvath. Report Submitted to the Southeast Archeological Center, Tallahassee, Florida.

Quitmyer, I. R., H.S. Hale, and D. S. Jones 1985a Paleoseasonality Determinations Based on Incremental Shell Growth in the Hard Clam Mercenaria mercenaria, and its Implications for the Analysis of Three Southeast Georgia Coastal Shell Middens. Southeastern Archaeology 4(1):27‑40.

225 1985b Seasonality Study Based on the Incremental Growth Data from the Quahog Clam (Mercenaria mercenaria). In Aboriginal Subsistence and Settlement Archaeology of the Kings Bay Locality, edited by W. H. Adams. Reports of Investigation 2: Zooarchaeology. University of Florida, Department of Anthropology, Gainesville, Florida.

Quitmyer, I. R., and D. S. Jones 1992 Calendars of the Coast: Seasonal Growth Increment Patterns in Shells of Modern and Archaeological Southern Quahogs, Mercenaria campechiensis, from Charlotte Harbor, Florida. In Culture and Environment in the Domain of the Calusa, edited by W. H. Marquardt, pp. 247‑264. University of Florida Press, Gainesville, Florida.

2000 The Over‑Exploitation of Hard Clams (Mercenaria spp.) from Five Archaeological Sites in the Southeastern United States. The Florida Anthropologist 53(2):158‑166.

2009 Still Listening: Four Decades of Hard Clam (Mercenaria spp.) Sclerochronology on St. Catherines Island, Georgia. Paper presented at the 74th Annual Meeting of the Society for American Archaeology, Atlanta.

2012 Annual Incremental Shell Growth Patterns in Hard Clams (Mercenaria spp.) from St. Catherines Island, Georgia: A Record of Seasonal and Anthropogenic Impact on Zooarchaeological Resources. In Seasonality and Human Mobility Along the Georgia Bight, edited by E. J. Reitz, I. R. Quitmyer, and D. H. Thomas, pp. 135‑148. American Museum of Natural History, Anthropological Papers, Number 97, New York.

Quitmyer, I. R., D. S. Jones, and C. F. T. Andrus 2005 Seasonal Collection of Coquina Clams (Donax variabilis Say, 1822) During the Archaic and St. Johns Periods in Coastal Northeast Florida. In Archaeomalacology: Molluscs in Former Environments of Human Behaviour, edited by D. E. Bar‑Yosef Mayer, pp. 18‑28. Oxbow Books, Oxford.

Quitmyer, I. R., D. S. Jones, and W. S. Arnold 1997 The Sclerochronology of Hard Clams, Mercenaria spp. from the South‑Eastern U.S.A.: A Method of Elucidating the Zooarchaeological Records of Seasonal Resource Procurement and Seasonality in Prehistoric Shell Middens. Journal of Archaeological Science 24(9):825‑840.

Quitmyer, I. R., D. S. Jones, and N. Borremans 1990 Annual Incremental Shell Formation in the Southern Quahog, Mercenaria campechiensis, from Suwanee Reef, Florida and its use in Determining the Paleoseasonality of Archaeological Shell Middens. Florida Scientist 53(1):7.

226 Rafferty, J. 1985 The Archaeological Record on Sedentariness: Recognition, Development, and Implication. In Advances in Archaeological Method and Theory, edited by M. B. Schiffer, pp. 113‑156. vol. 8. Academic Press, New York.

Rafinesque, C. S. 1815 Analyse de la Nature: ou Tableau de lʹUnivers et des corps Organisés. Palerme.

Randall, A. R. 2008 Archaic Shell Mounds of the St. Johns River, Florida. The SAA Archaeological Record 8(5):13‑17.

Rehder, H. A. 1981 National Audubon Society Field Guide to North American Shells. Knopf, New York.

Reitz, E. J. 2008 Nonhuman Vertebrate Remains. In Native American Landscapes of St. Catherines Island, Georgia, edited by D. H. Thomas, pp. 615‑665. American Museum of Natural History, Anthropological Papers No. 88, Part II, New York.

Reitz, E. J., B. Pavao‑Zuckerman, D. C. Weinand, and G. A. Duncan 2009 Mission and Pueblo of Santa Catalina de Guale St. Catherines, Island Georgia. A Comparative Zooarchaeological Analysis. American Museum of Natural History, Anthropological Papers Number 91, New York.

Reitz, E. J., and E. S. Wing 1999 Zooarchaeology. Cambridge University Press, Cambridge.

Rocek, T. R. 1998 Pithouses and Pueblos on Two Continents: Interpretations of Sedentism and Mobility in the Southwestern United States and Southwest Asia. . In Seasonality and Sedentism: Archaeological Perspectives from Old and New World Sites, edited by T. R. Rocek and O. Bar Yosef. Peabody Museum Bulletin 6. Peabody Museum of Archaeology and Ethnology, Harvard University, Cambridge, Massachusetts.

Rollins, H. B., and D. H. Thomas 2009 Geomorphology, Sea‑Level, and Marine Resources: St. Catherines Island, Georgia. Paper presented at the Fourth Caldwell Conference, St. Catherines Island, Georgia.

227 2011 Geomorphology, Sea Level, and Marine Resources: St. Catherines Island. In Geoarchaeology of St. Catherines Island, Georgia, edited by G. A. Bishop, H. B. Rollins, and D. H. Thomas, pp. 319‑337. American Museum of Natural History, Anthropological Papers Number 94, New York.

Russo, M. 1991 Archaic Sedentism on the Florida Coast: A Case Study from Horrʹs Island. Ph.D. Dissertation, Department of Anthropology, University of Florida, Gainesville. 1998 Measuring Sedentism with Fauna: Archaic Cultures along the Southwest Florida Coast. In Seasonality and Sedentism: Archaeological Perspectives from Old and New World Sites, edited by T. R. Rocek and O. Bar‑Yosef, pp. 143‑164. Peabody Museum Bulletin 6. Peabody Museum of Archaeology and Ethnology, Cambridge, Massachusetts.

2004a Measuring Shell Rings for Social Inequality. In Signs of Power, edited by J. L. Gibson and P. J. Carr, pp. 26‑70. University of Alabama, Tuscaloosa.

2004b Notes on Florida and South Carolina Shell Rings. Ms. on file, National Park Service, Southeast Archeological Center, Tallahassee, Florida.

2006 Archaic Shell Rings of the Southeast: National Historic Landmarks Historical Context. Ms. on file, National Park Service, Southeast Archeological Center, Tallahassee, Florida.

2008 Late Archaic Shell Rings and Society in the Southeast U.S. The SAA Archaeological Record 8(5):18‑22.

Russo, M., A. Cordell, L. Newsom, and R. Austin 1989 Phase III Archaeological Excavations at Edgewater Landing, Volusia County Florida. . Piper Archaeological Research, Inc., St. Petersburg, Florida.

Russo, M., A. S. Cordell, and D. L. Ruhl 1993 The Timucuan Ecological and Historic Preserve Phase III Final Report. Report submitted to the Florida Museum of Natural History, Gainesville

Russo, M., and G. Heide 2001 Shell Rings of the Southeast U.S. Antiquity 75:491‑492.

2002 The Joseph Reed Shell Ring. The Florida Anthropologist 55:67‑87.

Russo, M., G. Heide, and V. Rolland 2002 The Guana Shell Ring. Historic Preservation Grant No. F0126. Report submitted to the Florida Department of State Division of Historical Resources, Tallahassee.

228 Russo, M., and I. R. Quitmyer 1996 Sedentism in Coastal Populations of South Florida. In Case Studies in Environmental Archaeology, edited by E. J. Reitz, L. A. Newsom, and S. J. Scudder, pp. 215‑232. Plenum Press, New York.

Russo, M., and R. Saunders 1999 Identifying the Early Use of Coastal Fisheries and the Rise of Social Complexity in Shell Rings and Arcuate Middens on Floridaʹs Northeast Coast. Report Submitted to the National Geographic Society, Washington, D.C.

Russo, M., and D. Ste. Claire 1992 Tomoka Stone: Archaic Period Coastal Settlement in East Florida. The Florida Anthropologist 45(4):336‑346.

Sassaman, K. E. 2008 The New Archaic, It Ainʹt What it used to be. The SAA Archaeological Record 8(5):6‑8.

Saunders, R. 2002 The Fig Island Ring Complex (38CH42): Coastal Adaptation and the Question of Ring Function in the Late Archaic. Grant No. 45‑01‑16441. Report Submitted to the South Carolina Department of Archives and History, Columbia.

2004 The Stratigraphic Sequence at Rollins Shell Ring: The Implications for Ring Function. The Florida Anthropologist 57:249‑268.

Saunders, R., and V. Rolland 2006 Exploring the Interior of the Guana River Shell Ring (8SJ2554). Report Submitted to the Florida Division of Historical Resources, Tallahassee.

Saunders, R., and M. Russo 2010 [1986] Quahog (Mercenaria mercenaria) Seasonality at Meeting House Field. In Native American Landscapes of St. Catherines Island, Georgia, edited by D. H. Thomas. American Museum of Natural History, Anthropological Papers Number 88, Part II, New York.

Say, T. 1822 Account of some of the Marine Shells of the United States. Journal of the Academy of Natural Sciences of Philadelphia 4(1):124‑155.

Schroeder, W. C. 1924 Fisheries of Key West and the Clam Industry of Southern Florida. U.S. Bureau of Fisheries, Document 962.

229 Schumacher, C. F. 1817 Essai dʹun Noveau Systeme des Habitations des Vers Testaces. A Copenhague: De Iʹlmprimerie de Mr. Schultz, Copenhagen.

Shackleton, N. J. 1973 Oxygen Isotope Analysis as a Means of Determining Season of Occupation of Prehistoric Midden Sites. Archaeometry 15(133‑141).

Simons, E. H. 1986 A Guide for Identifying Otoliths from Archaeological Sites. Southeastern Archaeology 5(2):138‑145.

Smith, B. 1986 The Archaeology of the Southeastern United States: From Dalton to de Soto, 10,500‑500 B.P. In Advances in World Archaeology, edited by F. Wendorf and A. E. Close, pp. 1‑92. Academic Press, Orlando.

Steponaitis, V. P. 1986 Prehistoric Archaeology in the Southeastern United States, 1970‑1985. Annual Review of Anthropology 15:363‑404.

Stuiver, M., and P. J. Reimer 1993 Extended 14C Database and Revised CALIB Radiocarbon Calibration Program. Radiocarbon 35:215‑230.

Thomas, D. H. 2008a Marine Foraging on St. Catherines Island. In Native American Landscapes of St. Catherines Island, Georgia, edited by D. H. Thomas, pp. 74‑135. American Museum of Natural History, Anthropological Papers Number 88, Part I, New York.

2008b Aboriginal Foraging Strategies on St. Catherines Island. In Native American Landscapes of St. Catherines Island, Georgia, edited by D. H. Thomas, pp. 62‑73. American Museum of Natural History, Anthropological Papers Number 88, Part I, New York.

2008c Late Holocene Sea Levels and the Changing Archaeological Landscape of St. Catherines Island. In Native American Landscapes of St. Catherines Island, Georgia, edited by D. H. Thomas, pp. 42‑47. American Museum of Natural History, Anthropological Papers Number 88, Part I., New York.

2008d Native American Landscapes of St. Catherines Island, Georgia. American Museum of Natural History, Anthropological Papers Number 88, Part I, New York.

230 2009 Why This Archaeologist Cares about Geoarchaeology: The Past and Future of St. Catherines Island. Paper presented at the Fourth Caldwell Conference, St. Catherines Island, Georgia.

2011 Why this Archaeologist Cares about Geoarchaeology: some Pasts and Futures of St. Catherines Island. In Geoarchaeology of St. Catherines Island, Georgia, edited by G. A. Bishop, H. B. Rollins, and D. H. Thomas, pp. 25‑66. American Museum of Natural History, Anthropological Papers Number 94, New York.

Thompson, V. D. 2006 Questioning Complexity: The Prehistoric Hunter‑Gatherers of Sapelo Island, Georgia. Ph.D. Dissertation, Department of Anthropology, University of Kentucky, Lexington.

2007 Articulating Activity Areas and Formation Processes at the Sapelo Island Shell Ring Complex. Southeastern Archaeology 26(1):91‑107.

Thompson, V. D., and F. T. Andrus 2011 Evaluating Mobility, Monumentality, and Feasting at the Sapelo Island Shell Ring Complex. American Antiquity 76(2):314‑344.

Thompson, V. D., and J. A. Turck 2009 Adaptive Cycles of Coastal Hunter‑Gatherers. American Antiquity 74(2):255‑278.

Trinkley, M. B. 1985 The Form and Function of South Carolinaʹs Early Woodland Shell Rings. In Structure and Process in Southeastern Archaeology, edited by R. S. Dickens, Jr. and H. T. Ward, pp. 102‑118. University of Alabama Press, Tuscaloosa.

Tuross, N., M. L. Fogel, L. Newsom, and G. H. Doran 1994 Subsistence in the Florida Archaic: The Stable‑Isotope and Archaeobotanical Evidence from the Windover Site. American Antiquity 59(2):288‑303.

Van der Schriek, T., D. G. Passmore, A. C. Stevenson, and J. M. Rolão 2007 The Influence of Environmental Change on Mesolithic Settlement‑Subsistence and Shell Midden Formation along the Lower Tagus River, Portugal. In Shell Middens in Atlantic Europe, edited by N. Milner, O. E. Craig, and G. Bailey, pp. 165‑182. Oxbow Books, Oxford.

Walker, K. J., F. W. Stapor, and W. H. Marquardt 1995 Archaeological Evidence for a 1750‑1450 BP Higher‑than‑Present Sea Level Along Floridaʹs Gulf Coast. Journal of Coastal Research 17:205‑281.

231 Waring, A. J., Jr. 1968 The Archaic Hunting and Gathering Cultures ‑ the Archaic and Some Shell Rings. In The Waring Papers: The Collected Works of Antonio J. Waring, Jr., edited by S. Williams, pp. 243‑246. Papers of the Peabody Museum of Archaeology and Ethnology, vol. 58, Harvard University, Cambridge, Massachusetts.

1968 The Bilbo Site, Chatham County, Georgia. In The Waring Papers: The Collected Works of Antonio J. Waring, Jr., edited by S. Williams, pp. 152‑197. Papers of the Peabody Museum of Archaeology and Ethnology, vol. 58, Harvard University, Cambridge, Massachusetts.

Waring, A. J., Jr., and L. H. Larson, Jr. 1968 The Shell Ring on Sapelo Island. In The Waring Papers: The Collected Works of Antonio J. Waring, Jr., edited by S. Williams. , pp. 263‑278. Papers of the Peabody Museum of Archaeology and Ethnology, vol. 58, Harvard University, Cambridge, Massachusetts.

Waselkov, G. A. 1987 Shellfish Gathering and Shell Midden Archaeology. Advances in Archaeological Method and Theory, edited by M. B. Schiffer, pp. 93‑210. vol. 10. Academic Press, New York.

Weide, M. L. 1969 Seasonality of Pismo Clam Collecting at ORA‑82. University of California at Los Angeles, Archaeological Survey Annual Report 11:131‑141.

Wheeler, A., and A. K. G. Jones 1989 Fishes. Cambridge University Press, Cambridge, Great Britain.

Wheeler, R. J., J. J. Miller, R. M. McGee, D. L. Ruhl, B. Swann, and M. Memory 2003 Archaic Period Canoes from Newnans Lake, Florida. American Antiquity 68(3):533‑551.

Whetstone, J. M., L. N. Sturmer, and M. J. Oesterling 2005 Biology and Culture of the Hard Clam (Mercenaria mercenaria). Southern Regional Aquaculture Center, SRAC Publication Number 433.

Widmer, R. J. 1988 The Evolution of the Calusa. University of Alabama Press, Tuscaloosa.

Willey, G. R., and P. Phillips 1958 Method and Theory in American Archaeology. University of Chicago Press, Chicago.

232 Wing, E. S., and I. R. Quitmyer 1992 A Modern Midden Experiment. In Culture and Environment in the Domain of the Calusa, edited by W. H. Marquardt, pp. 367‑373. Institute of Archaeology and Paleoenvironmental Studies, Monograph Number 1, University of Florida, Gainesville, Florida.

Wyman, J. 1875 Fresh‑Water Shell Mounds of the St. Johns River, Florida. Peabody Academy of Science, Memoir 1(4):1‑94., Salem, Massachusetts.

Yesner, D. R. 1980 Maritime Hunter‑Gatherers: Ecology and Prehistory. Current Anthropology 21:727‑750.

1984 Population Pressure in Coastal Environments: An Archaeological Test. World Archaeology 16(1):108‑127.

233 BIOGRAPHICAL SKETCH

Alexandra Parsons graduated with a B.A. in Anthropology from University of Central Florida in Orlando in 2004. Alex has participated in archaeological fieldwork in Florida, Georgia,

Virginia, England, Spain, and Hungary. Alex began graduate work at Florida State University in 2005 under the guidance of Dr. Glen Doran and Dr. Rochelle Marrinan. Alex received her

M.S. in Anthropology (with a focus in Archaeology) from Florida State University in

Tallahassee in 2008. Her master’s thesis evaluated clam seasonality at the Grand Shell Ring, the only Mississippi period shell ring known to date. Alex will receive her Ph.D. in Anthropology from Florida State University in 2012 under the advisement of Dr. Marrinan.

Alex received several grants for her graduate research. These include a National Science

Foundation Dissertation Improvement Grant, two Eisele Foundation grants for dissertation research, and the John W. Griffin Student Grant for her master’s thesis research.

During her graduate tenure, Alex presented research papers at conferences including the Annual Meeting of the Southeastern Archaeological Conference, the Annual Meeting of the

Society for American Archaeology, the Caldwell Conference held on St. Catherines Island,

Georgia, and the Annual Meeting of the South Georgia Archaeological Research Team. Alex has also presented her research in public forums at Mission San Luis and at meetings of the

Panhandle Anthropological Society at Tallahassee (branch of the Florida Anthropological

Society). Alex has been involved with many public outreach events in the Tallahassee area, including visits to local elementary schools (for lectures and mock‑dig events), workshops provided for the public and for anthropology undergraduate students, a forensic workshop for the SCI‑GIRLS summer science camp, and she served as a judge for the Florida History Fair.

During her graduate career, Alex co‑authored two reports with Dr. Marrinan detailing zooarchaeological analysis; these include reports on the Spring Warrior and Garden Patch Sites and Pickalene Midden on St. Vincent Island. In 2008, Alex co‑authored a report detailing fieldwork in Virginia with her husband, Dr. Timothy Parsons. Alex also conducted clam

234 seasonality research at several sites in the Canaveral National Seashore, and her report detailing this work is currently under review.

Alex and Dr. Marrinan have submitted an article to be published in the Anthropological

Papers of the American Museum of Natural History. This paper evaluates coastal faunal assemblages in Coastal Georgia and northeast Florida. Alex’s current research includes evaluating clam seasonality at Turtle Mound and Castle Windy in Canaveral National Seashore.

She also plans to analyze the vertebrate faunal remains from Guana (excavated in 2001 and

2005) in the upcoming year.

Alex grew up in West Palm Beach, Florida and currently resides in Tallahassee, Florida.

She is married to Dr. Timothy Parsons, who also received his Ph.D. in Anthropology from

Florida State University. Alex is the proud mother of Josephine Parsons, born on March 31,

2012. Alex has been employed at the Southeast Archeological Center of the National Park

Service in Tallahassee, Florida since 2006.

235