Fort San Juan and the Limits of Empire: Colonialism and Household Practice at the Berry Site 0813061598, 9780813061597

Citation preview

Fort San Juan and the Limits of Empire Florida Museum of Natural History: Ripley P. Bullen Series

University Press of Florida Florida A&M University, Tallahassee Florida Atlantic University, Boca Raton Florida Gulf Coast University, Ft. Myers Florida International University, Miami Florida State University, Tallahassee New College of Florida, Sarasota University of Central Florida, Orlando University of Florida, Gainesville University of North Florida, Jacksonville University of South Florida, Tampa University of West Florida, Pensacola

This page intentionally left blank

Fort San Juan and the Limits of Empire Colonialism and Household Practice at the Berry Site

Edited by Robin A. Beck, Christopher B. Rodning, and David G. Moore

University Press of Florida Gainesville · Tallahassee · Tampa · Boca Raton Pensacola · Orlando · Miami · Jacksonville · Ft. Myers · Sarasota

Copyright 2016 by Robin A. Beck, Christopher B. Rodning, and David G. Moore All rights reserved Printed in the United States of America on acid-free paper This book may be available in an electronic edition. 21 20 19 18 17 16

6 5 4 3 2 1

A record of cataloging-in-publication data is available from the Library of Congress. ISBN 978-0-8130-6159-7 The University Press of Florida is the scholarly publishing agency for the State University System of Florida, comprising Florida A&M University, Florida Atlantic University, Florida Gulf Coast University, Florida International University, Florida State University, New College of Florida, University of Central Florida, University of Florida, University of North Florida, University of South Florida, and University of West Florida. University Press of Florida 15 Northwest 15th Street Gainesville, FL 32611-2079 http://www.upf.com

To the memory of Patsy Nesbitt Berry (1935–2012) and Charles Melvin Hudson Jr. (1932–2013)

This page intentionally left blank

Contents

List of Figures ix List of Plates xiii List of Tables xv Acknowledgments xvii I. Joara, Cuenca, and Fort San Juan 1 1. Introduction 5 Robin A. Beck, David G. Moore, and Christopher B. Rodning

II. Who They Were: Situating the Colonial Encounter 27 2. Joara in Time and Space 31 Robin A. Beck, David G. Moore, and Christopher B. Rodning

3. Recollections of the Juan Pardo Expeditions: The 1584 Domingo de León Account 58 John E. Worth

III. Where They Lived: Household Archaeology at Fort San Juan 81 4. The Built Environment of the Berry Site Spanish Compound 85 Robin A. Beck, David G. Moore, Christopher B. Rodning, Sarah Sherwood, and Elizabeth T. Horton

5. Wood Selection and Technology in Structures 1 and 5 150 Lee Ann Newsom

IV. What They Ate: Politics, Food, and Provisioning 233 6. People, Plants, and Early Frontier Food 237 Gayle J. Fritz

7. Fauna, Subsistence, and Survival at Fort San Juan 271 Heather A. Lapham

V. What They Carried: Material Culture and Household Practice 301 8. Spanish Material Culture from the Berry Site 303 Christopher B. Rodning, Robin A. Beck, David G. Moore, and James Legg

9. Native Material Culture from the Spanish Compound 341 David G. Moore, Christopher B. Rodning, and Robin A. Beck

VI. What They Left Behind: Fragments of the Colonial Encounter 369 10. Conclusions 373 Robin A. Beck, Christopher B. Rodning, and David G. Moore

References Cited 383 List of Contributors 411 Index 413

Figures

1.1. Route of Juan Pardo’s first expedition, 1566–1567 6 1.2. Hernando Moyano’s forays in the Appalachian Mountains, 1567 9 1.3. Towns visited by Juan Pardo’s second expedition, 1567–1568 11 1.4. Berry site excavations in 1986, looking north 20 1.5. Berry site excavations, 1986–2013 24 2.1. Berry site in topographic setting 33 2.2. Major Burke phase sites mentioned in text 34 2.3. Lamar-style complicated-stamped jar rims from Berry site 35 2.4. Lamar-style incised cazuela bowl from Berry site 36 2.5. Native towns and phases in Carolina Piedmont, A.D. 1450–1600 37 2.6. Plan view of Structures 1 and 2 at Ensley site 44 2.7. T. F. Nelson Mound, Caldwell County 46 2.8. Iron artifacts and Citico gorget from T. F. Nelson Triangle, Caldwell County 49 2.9. Large Burke phase settlements near Berry site 51 2.10. Histogram of Burke phase sites along Upper Creek/Warrior Fork 52 3.1. Domingo de León’s “mental map” of La Florida’s physical and social geography 74 4.1. Plan map of Berry site Spanish compound 86 4.2. Structure 3 after plow zone removal, 2003 87 4.3. Excavation units in Structure 1, 2003 89 4.4. Excavations in Structure 1, 2003 90 4.5. Structure 5 after plow zone removal, 2007 91 4.6. Structure 5 excavation units 92 4.7. Excavating plow scars over Structure 5, 2007 93 4.8. Rivercane (Arundinaria sp.) culms in Structure 5 95

x · Figures

4.9. Comparison of primary culm nodes in rivercane 96 4.10. Steelyard scale “hook” fragment from Structure 5, Unit 54 99 4.11. Small quartz crystal from Structure 5, Unit 54 100 4.12. Feature 91, central hearth in Structure 5 101 4.13. Plan map of Structure 5 at top of subsoil 102 4.14. Detail of posthole, post mold, and carbonized post in Structure 5, Unit 2 103 4.15. Structure 1 excavation units 105 4.16. East-west profile across south corner of Structure 1 107 4.17. North-south profile across south corner of Structure 1 107 4.18. Feature 119, central hearth in Structure 1 110 4.19. Excavations in entryway of Structure 1 112 4.20. Profile and cross section of Structure 1 entry trenches 114 4.21. Plan map of Structure 1 at top of subsoil, showing posthole distribution 115 4.22. Carbonized base of western central support post, OG#21 116 4.23. Excavation of OG#21/61 posthole 116 4.24. Plan map of Berry site Spanish compound showing features with European artifacts 118 4.25. Feature 23 119 4.26. Feature 23 profiles 120 4.27. Feature 25 121 4.28. Feature 25 profiles 122 4.29. Feature 38 123 4.30. Feature 64, 68, 69, and 83 cluster at base of plow zone 125 4.31. Features 64 and 68 profiles 125 4.32. Features 69 and 83 profiles 126 4.33. Feature 92 127 4.34. Feature 112 128 4.35. Western feature and midden complex 129 4.36. Feature 66 130 4.37. Feature 71 132 4.38. Plan map of Spanish compound, first-phase buildings and features 136 4.39. Feature 76 after excavation 138

Figures · xi

4.40. Plan map of Spanish compound, second-phase buildings and features 140 4.41. Geometry linking Structure 1, Structure 2, Structure 5, and Feature 76 144 4.42. Plan map of second-phase Spanish compound 145 4.43. Square postholes 1109 and 1176 146 4.44. Plan map of second-phase Spanish compound showing plaza and square post clusters 147 5.1. Structure 1 floor plan with organic samples plotted 155 5.2. Structure 5 floor plan with organic samples plotted 156 5.3. Histogram comparing roundwood diameters from Structures 1 and 5 190 5.4. Age classes for roundwood, including posts, from Structures 1 and 5 196 5.5. OG#21c (Structure 5) 201 5.6. OG#21h (Structure 5) 201 5.7. OG#25 (Structure 5) 202 5.8. OG#48b (Structure 5) 203 5.9. OG#3 (Structure 1) 204 5.10. OG#19i (Structure 1) 205 5.11. OG#44 (Structure 1) 206 5.12. OG#32b (Structure 1) 206 5.13. OG#84 (Structure 1) 207 5.14. Structure 1 timbers OG#78 and OG#81 208 5.15. OG#81 (Structure 1) 208 5.16. OG#10 (Structure 1) 209 5.17. OG#39 (Structure 1) 210 5.18. OG#20b (Structure 1) 213 6.1. Ubiquity values of maize and nutshell 243 6.2. Density values of maize and nutshell 246 6.3. Maize kernel-to-cob ratio values across contexts at Berry site 250 6.4. Percentages of food plant remains, early and late pit features 260 6.5. Box plots for maize, hickory nut, and acorn shell fragments, early and late pit features 261 6.6. Nutshell frequencies in Spanish compound compared to other sites in region 264

xii · Figures

6.7. Intersite comparison of kernel-to-cob ratios 266 7.1. Main taxa at Berry mound and Spanish compound 279 7.2. Main mammals at Catawba Meadows village, Berry mound, and Spanish compound 288 7.3. Meat-bearing limbs and butchery waste for white-tailed deer and black bear 292 7.4. Main mammals in early and late features in Spanish compound 295 7.5. Meat-bearing limbs and butchery waste in Spanish compound features for white-tailed deer and black bear 296 8.1. Spanish olive jar fragments 315 8.2. Spanish pottery 318 8.3. Chain mail links and fragments 322 8.4. Iron jack-plate fragment from posthole in Structure 5 323 8.5. Lead artifacts 324 8.6. Definitive Spanish nails and tacks 326 8.7. Possible Spanish nails (category B in Legg’s analysis) 327 8.8. Possible sheet iron pan 328 8.9. Iron clothing fasteners and ball button 329 8.10. Copper or brass aglets, tinkling cones, and beads 330 8.11. Copper or brass bell fragments 331 8.12. Miscellaneous copper or alloy fragments 332 8.13. Glass beads 334 8.14. Iron knife from Burial 1 336 8.15. Wrought iron fragment recovered during troweling of Structure 5 337 8.16. Unidentified iron fragments from structures and features 337 9.1. Geographic distribution of Burke series pottery 349 9.2. Vessel forms 353 9.3. Features 129 and 145 with exposed pottery 357 9.4. Pisgah vessels from Feature 112 358 9.5. Pisgah and Dallas vessels 359 9.6. Ceramic elbow pipes 366

Plates

Plates follow page 154. 1. Plan map of the Berry site Spanish compound 2. Photomosaic of Structure 5, base of plow zone 3. Photomosaic of Structure 5, top of Zone 3 4. Plan map of organic remains in Zone 3, Structure 5 5. Photomosaic of Structure 5, top of Zone 4/subsoil 6. Removing plow zone above Structure 1, 2008 7. Photomosaic of Structure 1, top of Zone 4 8. Plan map of organic remains in Zone 4, Structure 1 9. Plan map of organic remains in Zone 4, Structure 1, with small wood removed 10. Photomosaic of Structure 1, top of subsoil 11. Thin section of sediments from Structure 1, showing clay papules added through levigation 12. Thin section of sediments from hearth area in Structure 1, showing clay papules added through levigation

This page intentionally left blank

Tables

5.1. Archaeological wood taxa from Structures 1 and 5 159 5.2. Upright posts from Structure 1 162 5.3. Upright posts from Structure 5 163 5.4. Nonpost wood samples from Structure 1 165 5.5. Nonpost wood samples from Structure 5 176 5.6. Growth ring observations and harvest periods for Structure 1 samples 190 5.7. Growth ring observations and harvest periods for Structure 5 samples 197 5.8. Proportion classes for roundwood 200 5.9. Tool mark measurements 204 5.10. Tertiary construction elements from Structures 1 and 5 211 5.11. Charcoal from fuelwood contexts 215 6.1. Macrobotanical remains from the Berry site Spanish compound 244 6.2. Breakdown by context of maize fragments from the Berry site Spanish compound 249 6.3. Nonfruit seeds from the Berry site 255 6.4. Intersite comparison of maypops and grape seeds 265 7.1. Animal remains from the Spanish compound and mound contexts at the Berry site 280 7.2. Animal remains from the Catawba Meadows site 287 7.3. White-tailed deer and black bear body portions 290 7.4. White-tailed deer and black bear body portions in Spanish compound features 297 8.1. Sixteenth-century material culture from Santa Elena 307

xvi · Tables

8.2. Sixteenth-century items imported to La Florida 308 8.3. Supplies issued to Captain Juan Pardo’s expeditions 310 8.4. Supplies issued to Captain Juan Pardo’s forts 311 8.5. Sixteenth-century Spanish-related artifacts from the Berry site 316 8.6. Glass beads from the Berry site 335 9.1. Distribution of Native American artifacts from excavated features in the Spanish compound 344 9.2. Vessel number and form from Spanish compound features 351

Acknowledgments

Our joint research at the Berry site began in 2001, and we have returned to continue work there each subsequent year. Most of the data that we analyze and discuss in the pages that follow were recovered during the 2001 to 2010 seasons from that part of the site that we refer to as the Spanish compound. There are many institutions and individuals whose contributions have made our work not only possible but a joy, and we would like to acknowledge them here. First, we want to thank the entire Berry family for their stewardship of this incredible site, but we especially want to recognize James and Pat Berry, together with their son, Burton, and his family. Pat Berry passed away in 2012, and we miss the charm and wit that accompanied her near-daily visits to the site with James during our field seasons. If the Berry family has enabled the practical side of our work, then our greatest intellectual debt is to Charles Hudson. It is safe to say—in more ways than we can recount here—that Charlie was the force that brought all of us to Berry. Although none of us (Beck, Moore, and Rodning) was ever formally Charlie’s student, we have all benefitted immensely from his guidance, encouragement, and unbounded enthusiasm through the years. Our debt to him is both professional and personal. Sadly, Charlie, too, passed away before we were able to bring the book to publication. We therefore dedicate it to these two people—Pat and Charlie—who have meant the world to us and whom we miss dearly. Since 2001, several institutions have provided our academic homes, and each of these has supported our work in myriad ways: the University of Michigan, Warren Wilson College, Tulane University, the University of Oklahoma, Southern Illinois University–Carbondale, Northwestern University, and the University of North Carolina–Chapel Hill. Other support and funding for our work in the Spanish compound has come from the National Science Foundation (BCS-0542120); the National Geographic So-

xviii · Acknowledgments

ciety Committee on Research and Exploration; the National Park Service Certified Local Governments Grant to the City of Morganton, administered by the North Carolina Division of Archives and History, Department of Cultural Resources; and the Burke County Historical Society, the Historic Burke Foundation, and the Woodbury Foundation. Finally, the School for Advanced Research in Santa Fe, New Mexico, hosted several of our project members for a 2009 Research Team Short Seminar, the results of which include this book. Throughout our project, Warren Wilson College has served as our administrative host institution and supported the Archaeology Work Crew and Laboratory. Dave would especially like to thank Virginia McKinley, Doug Orr, Larry Modlin, John Casey, Ian Richardson, Mary Davis, Steve Solnick, Paula Garrett, Christa Bridgman, and Richard Blomgren for their support. At the University of Michigan, Rob would like to express his thanks to Joyce Marcus and Tom Fricke for all of their encouragement during the writing of this book. Chris acknowledges many supporters at Tulane University, including Will Andrews, Dan Healan, and Kit Nelson. Each field season at Berry, we have been fortunate to work alongside an exemplary staff, without whose commitment and dedication we could not have accomplished what we have. Many thanks to you all: 2002: Scott Ashcraft, Megan Best, Aaron Brummit, Emily Dale, Caroline Ketron, and James Lang; 2003: Megan Best, Karla Evans, Ellie Haywood, Crickett Hefner, Caroline Ketron, James Lang, and Jessica Smeeks; 2004: Megan Best, Rachel Briggs, Karla Evans, Crickett Hefner, Jesse Mitchell, and Johanna Vasek; 2005: Megan Best, Rachel Briggs, Will Gulley, Liz Horton, Lotte Govaerts, and Lauren Zych; 2006: Jessica Baker, Lotte Govaerts, Crickett Hefner, Liz Horton, Jesse Mitchell, Johanna Vasek, and Lauren Zych; 2007: Annie Blankenship, Rachel Briggs, David Cranford, Emily Dale, Elsbeth Dowd, Andrea Glenn, Lotte Govaerts, Lorie Hanson, Crickett Hefner, Liz Horton, Hannah Humphrey, Caroline Ketron, Patricia Overman, Emma Richardson, Merritt Sanders, and Johanna Vasek; 2008: Annie Blankenship, Rachel Briggs, Andrea Glenn, Lotte Go­ vaerts, Lorie Hanson, David Heavner, Crickett Hefner, Liz Horton, Hannah Humphrey, Caroline Ketron, Elizabeth Martin, Trevor

Acknowledgments · xix

Martin, Adam Moody, Kevin Moody, Robert Pagnuzzi, Emma Richardson, Laurel Sanders, Merritt Sanders, and Johanna Vasek; 2009: Rachel Briggs, David Cranford, Lorie Hanson, Trevor Martin, Emma Richardson, and Merritt Sanders; 2010: David Cranford, Allyson Gardner, David Heavner, Abra Johgart, Caroline Ketron, Theresa McReynolds, Casey Monahan, Kate O’Mara, Emma Richardson, and Merritt Sanders; 2011: Rachel Bissell, Allyson Gardner, Brian Geiger, David Heavner, Abra Johgart, Kat Kipfer, Theresa McReynolds, Casey Monahan, and Emma Richardson; 2012: Rachel Bissell, Jessica Dalton-Carringer, Brian Geiger, David Heavner, Abra Johgart, Kat Kipfer, Shaun Lynch, Theresa McReynolds, and Emma Richardson; 2013: David Heavner, Abra Johgart, and Kat Kipfer; 2014: Jess Cantrell, David Heavner, Abra Johgart, Kat Kipfer, Caroline Loveland, and Travis Williams. In addition, each year the Archaeology Crew at the Warren Wilson College Archaeology Lab is responsible for washing and cataloguing the field school materials. These undergraduate students have assisted with creating and curating all of our project documentation, maps, and photographs and have also conducted a variety of lab analyses on the Berry material. Their assistance is immeasurable, and we thank them all for helping to sustain the project year after year: 2001–2: Megan Best, Emily Dale, Rachel Horne, Zoey Keefer-Norris, Isabel Salazar, and Will Spoon; 2002–3: Megan Best, Emily Dale, Laurel Key, Michelle Kozma, James Lang, and Rose Orleans; 2003–4: Emily Dale, Laurel Key, James Lang, Rose Orleans, and Johanna Vasek; 2005–6: Candace Anthony, Gretchen Fitzgerald, Andrea Glenn, Will Gulley, Jesse Mitchell, and Johanna Vasek; 2006–7: Lindsay Butenhof, Gretchen Caverly, Andrea Glenn, Lotte Govaerts, Will Gulley, Peter Lanier, Niels Nugent, Emma Richardson, and Shannon Saville; 2007–8: Andrea Glenn, Niels Nugent, Emma Richardson, and Laurel Sanders; 2008–9: Lindsay Butenhof, Allyson Gardner, Abra Johgart, Niels Nugent, Laurel Sanders, and Amy Wagner;

xx · Acknowledgments

2009–10: Lindsay Butenhof, Allyson Gardner, Abra Johgart, Casey Monahan, and Kate O’Mara; 2010–11: Allyson Gardner, Abra Johgart, Mary Dawn Kidd, Casey Monahan, and Maureen Vaughn; 2011–12: Helen Albea, Ellie Ferguson, Abra Johgart, Kat Kipfer, Brandy Oliver, and Maureen Vaughn; 2012–13: Mariah Bruce, Eli Clare, Russell DePratter, Fern Hoffmann, Abra Johgart, and Claire Woerner; 2013–14: Russell DePratter, Fern Hoffmann, Abra Johgart, and Caroline Loveland; 2014–15: Fern Hoffmann, Abra Johgart, and Caroline Loveland. Our project was conceived from the start as a public archaeology project. We wanted to find innovative ways to engage more of the public in the actual work as well as to do the best job we could at sharing the results of our explorations with the general public. We have been rewarded and continue to be motivated by the opportunity of working with over five hundred field school students ranging from high schoolers to retirees, from vacationing business executives to schoolteachers seeking different ways to share their love of history and intellectual growth with their students. These are the folks who do much of the hard work, and often under challenging conditions. We have enjoyed working with every one of you, and without your efforts we could not have carried out a project of this magnitude. Spurred by the interest and support of a large contingent of friends in Morganton and Burke County, we supported the formation in 2008 of Exploring Joara Foundation Inc. (EJF), a not-for-profit organization whose mission is to support public education and outreach regarding archaeology in the western Piedmont region of North Carolina. We are grateful to all of the EJF membership for their interest and support, and we appreciate their efforts to spread their interest in archaeology to the wider community. We especially enjoy the Annual Field Day sponsored by EJF, when we host five hundred to one thousand visitors at the Berry site to show off the yearly excavations. We have many to thank in Morganton, including the City of Morganton, Western Piedmont Community College, and the Western Piedmont Foundation. We especially wish to thank several individuals who provided critical support throughout the years. Linda Wall was instrumental in our early years of investigations and more recently helped to establish the Carl and Linda Wall Center for Archaeological Research in Morganton. Mary

Acknowledgments · xxi

Charlotte Safford and many other staff, faculty, and administrators at Western Piedmont Community College have been ardent field school supporters, as have many other community members, but we would especially like to thank Larry Clark, Clark and Lee Erwin, Patti Matthias, Janet Wilson, Sam Avery, Jim Richardson, Frank and Shirley Sprinkle, Charles and Alice Carey, and Robert and Nicole Vasile. We also owe a debt of gratitude to our many friends and colleagues in the archaeological community who have generously offered their time, encouragement, advice, wisdom, and careful critiques: David Anderson, Doug Bolender, Tony Boudreaux, Jim Brown, John Chenowith (for suggestions regarding the folkloric use of iron), Steve Claggett, Charles Cobb, Steve Davis, Kathy Deagan, Chester DePratter, Tim Earle, Robbie Ethridge, Charles Ewen, Pat Gilman, Dolores Hall, Linda Hall, David Hally, Bennie Keel, T. R. Kidder, Larry Kimball, Jim Knight, Janet Levy, Matt Liebmann, Patrick Livingood, Clay Mathers, Alan May, Maureen Meyers, Paul Minnis, Jeff Mitchem, Dan Morse, Phyllis Morse, Grady Pitts, Brett Riggs, Matt Schmader, Gerald Schroedl, Scott Shumate, Marvin Smith, Stan South, Vin Steponaitis, David Hurst Thomas, Victor Thompson, Jason Ur, Trawick Ward, Steve Wernke, Tom Whyte, and Mark Williams. We would like to make a special acknowledgment of our friend George Stuart, whose passing in 2014 was a profound loss to each of us; we deeply appreciate the generosity of George and his wife, Melinda, during our visits to Boundary End Farm, their home in Barnardsville, North Carolina. We would like to thank the Catawba Indian Nation, particularly Beckee Garris and Winona Haire of the Catawba Cultural Center. Our wonderful editor, Judy Knight, kept us motivated and focused from the beginning of this book project through to its end, all the while showing patience and continued encouragement as our deadlines came and went. Marvin Smith and Charles Ewen, our readers, offered cogent advice and suggestions for improving the manuscript, and we appreciate both the careful reading they gave it and their remarkably speedy turnaround. Finally, our families have seen us through more than a dozen years of fieldwork at the Berry site. They have helped us to pack our bags each summer and then have held down homes in our absence, but just as often they have joined us in the field, our partners and our children alike. Their love for us, and their appreciation for what we do, make it worth doing: Laoma, Soledad, and Cadir Beck; Hope Spencer, Henry, Erik, and Leif Rodning; and Kaitlyn Moore. Our love and thanks to you all for showing us such patience and bringing us such joy.

This page intentionally left blank

I Joara, Cuenca, and Fort San Juan

Deep in the country north and west of Santa Elena, near an Indian town named Joara, was a place where the trail crested a hill and gave the captain his first glimpse of the rugged lands that lay just ahead. Whatever paths wended up and over those steep slopes would no doubt have been precarious for any men so laden with supplies, even in the best of times, but it was late December now, icy and cold, and Pardo shared his company’s disquiet as he regarded the rim of snow-filled mountains on the horizon. They had put a hundred leagues or more behind them already, most of it across fine, open country, but he knew—looking across at the high ridge—that any effort to push beyond its frozen passes with tired, overburdened men would be impractical at best, foolhardy at worst. He was neither. Even so, he would not be bogged down here with so little yet to show for himself. It was nearly dusk. He gave a nod to his sergeant, Moyano, that they should make camp now before night was upon them. In the morning they would arrive at the town the Indians called Joara, and while the company recovered he would take the counsel of his officers before deciding where they should go from there. Standing on the crest of the hill, he was suddenly struck by how much this land in the shadow of the mountains reminded him of Cuenca, of home. He shut his eyes and let himself believe, but for a moment, that the setting sun on his face was that of Castile.

*

*

*

What strikes you, walking across the Berry site, is how remarkably unassuming it is—how unremarkable. It occupies the eastern edge of a field like many

2 · Part I. Joara, Cuenca, and Fort San Juan

others in the western Piedmont of North Carolina, a wide expanse of rich bottomland hemmed in by low uplands, the foothills of the Appalachians. Upper Creek, a tributary of the upper Catawba River as unremarkable in name and appearance as Berry itself, borders the site to both the east and the south, a feat it achieves by making a deep, nearly ninety-degree bend to the west before joining with Irish Creek and flowing on toward the Catawba. The Berry site is bordered on its west by two ornamental tree farms with a cornfield in between; fields both to its north and to its east across Upper Creek are also planted in trees and shrubs. Of all this bottomland at the confluence of Upper and Irish Creeks, more than 200 acres (or 75 hectares), Berry is but a narrow strip that measures about five American football fields in length and just less than one in width, covering an area of 11 acres (or 4.5 hectares). Today, only the lower end of the site is annually plowed for small garden plots of corn, squash, beans, cucumbers, tomatoes, melons, potatoes, and even broccoli. The rest of the site lies fallow. There are no great earthen mounds to climb at the Berry site, no impressive plazas to cross. Compared with some of the better-known, more-photogenic sites in the Eastern Woodlands, you might justifiably walk out onto Berry and think to yourself, there is nothing to see here. But beneath the surface, out of reach of the plows, the story that the Berry site is telling is, indeed, remarkable. Berry is the site of the southeastern Indian town of Joara, also called Xualla, and it was there—on December 27, 1566, the feast day of Saint John the Evangelist—that the Spanish captain Juan Pardo established Fort San Juan and its adjoining colony of Cuenca, named after his hometown in the province of Castile. Fort San Juan was the first European settlement in the interior of what is now the United States. Pardo garrisoned his fort with thirty men, and for much of the eighteen months that followed they enjoyed good relations with their native hosts, who supplied the Spaniards with most of their daily food and provisions. Yet relations between the people of Joara and the men of Fort San Juan had deteriorated badly by the spring of 1568, when the Indians there and at the six other towns across the Carolinas and eastern Tennessee where Pardo built forts rose up and destroyed the Spanish garrisons. Never again would Spain attempt to settle the interior of La Florida, a territory that once included most of the American Southeast, creating a vacuum that English settlers would exploit during the centuries to come. Together, we have explored the archaeology of the Berry site for more than a decade now, focusing much of our attention on the northern end of the site where we first recovered sixteenth-century Spanish ceramics and hardware. Subsequent magnetometry and soil coring here revealed the presence of several burned buildings and numerous large pit features that seemed to con-

Part I. Joara, Cuenca, and Fort San Juan · 3

stitute a well-defined architectural compound. Our extensive excavations in these buildings and features have led us to identify this compound as Cuenca, the settlement that housed Pardo’s doomed men stationed at Fort San Juan. What remained elusive was any evidence of fortifications—where was the fort itself? Finally, in an area south of the compound and where we were working during our 2013 field season, we found a long section of dry moat measuring 10 feet (3 meters) across and nearly 6 feet (2 meters) deep. Exploring the features of Fort San Juan—the public face of the colonial encounter—will consume much of our research agenda in the years to come. What we offer in this volume is thus the private face of the encounter, as revealed through our years of work in the Spanish compound. Its features contain the remains of daily life at Cuenca, while its extraordinarily well-preserved buildings reveal much about relations between Indians and Spaniards and how these relations changed over the course of eighteen months. These data provide us an unparalleled view of household practice on an early colonial frontier. In chapter 1, we summarize Pardo’s two expeditions into the Carolinas, then outline a background to our research at Berry and discuss the research design that has framed this phase of our project.

This page intentionally left blank

1 Introduction Robin A. Beck, David G. Moore, and Christopher B. Rodning

Columbus’s landfall in the Bahamas in 1492 initiated the most dramatic century of cultural exchange in human history. Over two continents, the native peoples of the Americas from Tierra del Fuego to the St. Lawrence River met waves of explorers, settlers, proselytizers, and profiteers from Spanish, English, French, Portuguese, and Dutch centers of colonial aspiration. Of all these nations, Spain was by far the most ambitious in its early efforts at exploration and conquest (Bray 1993; Deagan 2003; Hoffman 1990; Hudson 1990, 1997; Lyon 1976; Thomas 1991; Thomas, ed. 1989, 1990). During the first decades of the sixteenth century, Spain launched several failed attempts to claim the American Southeast. The most famous—or perhaps infamous—of these was the expedition of Hernando de Soto, which cut its ravenous swath across the Southeast from 1539 to 1543. But of the more than six hundred soldiers who began the expedition, more than half—including Soto himself—would ultimately perish in La Florida. Similarly disastrous ventures during this period include the expeditions of Juan Ponce de León (1521), Lucas Vázquez de Ayllón (1526), Pánfilo de Narváez (1528), and Tristán de Luna y Arellano (1559–1560). Finally, after his expulsion of the French from La Florida in 1565–1566, Pedro Menéndez de Avilés successfully founded a pair of small colonies on the South Atlantic coast: San Agustín (modern St. Augustine), founded in September 1565 in what is now northeastern Florida, and Santa Elena, established in April 1566 on modern-day Parris Island, South Carolina. The latter was to be the principal site of Menéndez’s colonial project (Hoffman 1990; Hudson 1990; Lyon 1976; Paar 1999). When Philip II of Spain learned of these developments, he ordered reinforcements for Menéndez’s colony. In July 1566, Captain Juan Pardo, a member of the king’s private guard, arrived at Santa Elena with a contingent of 250 men and soon began to fortify

6 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

Figure 1.1. Route of Juan Pardo’s first expedition, 1566–1567 (after Beck 2013: 71, map 5).

the settlement. Because the Santa Elena colony was not prepared to feed this company of men for very long, Menéndez ordered Pardo to prepare half of his army for an expedition into the interior regions that lay behind the Atlantic coast. His task was to explore the area, to claim its lands for Spain while pacifying local Indians, and to forge an overland road from Santa Elena to the valuable silver mines in Zacatecas, Mexico. Pardo left the colony with 125 men on December 1, 1566, and in the months to follow his army would revisit a small portion of the route that Soto had taken across the Carolinas in 1540 (figure 1.1).

The Juan Pardo Expeditions, 1566–1568 Of this, the first of Pardo’s two expeditions, we have but a single eyewitness account—the brief relation written by Pardo himself. This document provides few details about social relations on the Carolina Piedmont in the post-Soto era, other than the names of places that the expedition vis-

Introduction · 7

ited and how their respective leaders received Pardo and his men. By joining Pardo’s account with accounts of the second expedition, however, we can reconstruct a basic itinerary of the first (DePratter et al. 1983; Hudson 1990). After departing Santa Elena, the entrada army traveled north for a few days across South Carolina’s lightly populated inner Coastal Plain. The first town of note was Guiomae, the same place as Soto’s Himahi (Hudson et al. 1984: 72). Two days later, they arrived at the important town of Canos, where Pardo (1990: 311) reported that he “found a great number of caciques and Indians.” Fortunately, Pardo’s scribe, Juan de la Bandera, made two records of the second expedition and noted in the shorter of these that the Indians called this town “Canosi and, for another name, Cofetazque” (Bandera 1990b: 301). Thus, the place referred to as Canos in the Pardo accounts is probably the same as Soto’s Cofitachequi, one of the most powerful polities that Soto visited east of the Appalachians. As at most of the towns that Pardo entered, he told the cacique and his subjects to build a house for the expedition and to fill it with maize. Pardo and most of the soldiers who accompanied him into the interior could neither speak nor understand any of the Indian languages they encountered. For this reason, the expedition was dependent on the translators and interpreters among the company. One of these was Domingo de León, and John Worth devotes chapter 3 of this volume to León’s background and account of the Pardo expeditions. Another was a young Frenchman, Guillaume Rouffi, whose story was as dramatic as that of any of his fellow explorers (DePratter et al. 1983: 126). In 1562, the sixteen-year-old Rouffi was among the thirty French soldiers Jean Ribault left on modern Parris Island to garrison the newly founded Charlesfort. When winter fell on the fort and the men began to starve, they built a small, crude boat and set sail across the Atlantic; only Rouffi chose to remain behind. Two years later, a Spanish ship’s captain found Rouffi living among the local Orista Indians, having married a daughter of an Orista chief (Rudes et al. 2004: 380). A Huguenot when he left France, Rouffi quickly converted to Catholicism and became Pedro Menéndez’s most trusted interpreter. Pardo continued north from Canos through the Catawba-Wateree Valley, pausing briefly at several small towns before leaving the main valley for the South Fork River, where he came to a town called Ysa. “Ysa” is a variant of “Esaw,” the name applied to one of the core peoples of the Catawba Nation in the early eighteenth century, and transposed as “Nassaw” it referred to one of the most important of all Catawba towns (Merrell 1989: 94). Two or three days north of Ysa, Pardo entered the province of Joara,

8 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

located along a tributary of the upper Catawba River in the Appalachian foothills. During the fifteenth and sixteenth centuries, the upper Catawba River was among the most densely populated regions of the North Carolina Piedmont (chapter 2, this volume; Beck and Moore 2002; Moore 2002). In 1540, Hernando de Soto had passed through this area, and all of the expedition accounts refer to its primary town as “Xualla,” a cognate of “Joara” (e.g., Beck 2013: 69–70). Soto’s army stayed at Xualla only two to four days, however, and had little apparent impact on the town. Pardo’s instruction from Menéndez had been to press across the mountains and forge a route to Mexico. At Joara, however, he saw that there was already snow on the peaks that lay ahead (Martínez 1990: 320), so he halted his westward push to continue exploring the Piedmont. Pardo remained at Joara for fifteen days, writing that its people “demanded Christians from me to catechize them” (1990: 312). There, he constructed a fort that he christened Fort San Juan. Because the area reminded him of his home in Spain, he bestowed the name “Cuenca” on both the Indian town of Joara and the new Spanish settlement (see figure 1.5). Pardo marched northeast through the upper Catawba Valley, spending several days at the towns of Guaquiri and Quinahaqui before leaving the Catawba River. After marching east across an unoccupied region between the Catawba and Yadkin rivers, he and his men arrived at Guatari, probably located in the central Yadkin Valley near modern-day Salisbury, North Carolina. Pardo (1990: 312) found thirty caciques and many Indians waiting to meet with the company at Guatari, a clear sign of this town’s prominence. By the early 1700s, a greatly reduced Guatari (known to the English as Watery) had moved to the lower Wateree Valley, giving this river its name. Pardo stayed at Guatari for just over two weeks, hastily departing when he received news of a possible French military threat to Santa Elena. He left four of his soldiers and the company cleric, Father Sebastian Montero, to catechize the Indians, and it was there that Montero established the earliest mission in the interior of North America (Gannon 1965). Pardo and his remaining army went back to the Catawba-Wateree Valley and passed through several villages south of what is now the line that divides modernday North and South Carolina before returning to Canos; they arrived in Santa Elena on March 7, 1567 (Bandera 1990a: 259). In the six months that Pardo was at Santa Elena, Hernando Moyano— sergeant at Fort San Juan—led devastating attacks on at least two native towns in the Appalachians that were probably enemies of Joara (figure 1.2). Our knowledge of Moyano’s two assaults derives from an account by

Introduction · 9

Figure 1.2. Hernando Moyano’s forays in the Appalachian Mountains, 1567 (after Beck 2013: 77, map 6).

Francisco Martínez, written in July 1567 between Pardo’s first and second expeditions. Thirty days after Pardo made his return to Santa Elena, Moyano sent news claiming that he had defeated “a cacique . . . named Chisca” (Martínez 1990: 320). He wrote, according to Martínez, that he had killed more than a thousand Chisca Indians and burned fifty huts, though the former number is probably exaggerated (Hudson 1990: 48). He also suggested that he could “press further on” if commanded to do so (Martínez 1990: 320). But before Pardo’s orders could arrive, Moyano and his men were accosted by “a cacique of the mountains” who vowed to eat the Spaniards and “a dog the sergeant had” (Martínez 1990: 320). Departing at once, Moyano and nineteen of his soldiers, accompanied by a party of warriors from Joara (Hudson 1990: 48), marched for four days across the mountains, where they were surprised to find “the enemies . . . enclosed by a very high wooden wall” (Martínez 1990: 320). After breaching the palisade, they

10 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

burned the town and, again, according to Martínez, killed 1,500 Indians; this figure, too, was likely exaggerated (Hudson 1990: 48). Until recently, Martínez’s account was our only source describing Moyano’s foray in 1567. But John Worth’s discovery of a 1584 account by Domingo de León, a translator who participated in Pardo’s second expedition, offers a link between Martínez’s account and the testimony of Luisa Méndez and Juan de Ribas before Governor Gonzalo Méndez de Canço at St. Augustine in 1600. Luisa Méndez, who was a native woman taken from the interior during Pardo’s second expedition, reported that there were three to five saltwater springs at the base of the mountains. She described how this salt was processed and recalled that these were the only such springs in all of that land (Hudson 1990: 87). Juan de Ribas, a member of the second expedition who was likely stationed at Joara, testified that Méndez, his wife, was the cacica of Guanaytique, also known as Manaytique (Hudson 1990: 190, 201). Significantly, Domingo de León notes that “Maniatique” and “Guapere” were the mountain villages assaulted and destroyed by Moyano in 1567 (Worth, chapter 3, this volume). Maniatique was likely located along the North Fork of the Holston River near present-day Saltville, Virginia (Beck 1997a). The Saltville locale was one of the most important brines in the American South and one of just three where rock salt is available (Barber and Barfield 2000). An “Indian path” illustrated on the Fry-Jefferson Map of 1751 extends from the upper Holston to the upper Catawba Valley. This path is similarly marked on the 1770 Collet and 1775 Mouzon maps (e.g., Cumming 1966: 23–27). Also, Michael Barber and Eugene Barfield (2000) propose that the Chilhowie High School site (44SM8) was the central town of a “petty chiefdom” in the Saltville locale. Sixteenth-century shell gorgets have been identified in local collections, as have other exotic goods (Barber and Barfield 2000; Wedel 1951: 115, 121). While this location of Maniatique does depend on the assumption that the springs noted by Méndez were located within the bounds of her territory (that is, Maniatique), it seems unlikely that Méndez, a young woman when taken from the interior, would have had such detailed knowledge of a distant brine. After defeating the second of these villages, probably located along the upper Nolichucky or Watagua River in present-day eastern Tennessee, Moyano left along “the road of a great chief that was in that head of the mountain range, who is called Chiaha” (Martínez 1990: 320). In four days’ time, he and his men arrived at a town of Chiaha. Located between “two heavily flowing rivers,” this town was defended by a palisade. Within “were

Introduction · 11

Figure 1.3. Towns visited by Juan Pardo’s second expedition, 1567–1568 (after Beck 2013: 79, map 7).

more than three thousand warriors . . . [and] no other persons, neither women nor children” (Martínez 1990: 320). Although the number of native warriors here is probably exaggerated, Soto met a comparably sized native force at Mabila (Biedma 1993: 233). In any event, Moyano encountered no hostility. Instead, the inhabitants gave Moyano and his company “lots to eat” (Martínez 1990: 320). This outlying town of Chiaha was situated near the junction of the French Broad and either the Nolichucky or the Pigeon River. Moyano left this town by a road “to that chief already mentioned [that is, Chiaha]” (Martínez 1990: 320), and two days later he entered the principal town of that chief. There he quickly built a small fort and waited for Pardo to begin his second expedition. Pardo left again from Santa Elena at Menéndez’s command on September 1, 1567, taking 120 soldiers. As noted, the second expedition was accompanied by an official scribe, Juan de la Bandera, who kept two accounts of the march (figure 1.3); the longer of these is especially useful because of the

12 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

data it offers on interpolity relations in the Piedmont (Beck 2013). Pardo returned to Canos (Cofitachequi) on September 10. Bandera notes that an impressive group of oratas were awaiting Pardo at Canos and that all of these had helped to build a large house for the Spaniards there or else had provided some maize. In the longer of his two accounts, Bandera distinguishes between the terms mico and orata: while the latter refers to a town headman—he recorded no fewer than 120 oratas during the second expedition—the former applied only to a native leader with regional authority (e.g., Hudson 1990: 63). In fact, Bandera explicitly observes that mico was the term used for great lords (un gran señor) and that orata referred to lesser lords (un menor señor) (Bandera 1990a: 215). He recorded only three micos in his longer account of the expedition: Joara Mico, Guatari Mico, and Olamico (Chiaha). Thus, there was no mico present at Cofitachequi in 1567, though this was the only town in the Carolina Piedmont to which Soto had applied the term in 1540. Still, Canos was an important place, given that at least thirteen oratas met there to receive Pardo, including chiefs from Ylasi, Vehidi, Ysa, and Cataba (Bandera 1990a: 260). Ylasi is perhaps the Ylapi of the Soto accounts, to which one of Soto’s officers, Baltasar de Gallegos, led most of the men of the expedition to take some maize that belonged to the cacica of Cofitachequi. Ylasi seems not to have been a subject of Canos in 1567, for while its orata did want credit for helping to build the house for the Spaniards there, he was keeping his maize contribution in a house he had built for that purpose in his own town (Bandera 1990a: 261). “Vehidi” may refer to the people whom the English knew as the Pee Dee Indians (Hudson 1990: 77), who left their name to the Pee Dee River. Two others—the aforementioned Ysa Orata and the one named Cataba—were key constituents of Catawba coalescence during the eighteenth century; Bandera’s account actually represents the first documented use of “Catapa”—a cognate of “Catawba,” a name that would come to represent most Carolina Piedmont peoples after 1720. Pardo’s company continued north from Canos and visited several towns—Tagaya, Gueça, Aracuchi, and Otari—on the lower reaches of the Catawba River, just north of the Fall Line. Each of these towns was a subject of Cofitachequi in 1540, but whether any remained so in 1567 is unclear. In any event, each had built its own house for Pardo. In fact, all of the towns that the army visited on the Wateree and Catawba rivers built similar houses for the expedition, suggesting that all viewed the entrada as an opportunity to demonstrate their relative autonomy from other towns

Introduction · 13

(Beck et al. 2010). After leaving Otari, Pardo and his men traveled through an uninhabited stretch of the lower and middle Catawba before resting at Quinahaqui and Guaquiri. At Quinahaqui, Pardo again received the leaders of Ysa and Cataba, as well as another whose name, Uchiri (Usheree or Ushery), may identify a community of Yuchian speakers in the Carolina Piedmont. On September 24, Pardo arrived at Joara, where he learned that Moyano “was gone from the fort . . . and that the Indians had him under siege” (Pardo 1990: 313–314). Fearing for Moyano’s outnumbered contingent, he departed at once for Chiaha, leaving Corporal Lucas de Canicares in command of Fort San Juan. Bandera writes that the company marched for three days before they came to a town called Tocae, “a place which is over the top of the ridge [of mountains]” (Bandera 1990a: 266). Chester DePratter and colleagues (1983: 143) suggest that Pardo and his men crossed into the Blue Ridge Mountains at Swannanoa Gap and that Tocae was located near the modern city of Asheville. Pardo stopped at Tocae for just four hours. The following day, October 2, he came to Cauchi, located along the Pigeon River, where he met with several leaders from Cherokee towns to the south and southwest (Hudson 1990: 97). After leaving Cauchi, Pardo marched northwest through the Pigeon River Gap into what is now eastern Tennessee. On October 7, the expedition finally arrived at Olamico, the principal town of Chiaha, where they discovered that Moyano and his contingent were “hard pressed” but safe (Pardo 1990: 314). On October 13, the reunited army departed Olamico. Three days later, at a town called Satapo, Pardo learned of a plot to ambush the expedition, and after conferring with his officers, he decided to turn back. At Olamico, he established Fort San Pedro, where he left twentyseven men. At Cauchi, he founded Fort San Pablo with eleven men (Bandera 1990a: 269–278). Finally, on November 6, he and his reduced expedition arrived at Joara, “where he made a halt and remained twenty days because the people of his company were tired and poorly provided, that they might have a place to rest and to provide themselves” (Bandera 1990a: 277). While the company rested, no fewer than twenty-five oratas and their mandadores, or principal men, came to see Pardo at Joara, a clear sign of the burgeoning status enjoyed by this chiefdom and its leader, Joara Mico. These visiting oratas arrived at Joara in three groups. In the first were leaders from the aforementioned villages of Quinahaqui, Guaquiri, Catape (or Cataba), and Ysa Chiquito (Bandera 1990a: 277), the latter apparently a

14 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

hamlet of Ysa. In the second party were eighteen caciques of towns whose locations are unknown. Some of these chiefs had met with Pardo at Cauchi as he was returning across the Appalachians, and Charles Hudson (1990: 88–89) suggests that these leaders may have traveled to Joara from Cherokee towns on the upper Broad, Saluda, and Savannah rivers. It is also possible that some of these oratas came from Burke phase towns on the upper Catawba and Yadkin rivers. Bandera does specifically write that five of the oratas in this group were “caciques of ” Joara Mico (Bandera 1990a: 278), such that their towns likely formed the core of the Joara polity (Beck and Moore 2002: 201). Finally, just before the army left Joara, two leaders named Chara Orata and Adini Orata—though nominally subjects of Guatari—came to switch their allegiances to Joara. To prevent conflict between the Guatari and Joara polities, Pardo convinced these men to continue giving their obedience in the place they were accustomed. Before departing from Joara, Pardo outfitted Fort San Juan with more materials—and with a broader range of supplies—than any of the other garrisons. He again manned the fort with thirty soldiers but made Alberto Escudero de Villamar its commander and gave him authority over all of the forts between Chiaha and Santa Elena; together, these actions clearly indicate that Pardo conceived of Fort San Juan as the most important of his scattered outposts. Pardo and the rest of his company left on November 24, marching south toward Ysa. Along the way, they made numerous halts to investigate crystal mines that some of the soldiers at Joara, including the newly demoted Moyano and a silversmith named Andrés Suarez, had begun to work during the months they were stationed at Fort San Juan (Bandera 1990a: 279–283). In fact, Bandera goes to great lengths in his long relation to establish which men of the company had a legal claim to these mines should they prove to contain valuable gems. Juan de Ribas, a member of the second expedition who provided testimony before the governor of La Florida in 1602, recalled that the Spaniards at Fort San Juan gambled for “tiny diamonds” they collected near the fort (Hudson 1990: 163). While these stones were only quartz, such tales nourished the enduring myth of Los Diamantes, a fabled mountain of diamonds in the interior (Hudson 1990: 189). After a long pause at Ysa, where Pardo and the company stayed for some two weeks, they finally arrived at the town of Guatari on December 15. As with the Cofitachequi chiefdom when Soto passed through the Carolinas in 1540, the mico at Guatari was a woman. Thus, two of the three micos who met with Soto and Pardo in the Carolinas were women. On his arrival,

Introduction · 15

Pardo told the cacica to summon her subjects to help the Spaniards build a fort. Bandera’s list of formal gifts that Pardo made that day indicates that seven of the subject oratas came at her behest (Bandera 1990a: 284). While this number might seem low, it is quite close to the number of local headmen that Bandera describes as “caciques of ” Joara Mico. Pardo completed Fort Santiago on January 6 and stationed sixteen men there. He left the next day with the rest of his army and marched for five days before reaching Aracuchi, located along the lower Catawba River. They then turned east and made a detour to the Pee Dee River and the aforementioned town of Ylasi, arriving there in five days. Heavy rain delayed the departure, but they left for Canos on January 21, arriving on January 23. Pardo stayed there for eighteen days, during which time he built Fort Santo Tomás and left it garrisoned with thirty men (Bandera 1990a: 292). After leaving Canos and the Carolina Piedmont, he made a strong house at the Coastal Plain town of Orista, then returned to Santa Elena with less than a dozen men on March 2, 1568. Shortly after Pardo’s arrival at Santa Elena, relations between these forts and their Indian hosts took a deadly turn for the worse. By May 1568, news reached Santa Elena that Indians had attacked all of the interior forts and that all had fallen. Whether all of these forts were surprised at the same time, in an action prefiguring the well-known Pueblo Revolt of 1680, is unclear, but clearly none remained by June 1568 (Hudson 1990: 176). Among the several factors that appear to have had a role in the Indians’ decision to destroy the garrisons, two stand out: the Spaniards’ demands for food and their improprieties with Indian women. With respect to the provisioning of the forts, there was apparently a great deal of miscommunication between Spaniards and Indians about the presentation of food. By May 1568, the soldiers at the forts may have had few trade goods left to give in exchange for maize and the other foodstuffs they obtained from their hosts, and it is quite significant that Bandera’s detailed lists of supplies that Pardo left at each fort make no specific mention of trade goods (though some of the chisels listed among the garrison’s supplies may have continued to serve this purpose for a time; see table 8.4). From the Spaniards’ perspective, this was as it should be. The native peoples of the Piedmont were now their subjects, and as such they owed the Spanish Crown—and these men who served as its military stand-ins—continued sustenance as tax or tribute. And yet from the Indians’ perspective, Pardo’s soldiers were no longer fulfilling their end of this exchange relationship, and hence they owed the Spaniards nothing.

16 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

Sexual politics also seems to have played a significant role in the forts’ destruction. Before the army left from Guatari, for example, Bandera (1990a: 285) notes that Pardo “commanded him [the corporal placed in charge of Fort Santiago] in the name of His Majesty . . . that no one should dare bring any woman into the fort at night and that he should not depart from the command under pain of being severely punished.” However, Teresa Martín, an Indian woman taken back to Santa Elena on Pardo’s second expedition who was the wife of Juan Martín de Badajoz—who somehow managed to survive the attack on the forts—testified before Governor Canço in 1600 that the men waited “three or four moons” for Pardo to return to the interior. When he failed to do so, they had begun to commit improprieties with native women, angering their men (Hudson 1990: 176). A Jesuit priest, Juan Rogel, blamed the destruction of Pardo’s garrisons on the Spaniards’ continued mistreatment of their hosts. In a letter to Havana dated July 25, 1568, Rogel notes, “[S]ince I intended to learn the Indians’ dispositions, the Captain [Pardo] took me to a town called Escamacú, five leagues distant from [Santa Elena]” (quoted in Ugarte 1935: 85). Immediately upon their arrival, Pardo had to appease the townspeople for “certain wrongs” inflicted by soldiers stationed nearby at Fort Orista. But that night, he reported, We heard the outcries of some Oristan Indian women who were imploring the garrison at Escamacú to restrain the soldiers who had raided Oristan and captured it. As we happened to be present, the Captain sent a soldier of our company to reprehend them. . . . This is why I suspect that the forts were lost and our soldiers slain because of the ill-treatment given by our troops. If they wrong them while their Captain was so near what would they not do when they were inland two hundred leagues away from him? (quoted in Ugarte 1935: 85) Whatever the combination of causes, all six of Pardo’s garrisons and most of the men he stationed along his path were lost, and with them was lost Spain’s only attempt to colonize the deep interior of La Florida. Indeed, this loss precipitated the shrinking of Spanish claims across the Southeast. Ajacán, a place that English settlers would later name Virginia, was lost in 1571 following a brief and disastrous attempt at missionization. Santa Elena was finally abandoned in 1587, after which Spain’s southeastern territory was largely restricted to St. Augustine and the boundaries of present-day Florida. No other Europeans, however, would travel so far into the southern interior until the last quarter of the seventeenth century. For nearly 450

Introduction · 17

years, the remnants of Joara and Fort San Juan lay buried in the field that we now know as the Berry site. But how have we been able to identify this site as the town of Joara and its compound of burned buildings as Pardo’s Cuenca? Here we need to recount the confluence of scholarship and good timing that brought us all to the Berry site and allowed us to recognize the archaeological signature of the long-lost colony.

Anchoring Spanish Entradas in the American Southeast Spanish entradas under the leadership of Hernando de Soto (1539–1543), Tristan de Luna (1559–1560), and Juan Pardo (1566–1567) all produced several descriptions of Late Mississippian chiefdoms in the American Southeast. Until recently, though, it was difficult to make the most of these accounts because historians and archaeologists were unable to anchor these descriptions to the cultural geography of the precolonial South with any real accuracy. In 1939, John Swanton, who was among the foremost ethnologists of his era, completed what was at the time the most detailed reconstruction of Soto’s long path ever undertaken. But southeastern archaeology was still in its infancy as a discipline when Swanton was tracing Soto’s route, and because his reconstruction made little use of archaeological data, it was not convincingly tied to the sixteenth-century landscape of native towns and polities that these Spanish explorers described. Then, in the early 1980s, Charles Hudson and his colleagues at the University of Georgia and the University of Florida began the project of combining the documentary and archaeological records, knowing that their effort provided the best means of linking the precolonial world to the Native South of the eighteenth century. Hudson’s group used a hitherto littleknown narrative of the second Pardo expedition to reconstruct the Spaniard’s route with great precision (DePratter et al. 1983; Hudson 1990). While they were not the first to realize that Pardo had followed much of the same path that Soto’s army had used to cross the Piedmont in 1540, this detailed account—the long report of Juan de la Bandera—offered an important link reconstructing the Soto path, as well. By this time archaeologists had established better cultural chronologies for most of the Southeast, so that Hudson and his team could finally situate these early Spanish documents in archaeological space. This allowed them to connect specific towns and chiefdoms named in these accounts with archaeological sites and cultural complexes (Hudson 1990, 1997; Hudson et al. 1985). Three lines of evidence proved especially useful for tracing the routes of

18 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

early expeditions such as those of Soto and Pardo (e.g., Beck 2013: 65–66). First, of course, are accounts written by expedition members. Such accounts offer invaluable details about how far an expedition traveled on a particular day; describe physiographic features such as streams, swamps, and mountain passes; and record the names of indigenous towns and note their geographic relationships both to these landscape features and to one another. Yet the distance traveled in a day may be deceiving to exhausted soldiers marching through dangerous, rough, and unfamiliar country. Moreover, our understanding of these distances now depends on knowing which unit of measure the writer used. And while many early route historians used the recorded names of native towns and their known eighteenth-century locations to trace the routes of sixteenth-century Spaniards, we now know that these towns moved frequently, and sometimes across large distances, during the first two or three centuries of the colonial era. Later sources should be considered poor guides to sixteenth-century town and polity locations until they are proven otherwise. We must therefore turn to archaeology for corroborating evidence of Spanish explorations across the Carolina Piedmont. Explorers such as Soto and Pardo carried specific kinds of goods—particularly glass beads strung into necklaces, brass or copper bells, and iron knives and chisels—to use as gifts for Indian leaders who provided support to the Crown. Some of these goods are temporally diagnostic, such that their recovery on a site can help to date it to the period of these early expeditions (e.g., Smith 1987, 2000). Yet native leaders traveled long distances to obtain items of European manufacture from the Soto and Pardo expeditions, such that the identification of these sorts of trade goods at a site need not mean that it was directly along the Spaniards’ path. More convincing are items such as European ceramics, nails, lead shot and gun parts, crossbow bolts, and fragments of armor that rarely entered exchange relationships between Spaniards and native groups but instead suggest the actual presence of Europeans at a site. Finally, the distribution of Native American towns and villages offers a key line of evidence in reconstructing the paths of Spanish expeditions. The Soto and Pardo accounts alike refer to large, concentrated populations in many parts of their respective routes, and indeed, each expedition was dependent on such populations for their basic subsistence needs. The dependence of the Spanish expeditions on native food supplies, particularly on maize, means that where these Spaniards marched was entirely dependent on where native peoples lived—specifically, where they resided in sufficient numbers to have large stores of sur-

Introduction · 19

plus maize that the expeditions could appropriate, whether by persuasion or by force. Although the research done by Hudson’s group revived debate among scholars throughout the region traversed by these expeditions, in few places did the revisions offered by Hudson and his colleagues require as complete a redrawing of earlier efforts as in the southern Appalachians and adjacent parts of the Piedmont. The distinctions between what have come to be referred to as the Swanton Route and the Hudson Route through this place hinged on whether Soto and Pardo entered the Appalachians via the upper Savannah River (Swanton) or via the upper Catawba (Hudson). Just as Hudson was beginning to spearhead the effort to reconstruct a more accurate route, archaeological research in the Savannah Valley (e.g., Anderson 1994) was suggesting an absence of sixteenth-century towns and villages such as those reported in accounts of the Spanish expeditions. At the same time, the Catawba Valley was virtually a terra incognita, with archaeologists unable even to say whether a Native American population lived in the valley during the mid-sixteenth century; identifying sites such as the native towns of Ysa, Guaquiri, and Joara seemed a remote possibility, if attainable at all. It was with the goal of testing Hudson’s Catawba hypothesis and developing a knowledge of the valley’s late precolonial population that David Moore began excavations at the McDowell and Berry sites as part of his dissertation field research in 1986 (Moore 2002). Hudson understood that any confirmation of town locations by archaeologists would offer a critical linchpin for his redrawn map of the sixteenthcentury South. He and his colleagues had suggested that the McDowell site (31MC41), located near Marion, North Carolina, was the town of Joara (De Soto’s Xualla). Moore therefore intended to focus his project on the McDowell site, but—because of unforeseen circumstances—he spent more time at the Berry site (31BK22, figure 1.4), about 20 miles (40 kilometers) east of Marion near present Morganton. This seemingly unfortunate turn of events would prove in time to be a boon. In the end, Moore excavated extensively at each site, and although he was unable to confirm that either was the location of Joara, he did show that both were large towns with earthen mounds and that the upper Catawba River valley had a substantial late precolonial population that was consistent both with descriptions in the Spanish accounts and with Hudson’s Catawba hypothesis (e.g., Levy et al. 1990; Moore 2002). In 1994, two independently unfolding events would confirm the general path of Hudson’s revised route and suggest that Berry, not McDowell,

20 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

Figure 1.4. Berry site excavations in 1986, looking north. The Spanish compound is in the upper right portion of the image (cf. figure 1.5).

was the location of Joara and Fort San Juan. First, as noted, John Worth’s (see chapter 3, this volume) translation of Domingo de León’s account offered a detailed description of the route taken by Pardo and his company, a description that only matches a course through the Catawba-Wateree Valley. Second, Moore and Robin Beck reported the discovery of sixteenthcentury Spanish pottery and hardware from the Berry site (see chapter 9, this volume). During surface reconnaissance of Berry in early 1994, Beck had discovered several sherds of olive jar and a wrought iron nail in the general area of Moore’s 1986 excavations. As a result, Beck and Moore reexamined the 1986 collections and identified additional sherds of olive jar, as well as molten lead sprue. Subsequent visits to the site yielded more olive jar sherds and a single fragment of Caparra Blue majolica, found on New World sites that date to the period from 1492 to 1600 (Deagan 1987: 92). This assemblage of Spanish artifacts from the Berry site closely matches the list of supplies that Bandera reported were left for the soldiers at Fort San Juan. The only other site in the interior Southeast with a similar assemblage of early Spanish artifacts is the Governor Martin site in present-day downtown Tallahassee, Florida, the location of Soto’s winter camp in 1539

Introduction · 21

at Apalachee (e.g., Ewen and Hann 1998). At the Southeastern Archaeological Conference, Moore and Beck (1994) identified Berry as the site of Joara and Fort San Juan, a location that fell within about 20 miles of the site that Hudson and his colleagues had proposed in 1983. The identification of Berry as Joara spurred Beck (1997b) to revisit the routes of Soto and Pardo across the Appalachian Summit. Specifically, if Berry was Joara, then the Spaniards likely traveled into the mountains by different routes, with Soto taking a northern course from Berry into the Nolichucky drainage and Pardo taking a westerly course into the French Broad drainage through the Swannanoa Gap. In 1996, as a part of his master’s project at the University of Alabama, Beck (1997a) conducted a systematic archaeological survey of Upper Creek/Warrior Fork, the tributary of the upper Catawba on which the Berry site is located. This survey identified five large sites or site pairs that seem to have been contemporaneous with Berry (see chapter 2, this volume). In 1997, Beck and Thomas Hargrove conducted a gradiometer survey of Berry that revealed evidence of at least four large subsurface anomalies believed to represent the remains of burned buildings; auger tests confirmed the presence of burned materials in discrete areas corresponding to the anomalies. And in 2001, Moore, Beck, and Christopher Rodning—then a doctoral student studying Cherokee archaeology in southwestern North Carolina—established a joint archaeological project and began new excavations at Berry to expose the site areas thought to contain the remains of Fort San Juan (Beck et al. 2006; Moore et al. 2004). This work would soon lead us into the household archaeology of a colonial encounter.

Household Practice at Cuenca and Fort San Juan Stephen Silliman (2005) has proposed disentangling the concepts of culture contact and colonialism in historical archaeology. He suggests that while the term contact may be suitable for initial or first encounters, the term colonialism better suits long-term contexts of interaction between Europeans and the native peoples of North America. Initial European contact, as such, occurred in North Carolina during May 1540, when the Hernando de Soto expedition marched across the Carolina Piedmont. We conceive the founding, occupation, and ultimate destruction of Fort San Juan—which took place from 1566 to 1568—as constituting a colonial encounter, during which the native peoples of the upper Catawba Valley engaged in daily, sustained interaction and exchange with representatives of a colonial power.

22 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

This encounter informs the broader processes of colonialism, especially with respect to similar frontier and borderland settings. Archaeologists have moved away from unidirectional studies of acculturation, focused on how indigenous groups passively accepted elements of European culture, toward more-balanced and agent-centered approaches that conceptualize the myriad ways by which both Europeans and native peoples actively negotiated identity in colonial settings (Cusick 1998; Deagan 1983, 2003, 2004; Ewen 1991; Lightfoot 1995; Lightfoot et al. 1998; Scott 1990; Silliman 2005; Voss 2008). Kathleen Deagan (2003: 8) notes, for example, with respect to the Spanish Empire, that “the goals of establishing civilized Christian life as dictated by the Church and the Crown were apparently adjusted most strikingly (and perhaps even largely ignored) in rural and frontier areas of the empire. . . . In fact, there is some indication that the Spaniards who lived in these communities made far greater adjustments to the American mode of life than vice versa.” While Gil Stein (1998, 1999) focuses his distance parity model of interregional interaction on commodity exchange between “core” regions and their peripheries, we believe that this model has broader implications for colonial exchange along frontiers, including those that Deagan discusses in the passage quoted above. Distance parity, a recent alternative to world system approaches that assume cultural hegemony in core-periphery relations, posits that the ability of a “core” region to project its power and cultural influence into peripheries diminishes with distance, leading to moresymmetrical exchange relations (Stein 1998: 229). As Deagan’s passage suggests, conventionally expected patterns of exchange—both of material commodities and of cultural practices—may actually be less symmetrical than reversed in frontier settlements such as Fort San Juan. That is, the more distant or isolated a colony is from its “core” (or from fellow colonies), the more dependent it will be on exchange and interaction with local indigenous groups, assuming relative technological and organizational parity between the settlers and their native hosts. Along frontiers, households are particularly well-suited to provide archaeological data on colonial encounters (e.g., Deagan 1983, 1985a, 2004; Ewen 1991; Hoffman 1990; Lightfoot et al. 1998; McEwan 1995; McEwan, ed. 1993). Kent Lightfoot and colleagues note, “From an archaeological perspective, we believe the study of change and persistence in multi-ethnic contexts pertaining to the construction of social identities may be best addressed by considerations of daily practices involving domestic life and the organization of space” (1998: 202). In light of this, our archaeologi-

Introduction · 23

cal research at the Berry site has thus far emphasized the excavation and analysis of household contexts at Fort San Juan, specifically within the architectural complex that we refer to throughout this volume as the Spanish compound (figure 1.5). Today, this compound consists of five burned buildings and dozens of features (plate 1), but the men of Fort San Juan would have passed much of their time there: it was where they slept, took their meals, developed personal relationships with one another and with the men and women of Joara, and followed the habitual, routine practices of daily life. We have designed our research—both its field and its laboratory analyses—with respect to three domains of household practice: (1) house construction, (2) household organization, and (3) food preparation and consumption (e.g., Lightfoot et al. 1998: 209–215). Our particular questions inform these three broad domains and together address our key problem: how did these Spaniards and their native hosts construct and maintain social relations through household practice? Domain 1: House Construction 1. How was the labor of house construction organized? 2. What kinds of material culture were used to build and maintain houses? 3. What kinds of techniques and practices were used to build and maintain houses? 4. In what season(s) of the year were houses built? Because of their unusual state of preservation, the burned buildings at Berry offer unparalleled opportunities to investigate materials and techniques of house construction on a Spanish colonial frontier during the mid-sixteenth century. Detailed analysis of timbers and architectural furniture permit us to determine the season of year that these structures were built, the types of wood used for different structural elements, the kinds of tools used to prepare and shape materials (such as metal versus stone tools), and the materials used to join different elements (for example, nails). The Pardo accounts suggest that Spaniards built the fort, while native Joarans built at least one of the associated houses. However, our research suggests that both Spanish and native construction practices were incorporated in the buildings we have excavated so far. Our evidence on house construction thus sheds light on whether Spaniards and Joarans worked together on the same buildings or their work parties were culturally segregated.

Figure 1.5. Berry site excavations, 1986–2013, indicating the Spanish compound, mound areas, and the Fort San Juan moat. Note that the north arrow on this and all subsequent maps indicates grid north; the 1986 excavations are oriented toward magnetic north.

Introduction · 25

Domain 2: Household Organization 5. What kinds of activities took place in and around houses? 6. How were household activity areas spatially organized? 7. What kinds of material culture were incorporated into household contexts? 8. How did gender mediate the organization of household activity areas? Spanish soldiers lived at Fort San Juan for eighteen months, until its destruction, and each of the burned buildings there likely housed several soldiers throughout this period. Excavation and analysis of undisturbed house floors and outdoor features provide data on how early colonial Spaniards incorporated native lifeways into their daily routines and practices. These house floors and features also provide evidence about gender relations at Fort San Juan, particularly the extent to which native women participated in household activities inside the compound. Throughout the Spanish Americas, native women were, as Deagan (2003: 8) observes, “a potent force” in the integration of Spanish and native communities; indeed, Spaniards made some of their most significant adjustments to American life in household settings, guided by the practices of indigenous women (Deagan 1983, 1996, 2003; Ewen 1991; McEwan 1991; Reitz and McEwan 1995). The accounts of the Pardo expeditions hint at relations between native women and Spanish soldiers stationed at the interior forts, and some of Pardo’s soldiers later married Indian women taken to Santa Elena at the end of the second expedition. One of these, Teresa Martín, gave official testimony in 1600 that when Pardo did not return to the interior forts within “three or four moons,” some of his men began to commit indiscretions with local women, angering their men (Hudson 1990: 176). Martín was identified as a native of “Juacan” (Joara) and was married to one of the thirty soldiers left at Fort San Juan after Pardo’s first expedition, Juan Martín de Badajoz; Hudson (1990: 176) suggests that they may have married at Joara and so perhaps lived together in Fort San Juan. Domain 3: Food Preparation and Consumption 9. Where did food preparation activities take place? 10. What kinds of food were prepared and consumed in household contexts? 11. What kinds of material culture were used in food preparation and consumption?

26 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

12. How were the wastes associated with food preparation and consumption discarded? Many of our specialist analyses focus on food, food preparation, and food serving remains from floor and pit feature contexts. Food is one of the key means through which people maintain their social identities. These specialist analyses, together with detailed contextual data, permit us to examine how Pardo’s soldiers provisioned themselves while occupying the fort, as well as the degree to which they received provisioning from the town of Joara. They also shed light on the kinds of food consumed in the fort and the kinds of tools used in food preparation. Furthermore, different kinds of food were associated with different preparation and consumption practices. Some foods were prepared by Joaran women within the Spanish compound, others by Joaran men outside the compound, and some by the soldiers themselves.

*

*

*

Too often in archaeology, the nature of our evidence demands that we pre­ sent the colonial encounter as a homogenous clash of cultures. Rarely can we disentangle the individual decisions, practices, and actions that constitute the encounter from those palimpsests of decisions, practices, and actions that constitute the archaeological record. It is here that our volume on Joara and Fort San Juan offers significant insights about how such encounters actually unfold, in decisions about which woods and tools to use in building a house; about which foods to prepare and consume (or which to withhold); about which things to bring and which to leave as you embark on the march to a foreign land. Yet Berry provides more than a snapshot of the eighteen months that Pardo’s men occupied Joara. Instead, the temporal resolution that we have identified in the archaeology of the Spanish compound offers something akin to a moving picture, or as close to such as we can likely hope to see in an archaeological context. This book thus draws on the tools of scientific analysis and historical narrative. We present our data to tell a story, one that we hope will be of interest to all scholars of colonialism and empire—archaeologists and historians alike.

II Who They Were Situating the Colonial Encounter

It began with a rumor, as it had those many moons before when the mico himself was still but a boy. Then as now, word had come to Joara from the south, from Cofitachequi: strange men were upon the land, taking and devouring whatever they wanted, hundreds of men, with war dogs more vicious than wolves and weapons that shone and flashed with the light of the sun. Yet what had terrified and enthralled the mico most, all those moons before, was the tale that some of these strangers rode upon the backs of great beasts—beasts that stood taller on all fours than any man on two legs could reach. He had not truly believed it; he was still a boy but no longer a child. Then the strangers had arrived one evening as the people of his town were preparing to eat—an advance party, it turned out, for the hundreds of others who followed—and each man rode upon just such a beast as rumor described, terrible, beautiful, and real. That night, his mother had led him away to another town nearby, and they had not returned until the strangers and their beasts had departed over the mountains beyond Joara. Now, after such a long time, there was a new rumor: strangers were upon the land once more, though fewer in number and all on foot. His scouts had confirmed it, so he would begin to prepare his people. He was not afraid of the strangers—only disappointed in their manner of march. He would have liked to see those great and terrible beasts again.

*

*

*

In December 1566, when their respective paths through history crossed for a second time, few of their contemporaries would have doubted that Spain and Joara were in ascendance. Under Philip II, Hapsburg Spain was at the peak of

28 · Part II. Who They Were: Situating the Colonial Encounter

its power and was the foremost nation in Europe. It was a time referred to as El Siglo de Oro, the Golden Century. Sailing under the unified banners of Aragon and Castile, Christopher Columbus had first come ashore in the Bahamas just seventy-five years before. Three decades later, in 1521, Hernán Cortés had toppled the powerful Aztec Empire and made its capital of Tenochtitlan—modern Mexico City—his own. Francisco Pizarro, with but 106 foot soldiers and 62 horsemen, had defeated the Inca ruler Atahualpa in 1532 and claimed the western slope of South America, from the Pacific Ocean to the Andes, for the Spanish Crown. By the time of Philip’s reign, his imperial dominion reached to every continent then known. Beyond Middle America, Central America, and the American Southwest, the Caribbean, La Florida, and Peru, there were Spanish possessions in western and eastern Africa, on the Indian subcontinent and modern Sri Lanka, and as far east as the Philippines—so named after Philip himself. As familiar as we are with the old adage that “the sun never sets on the British Empire,” history shows that it was first used in reference to the empire of Spain under Philip and his father, Charles I. Joara, too, despite the enormous differences in scale, was a polity on the rise. When Hernando de Soto and his expedition passed through the Carolina Piedmont in 1540, there is little to suggest that the headman of Joara (Xualla) had authority over any community but his own. Instead, Cofitachequi was the preeminent Piedmont town and the only place in the region described as holding political and economic power over neighboring towns and villages. Much, then, had changed between the time of the Soto and Pardo entradas. As Cofitachequi’s stature had waned, due in part perhaps to the treatment of its cacica during Soto’s passage, Joara and Guatari—each identified as the seat of a mico a quarter century later—had begun to fill the void left by their neighbor to the south. Joara Mico held authority over several other towns along the upper Catawba and its tributaries and was able to embroil the soldiers under Moyano in his conflicts with rival towns and polities across the Appalachians. Peoples in the Piedmont had taken notice of the ambitious mico—some towns, such as Chara and Adini, sought to give him their allegiance; others, such as Ysa and Quinahaqui, feared that the Spaniards would place them under his sway (Beck 2013: 154–155). Thus, even as Joara offered Spain a place to anchor its colonial ambitions deep in the interior of La Florida, so too did Spain provide Joara with the political capital to expand its own domain. In the chapters that follow, we consider the problem of who they were by establishing the cultural and historical contexts of the colonial encounter. First, the volume editors (Beck, Moore, and Rodning) synthesize Late Mississippian archaeology in the upper Catawba

Part II. Who They Were: Situating the Colonial Encounter · 29

Valley, focusing in particular on the Burke phase (circa A.D. 1400–1600). The Burke phase was a Late Mississippian cultural frontier representing the northeastern limits of the Mississippian South. Burke phase peoples were maize farmers who erected platform mounds and participated in far-flung exchange networks with other groups across the southern Appalachians. By the midsixteenth century, these peoples were also organized in regional polities or chiefdoms, one of which was centered at the Berry site. Here we discuss what we currently understand of Burke phase communities, settlement patterns, mortuary patterns, and mound construction, using these lines of data to prepare the stage for Pardo’s arrival at Joara—the Berry site—in December 1566. In chapter 3, “Recollections of the Juan Pardo Expeditions,” John Worth provides us with a newly discovered and translated account of the Pardo expeditions, written by interpreter Domingo Gonzáles de León in 1584. León had arrived in La Florida with the adelantado Pedro Menéndez de Avilés in 1565, helping to found the colonies of St. Augustine and Santa Elena. He was a member of Pardo’s second expedition, so he took part in the events that he describes in his account. Given that there are only four other existing accounts of Pardo’s journey, Worth’s discovery of the León document provides an invaluable primary source that offers information on towns and polities in the interior, including Joara, not contained in other sources. Particularly important—and alone among the primary accounts of the expeditions—is that León’s recollections situate the social geography of towns and villages along Pardo’s path in the physical geography of the region, leaving no room for doubt as to the path of Pardo and his men into the deep interior. This chapter marks the first publication of León’s account.

This page intentionally left blank

2 Joara in Time and Space Robin A. Beck, David G. Moore, and Christopher B. Rodning

Although the Pardo accounts provide us an extraordinary snapshot of political relations in the upper Catawba Valley during the mid-sixteenth century, they tell us practically nothing of the cultural and historical contexts in which these relations unfolded, contexts that probably stretched back more than a century before the Spaniards’ arrival and forward as much as a century after the fall of Pardo’s forts. Even when these accounts are brought together with those of the Soto expedition, our sum total of written sources can illuminate only a bit more than a quarter century of Joara’s social history along the upper Catawba River. To understand Joara in a broader context we must turn to archaeology. We acknowledge in doing so that the kinds of things we can know about the people of Joara are dramatically different from the kinds of things we can know about the Spaniards who settled at Santa Elena or the thirty men who garrisoned Fort San Juan. The people of Joara left us no written accounts that tell their own history—who they were on their own terms—much less that describe their range of thoughts, actions, and motivations throughout this colonial encounter. We will never know the specific year of Joara’s founding or that of its final abandonment. We do not know the name of a single native inhabitant of Joara, for even “Joara Mico” was a title rather than a personal name. Unfortunately, these and many other details are lost to us today. Yet if the intimate details of names, dates, and motivations are all but out of reach for Joara’s deep past, then what we can do through archaeology is to unmask the broader patterns and processes of history, the cumulative outcomes of the personal scale. By mapping the distributions of specific styles of decorated pottery—and how these regional distributions changed over time—we can begin to learn how the native peoples of the upper Catawba Valley settled their landscape and transformed it in socially constructed

32 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

and meaningful ways. By studying how they placed their earthen mounds and buried their dead, we might identify evidence of social inequality and hierarchy, both within and between communities. When we combine these kinds of evidence with the written accounts from the Pardo expeditions, we can gain a fuller understanding of the historical context in which the people of Joara received these foreigners as their allies and guests, if only for a matter of months.

Archaeology of the Upper Catawba and Yadkin Valleys Our study area, the place that Pardo and his men knew as the province of Joara, is located on the upper reaches of the Catawba and Yadkin rivers in western North Carolina (figure 2.1), on the inner boundary of the Carolina Piedmont in the foothills of the Appalachian Mountains. The earliest archaeological investigations in this area, although among the best-known archaeological undertakings documented in western North Carolina, remain poorly understood. These were the famous mound investigations conducted by John P. Rogan—under the auspices of the Bureau of Ethnology—and reported by Cyrus Thomas (1887, 1891, 1894). Thomas (1891: 152) reported seventeen mounds in Caldwell County, the largest number for any county in North Carolina; however, only those excavated sites described in the 1887 and 1894 publications have actually been confirmed. Thomas and W. H. Holmes (1903) also discuss at least eight mounds reported by James Mooney and J. M. Spainhour in McDowell, Burke, and Catawba counties along the upper Catawba. Four of these mound sites have been assigned tentative site number designations in the North Carolina State Site Record. Only one, located at the Berry site (31BK22, also reported as 31BK2), has been verified as a human-made, earthen mound (Moore 2002: 214–222). The remnant of one additional mound, possibly one mentioned by Mooney, has been identified at the McDowell site (31MC41) in McDowell County (Moore 2002: 197, 206). Despite the difficulties in relocating most of these reported sites, the distribution of mounds in the upper Catawba and Yadkin valleys is clearly unlike that of any other area in the North Carolina Piedmont. Several of the mound sites were excavated during the nineteenth century. Thomas (1887, 1894) reported on the T. F. Nelson Mound (31CW1) and T. F. Nelson Triangle sites, the R. T. Lenoir Burial Pit, the Davenport Jones Mound, and the Broyhill-Dillard Mound (31CW8). These sites were all located in the uppermost Yadkin Valley (figure 2.2), but little is known

Joara in Time and Space · 33

Figure 2.1. The Berry site in its topographic setting. The Appalachian Mountains are to the northwest of Berry (in the upper left corner of the map).

of them besides Thomas’s brief description of mound excavations and burials. His published reports describe complex mortuary facilities with multiple burials, many of which were accompanied by large quantities of artifacts. Significantly, iron objects obtained from trade with Europeans accompanied at least two of these burials (we return to a discussion of these Caldwell County mounds later in this chapter). At least one mound was also excavated at the Michaux Farm site (31BK17) in the early 1870s (Spainhour 1873, in Moore 2002: 52–56). Although the written descriptions are minimal, they do confirm the presence of at least three human burials in the mound. Unfortunately, these early archaeological explorations did not lead to sustained research, and most later projects were relatively limited in scope. Robert Keeler (1971) provided the first site survey data for the upper Catawba and also defined the Burke ceramic series, describing the soapstonetempered wares that are characteristic of Mississippian sites in the area. Kenneth Robinson (1996) and Robin Beck (1997a) conducted more-recent significant surveys in McDowell County and along Upper Creek in Burke

34 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

Figure 2.2. Major Burke phase sites mentioned in the text (after Moore 2002: 101).

County, respectively. Larger-scale excavations were conducted at the TylerLoughridge (31MC139) site (Robinson 1996), the McDowell (31MC41) site (Moore 1999, 2002; Ward 1977), and the Berry site (Moore 1999, 2002). Likewise, few systematic archaeological investigations were undertaken in the upper Yadkin Valley in the years following the Bureau of Ethnology’s early mound explorations. However, Richard Polhemus (see Moore 2002: 120–123) excavated at 31CW8 in 1964, while teams from Wake Forest University (Idol 1995, 1996; Rogers 1990; Rogers and Wilson 1993) and Appalachian State University (Kimball et al. 1994) recently directed site surveys and conducted excavations at several sites. When Hudson and colleagues proposed the new Soto and Pardo routes, archaeologists could not even say whether a large sixteenth-century population, such as that described for Joara, existed in the Catawba and Yadkin valleys. Today, not only are we certain that such a population was present, but we also know that this

Joara in Time and Space · 35

area featured a large population for at least a century prior to the European invasion. This large population is represented by an abundance of sites exhibiting Burke (soapstone tempered) and Cowans Ford (sand tempered) ceramics. Moore (2002: 167) has described both the Burke and the Cowans Ford ceramic series as regional variants of the Lamar ceramic tradition (Hally 1994b; Williams and Shapiro 1990). The Lamar tradition—identified by distinctive, carved paddle–stamped designs and particular rim forms on jars (figure 2.3) and by incised designs on carinated or shouldered bowls (figure 2.4)—is a Late Mississippian cultural expression that developed from the antecedent Savannah tradition about A.D. 1350 and would eventually encompass most of the southern Appalachian and South Atlantic areas (figure

Figure 2.3. Lamar-style complicated-stamped jar rims from the Berry site (after Beck 2013: 39).

36 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

Figure 2.4. Lamar-style incised cazuela bowl from the Berry site (after Beck 2013: 40).

2.5). The peoples who made Lamar pottery shared a range of lifeways beyond their techniques for decorating ceramic vessels. Most Lamar peoples depended on agricultural crops—especially maize, beans, and squash—and supplemented their farming by hunting, fishing, and gathering. Although some of these peoples built nucleated towns with defensive palisades, open plazas, and earthen mounds, others occupied small hamlets and farmsteads dispersed along major and minor waterways. For the purposes of this chapter we are particularly concerned with ceramics of the upper Catawba and Yadkin valleys, where Burke pottery predominates. W. H. Holmes (1903: 143–144) first reported these ceramics (subsequently named “Burke series”) in his Aboriginal Pottery of the Eastern United States. Holmes referred to several vessels excavated by the Bureau of Ethnology from mounds in the upper Yadkin Valley and suggested that different vessels related to the same occupation reflected distinct ceramic traits from the north, west, and south. Significantly, he also illustrated several examples of Burke ceramic vessels in plate 129 (Holmes 1903: 145). Five of those vessels (a, b, c, e, and f), are fine examples of Burke Lamar pottery (Moore 2002: 315–321). The other vessel (d) is cob impressed and sand tempered and is likely related to Cowans Ford wares; cob-impressing is a minority surface treatment in the Burke series, though it also occurs in Late Woodland assemblages in the central North Carolina Piedmont.

Figure 2.5. Native towns and phases in the Carolina Piedmont, A.D. 1450–1600 (after Beck 2013: 47).

38 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

While pottery of the Burke ceramic series predominates on sites in the upper Catawba and the extreme upper Yadkin Valley, it is important to recognize that other ceramics are found in the project area; these other ceramics include Cowans Ford (Moore 2002: 132–151, 265–267), Pisgah (Dickens 1976: 171–201), McDowell (Moore 2002: 72–73), and Dan River (Coe and Lewis 1952), in decreasing order of frequency. However, all late precolonial period (A.D. 1400–1600) Catawba Valley phases are characterized by the overwhelming presence of Burke and/or Cowans Ford ceramics, with the latter predominating along the middle and lower Catawba Valley in North Carolina (Moore 2002; Riggs 2010). Aside from the Pee Dee pottery (Boudreaux 2007b; Oliver 1992; Reid 1967) recovered in the southeastern Piedmont, Lamar ceramics are uncommon elsewhere in the North Carolina Piedmont, where the Siouan pottery tradition is represented. Yet Lamar attributes (burnished vessels, curvilinear complicated stamping, cazuela bowls, and added fillet strips on jar rims) have been reported on some Caraway (Coe 1964: 34, 1995: 160–166; Ward and Davis 1999: 137) and Oldtown (Wilson 1983: 425–454) ceramics in the central Piedmont. It is now clear that the Lamar tradition, in the form of Burke and Cowans Ford ceramics, extended north throughout the Catawba Valley during the late prehistoric period. The Catawba Valley thus constitutes a continuous Mississippian frontier on the western edge of the Piedmont Siouans. Given the distribution of Burke and Cowans Ford ceramics, Moore (2002: 8) coined the term “Catawba Valley Mississippians” for Mississippian peoples of the Catawba Valley and its tributaries during the centuries prior to A.D. 1650. Moore based this term on Leland Ferguson’s (1971) South Appalachian Mississippian, which did include Lamar culture. While we recognize that there can be no absolute correlation between cultural groups and ceramic types, the presence of Lamar ceramics has usually been recognized across the Southeast as the primary diagnostic of Lamar culture (Hally 1994b: 144). Given the ubiquity of both Burke and Cowans Ford ceramics in the Catawba River valley, we suggest that this area was the northernmost extent of Lamar culture from the fourteenth through the seventeenth centuries, and we view these ceramics as the primary unifying element of Catawba Valley Mississippian. This period of Late Mississippian occupation along the upper Catawba River is known as the Burke phase, but before turning to a discussion of the phase itself, we need to address the problem of chronologies in archaeology. In this and subsequent chapters, we refer to the “Burke phase” and the “Burke ceramic series” when writing about the archaeology of Joara and

Joara in Time and Space · 39

contemporary sites on the upper Catawba River and its tributaries. While archaeologists usually take the relationship between cultural phase and ceramic style for granted, nonspecialists may find it perplexing. So what are cultural phases and what is their relationship to ceramics? And what have either to do with how archaeologists think about and measure time? Archaeologists (unlike historians who work with text) usually lack sufficient information to pin a specific date to the contexts and events our excavations reveal. Yet such a limitation does not mean that those contexts and events revealed through careful excavation are any less historical than those contexts and events revealed through analysis of written text. What this does mean is that archaeologists working without access to texts need to think about time and chronology differently from most historians and to devise methodological tools and concepts that recognize these differences. Both the concept of cultural phase and the common practice of using ceramic styles as phase markers are products of this need. Archaeologists have long recognized that changes in artifact styles can allow us to divide time into smaller, more manageable units. Radiocarbon dating, which was developed in the mid-twentieth century, is still quite costly and time consuming, so that obtaining a radiocarbon date from every context on every site is neither feasible nor practical. Even if we could, this technique still offers a less precise means of establishing chronology than most historians would demand. Thus, in most circumstances archaeologists can date a specific context only to a block or unit time (for example, 700–450 B.C. or A.D. 1400–1600) rather than to any particular year. And because of the expense of radiocarbon dating, we need another substance that, like wood charcoal and carbonized seeds, is both time sensitive and ubiquitous enough to be recovered from a wide range of contexts. This is why ceramics, even broken potsherds, are important for archaeologists. In many places, ceramic styles changed frequently enough to make them ideal chronological markers, especially when they are matched with a suite of radiocarbon dates. Moreover, different styles often have discrete geographic as well as temporal distributions. When a particular ceramic style was produced for a brief period of time over a limited area, we may use it to mark a cultural phase. The Burke phase is the best locally documented example of Lamar Mississippian society. Moore (2002: 179–181) defined the Burke phase to denote a regionally and temporally distinctive distribution of ceramics and earthen mounds. The phase is named after the Burke ceramic series; its core area is located on Upper Creek/Warrior Fork and Johns River in Burke

40 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

County and along the extreme upper Yadkin Valley in Caldwell County near the Nelson Mound and Triangle sites (see figure 2.2). Based on survey evidence (e.g., Beck 1997a; Keeler 1971; Robinson 1996; Williams 2011), we estimate that no fewer than fifty Burke phase sites, and likely many more, are situated in this area. The distance from the Berry site on Upper Creek and the Michaux Farm site on Johns River to the Nelson Mound and Triangle along the upper Yadkin is less than 25 kilometers, but until recently, there was no basis for archaeologists to associate sites in the two areas. However, a reanalysis of much of the excavated materials from the Yadkin River sites found that nearly all of the sherds and whole vessels were Burke ceramics (Moore 2002: 100–124). The assemblages from nearly all of the sites along Upper Creek/Warrior Fork and the Johns River likewise consist overwhelmingly of Burke ceramic sherds (Beck 1997a; Moore 2002: 74–89). The distribution of sites with high frequencies (typically more than 90 percent) of Burke pottery corresponds almost exactly with the known distribution of earthen mounds and mortuary facilities described earlier. Earthen mounds are not generally found in the North Carolina Piedmont, the only exceptions beyond our study area being Town Creek and a purported mound at 31DV1 (Ferguson 1971: 229), situated along the middle Yadkin River. Also, mounds associated with the Pisgah and Qualla phases in the Appalachian summit are at least 125 kilometers southwest of the project area (Dickens 1976; Keel 1976). Evidence on Burke phase mortuary practices is restricted to the Berry site, where two burials were excavated (Moore 2002), and mortuary facilities reported by Thomas along the upper Yadkin River. These mortuary data are discussed below in the context of Burke phase chiefdoms. The temporal placement of the Burke phase is based largely on radiocarbon dates and the surface treatments of Burke and Cowans Ford ceramics. Moore (2002: 168–169) has reported that both of these series are most similar to the Tugalo phase ceramics of the upper Savannah Valley; the Tugalo phase is dated to A.D. 1450–1600 (Hally 1990, 1994b), and radiocarbon dates support a similar chronological range for the Burke and Cowans Ford ceramics. Sixteen radiocarbon dates (Beck and Moore 2002: 195, table 1) are associated with Burke or Cowans Ford ceramics at six sites on the Catawba and Yadkin, and three more dates are associated with Burke-related pottery in western North Carolina and eastern Tennessee (Moore 2002: 174–177). Most of the Catawba Valley dates range at one sigma from the fourteenth to the fifteenth century A.D. Radiocarbon dates from the Broyhill-Dillard Mound, located in the Yadkin Valley, and from 40JN89 and the Ward site

Joara in Time and Space · 41

in the Watauga Valley, range at one sigma into the seventeenth century. Radiometric data alone thus suggest that Burke and Cowans Ford ceramics were used for more than three centuries, from the fourteenth century to perhaps as late as the seventeenth century. Spanish artifacts associated with Burke pottery from the Berry site (see chapter 10, this volume) and the Nelson Mound and Triangle site also support a mid- to late sixteenth-century date for Burke ceramics. Radiocarbon dates support a similar temporal range for the Cowans Ford ceramics of the middle and lower Catawba, though these may extend into the seventeenth century at some sites. We are not yet able to establish phases of shorter duration for the region; the limits of the Burke phase are designed to incorporate the period of this region’s most intensive Mississippian occupation. Undoubtedly, the regional chronology will be refined by future research. Likewise, it is difficult to describe the antecedents to the Burke phase at this time. Several Late Woodland period Connestee sites have been recorded in the extreme upper valley (Robinson 1996), and the Woodland period Lewis site (31MC157) in McDowell County is the only known site in this area that features Napierlike ceramics (Moore 2002: 283–286). In the core Burke phase region, small numbers of Woodland sherds have been identified at Berry and most of the other sites on Upper Creek/Warrior Fork and along the upper Yadkin River (Beck 1997b; Keeler 1971: 36–37; Moore 2002: 280–287). In general, however, Woodland period occupations of the region appear to have been ephemeral, and we are unable to identify specific Woodland components. Along with the dearth of Woodland components, there are no clear Early Mississippian or early Lamar sites in the area. One possible exception is the Pitts site (31BK209), located about 2 kilometers upstream from the Berry site (Moore 2002: 63). The Pitts site ceramic assemblage consists of 65 percent Burke pottery and 35 percent Pisgah pottery (Moore 2002: 80). Pisgah sherds date to approximately A.D. 1000–1450 in the Appalachian Summit (Dickens 1976: 195–198) and are rarely found at other Burke phase sites. The high frequency of Pisgah pottery at Pitts may reflect a thirteenthto fourteenth-century occupation (Moore 2002: 178), preceding the later, intense Burke phase occupation at Berry and throughout the core Burke area. David Hally (1994b: 147) has pointed out that Lamar ceramics change slowly over time and that these changes generally “occur in the same relative order and are roughly contemporaneous throughout the Lamar area.” We have been unable to identify the presence of early Lamar components in the upper Catawba and Yadkin valleys. In the Burke phase area, we have

42 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

thus found little evidence of significant Late Woodland or Early Mississippian occupations. Instead, large numbers of middle Lamar Burke phase sites seem to appear abruptly in the region. We feel, in sum, that the distribution of Burke ceramics represents a movement of Lamar-related peoples into the upper Catawba Valley in the fourteenth and fifteenth centuries A.D. It is intriguing in this respect that Burke ceramics are less similar to those of the Wateree Valley, downstream from the Catawba in the same basin, than to those from the upper Savannah. As noted, Burke ceramics are most similar to those of the upper Savannah’s Tugalo phase (A.D. 1450–1600). Tugalo ceramics, in turn, were derived from those of the earlier Rembert phase (Hally 1990), when the entire middle Savannah River was abandoned. The Burke and Tugalo phases may thus hold a common antecedent in the preceding Rembert phase (Beck 2013: 52). Of course, we are much less interested in historical relationships among pots than in such relationships among people, and in making this link between the Rembert phase and the Catawba Valley Mississippians, we suggest that some of those groups who left the middle Savannah River after A.D. 1350 made new communities for themselves along the upper and middle courses of the Catawba. Processual archaeologists once rejected migration (also called site unit intrusion) as an explanation for social change (Smith 1984: 30), but a surge of new approaches, coupled with new data, has recently invigorated its explanatory value in southeastern archaeology, particularly with respect to Mississippian studies (Alt 2006; Blitz and Lorenz 2002, 2006; Pauketat 2003; Stoltman 2000; Williams 1994). This new research shows that identifying such large-scale movements of Mississippian groups over the southeastern landscape is essential for understanding how different historical trajectories unfolded. The most important identified Burke phase site is Berry (31BK22), and at just over 4.5 hectares it was one of the largest towns in the Catawba Valley and one of the few to have had a mound. The Berry site was briefly described by Cyrus Thomas (1891: 151) in Catalogue of Prehistoric Works East of the Rocky Mountains as a “mound on the west Bank of Upper Creek 8 miles north of Morganton (about 15 feet high and unexplored).” Both the earthen mound and its associated site have been continually plowed since the late eighteenth century, and in 1964 the mound was bulldozed to provide fill for a low-lying area north of the site that was susceptible to flooding and was the likely location of the original borrow pit. In 1986, David Moore directed excavations at Berry, focusing on a topographic rise at the

Joara in Time and Space · 43

northern end of the site that both the landowners and collectors identified as the remnant mound. His excavations here included a 10 × 70 foot trench that revealed deposits that he referred to as “basket-loaded mound fill” (Moore 2002: 214–216, plate 36), and he identified these deposits as the mound reported by Thomas. Moore’s trench revealed deposits of a form usually associated with indigenous earthen construction (e.g., Sherwood and Kidder 2011), and these likely represent some kind of earthen mound. We cannot yet say whether this is indeed a remnant of the feature that Thomas recorded, but throughout this book we refer to Moore’s excavation area as the mound area, to these fill deposits as mound contexts, and to all associated plant and faunal remains as mound assemblages. The only other Burke phase towns to have been the focus of extensive excavations are the Ensley (31BK468) and Catawba Meadows (31BK17) sites (see figure 2.2). Ensley is 2.5 kilometers east of Berry on Johns River. The work at Ensley was directed by Moore and Beck in 2005 as part of a salvage project to repair a flood-damaged section of riverbank where the erosion was unusually severe. Excavations revealed two buildings on the edge of the river: one was square and covered approximately 120 square meters; the second, northeast of the first, was round and measured 6 meters in diameter (figure 2.6). The size of the square building, Structure 1, along with analyses of its plant remains (Martin 2009) and lithic artifacts (Moody 2010), suggests that it functioned as a public building or town house (Moore and Beck 2005). If so, then such structures may have been common features in most Burke phase towns. Excavations at Catawba Meadows were also begun as a salvage operation to identify archaeological remains likely to be damaged by planned construction on the site. In 2004, archaeologists directed by David Moore completely excavated two large pit features identified through backhoe trenching that together yielded more than 1,500 Burke phase sherds and quantities of plant and animal remains. Then, from 2010 to 2011, Merritt Sanders, a graduate student at Tulane University, excavated a portion of a subterranean structure first identified during the earlier trenching operations (Geiger 2012). Aside from the five burned structures in the Berry site’s Spanish compound, this building—measuring about 8 meters on a side—is the only domestic architecture yet excavated on a Burke phase site.

Figure 2.6. Plan view of Structure 1 and Structure 2 at the Ensley site. Johns River is immediately beyond (northeast of) Structure 2.

Joara in Time and Space · 45

Burke Phase Chiefdoms The Berry site is one of the largest sites in the upper Catawba Valley, and investigations there have provided most of the excavation data for the Burke phase. Several distinct lines of evidence, both archaeological and documentary, have led us to identify Berry as the location of Joara, the center of a Late Mississippian chiefdom (Beck and Moore 2002; Beck et al. 2006). We present the evidence for our identification of Berry as Joara in subsequent chapters, but here we want to focus attention on our understanding of Burke phase chiefdoms. We use the chiefdom concept in reference to those historical episodes in which the leadership of one community forged a regional hierarchy that relegated neighboring towns to a tributary status (Beck 2003, 2013; Carneiro 1981; Earle 1978; Wright 1977). Our survey of the chiefdom literature suggests that the presence of a regional or multicommunity hierarchy has supplanted other features at the heart of the chiefdom concept (Anderson 1994, 1996; Beck 2013; Drennan 1991; Earle 1987, 1991, 1997; Hally 1996; Junker 1999; Milner and Schroeder 1999; Redmond 1998; Spencer 1987, 1990, 1994; Sturtevant 1998). Let us emphasize that we do not equate chiefdom political organization with evidence of Mississippian lifeways; not all Mississippian communities, that is, were necessarily integrated by regional polities. Four lines of evidence, however, do suggest that Burke phase populations were organized—at different times perhaps—into one or more such chiefdoms: mound construction and distribution, mortuary practices, settlement patterns, and sixteenth-century accounts. We have already described the unusual number of earthen mounds reported in the upper Catawba and Yadkin valleys. Although mound construction itself is not necessarily indicative of chiefdoms, it is significant that the Mississippian populations in this region (in contrast to contemporary Piedmont populations living to the north, south, and east) regularly participated in large-scale, corporate labor projects. Likewise, mortuary data alone are rarely used to identify chiefdoms in the archaeological record. We believe, however, that Mississippian mortuary practices in this region provide evidence of status differentiation during the Burke phase and that these practices, in combination with settlement pattern and mound construction data, can be used as another line of archaeological evidence to infer the existence of Burke phase chiefdoms. Two kinds of earthen constructions have been called “mounds” in the upper Catawba and Yadkin River valleys. The first, which includes the mounds reported in the late nineteenth century at the Berry and Michaux

46 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

Figure 2.7. T. F. Nelson Mound, Caldwell County (after Thomas 1894: 334, fig. 207).

Farm sites, was probably a form of substructure mound. Unfortunately, both of these mounds were significantly reduced in size prior to any modern documentation; the Berry mound was reportedly at least 4 meters tall before being bulldozed, while the mound at Michaux Farm was described as having considerable height. Three burials were excavated in the Michaux Farm Mound, but no other information was reported. Now it is all but impossible to determine whether structures stood on the summit of either mound. Examples of the second mound type include the Nelson Mound and Triangle, the Davenport Jones Mound, and the Lenoir Burial Pit. Cyrus Thomas’s reports (Thomas 1887, 1891, 1894) indicate that different numbers of individuals were interred at these sites in large, geometric-shaped pits. At the time of excavation, no mounds were visible above the pits, nor is there any record of raised mounds existing at these places. Plowing, however, might have eliminated any traces of smaller earthworks. Although each of these mound types probably represents distinct beliefs and practices, for our purposes here we consider both as large-scale, corporate labor projects and refer to them generally as mounds.

Joara in Time and Space · 47

The Caldwell County mounds on the upper Yadkin River provide the most evidence for Burke phase mortuary practices. Each mound contained multiple interments often accompanied by large numbers of artifacts, including shell gorgets, ceramic vessels, copper beads, perforated spatulates, and—in one case—iron artifacts. Descriptions of these mounds bear little resemblance to descriptions of platform or substructure mounds excavated in the nearby Appalachian Summit at Garden Creek (Dickens 1976; Keel 1976), Peachtree (Setzler and Jennings 1941), and Coweeta Creek (Egloff 1971; Rodning 2004, 2009). In contrast, the Caldwell County mound sites seem to represent formalized interment facilities, similar perhaps to the mortuary mound excavated at the Irene site in Georgia (Caldwell and McCann 1941). In addition, the number of exotic artifacts—in the form of European metal implements, shell masks and gorgets, spatulate axes, shell beads, and copper and mica plates—suggests high-status burials. Lack of excavation documentation makes it difficult to understand the character of the Caldwell County mounds; even so, they cast important light on mortuary practices during the Burke phase. The Nelson Mound and Triangle sites offer a typical example (figure 2.7). Thomas (1887: 61–63) described the Nelson Mound as follows: This mound, so insignificant in appearance as scarcely to attract any notice, was located on the farm of Rev. T. F. Nelson, in Caldwell county, North Carolina, on the bottom land of the Yadkin, about 100 yards from the riverbank. It was almost a true circle in outline, 38 feet in diameter, but not exceeding at any point 18 inches in height. The thorough excavation made revealed the fact that the builders of the mound had first dug a circular pit, with perpendicular margin, to the depth of 3 feet, and 38 feet in diameter, then deposited their dead in the manner hereafter shown, some in stone cists and others unenclosed, and afterwards covered them over, raising a slight mound above the pit. Sixteen burials, ten found in stone vaults, were distributed relatively evenly across the pit, while Burial 1, placed perpendicularly on its feet within a stone vault, held the central position in the pit. Use of space within the pit was orderly and planned and may indicate a single episode of multiple interment, although a more long-term use of the interment complex could also have been orderly and planned. Thomas (1894: 335) reported evidence for burning associated with many of these burials, conveying an impression of repetitive interments with ritual fires. Interment in this facility may have

48 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

involved defleshing of remains, application of paints or pigments, alignment of cairns, and ritual firing after completion. The Nelson Triangle, located about 75 meters south of the mound, is described as having been constructed in a manner similar to the mound with burials placed in a triangular pit that measured 48 feet on two sides and a base of 32 feet (Thomas 1887: 63). Burials 1–9 were single extended interments, while Burials 10–15 were placed in stone vaults like those in the nearby mound. Burials 11 and 14 contained two individuals each, and the only items associated with any of Burials 1–15 were a broken pipe and two polished celts. In contrast to the individual and dual burials was a mass interment at location A, where investigators found ten or more skeletons, all believed by the excavators to have been deposited at the same time. One individual, identified by Thomas (1887: 64) as the “principal personage of the group,” was surrounded by the other nine and was accompanied by a large inventory of materials, including an engraved shell gorget (figure 2.8c), a shell bead necklace, five copper beads, shell and copper bead bracelets, and one engraved shell filled with beads of different sizes. Also, four iron objects— two probable celts, a blade, and an awl-like tool (figure 2.8a,b)—were located at his right hand. Scattered among the surrounding nine individuals were polished celts, discoidal stones, copper arrow points, and plates of mica (Thomas 1887: 65–66). According to Thomas, the construction of both the Davenport Jones Mound and the Lenoir Burial Pit was similar to that of the Nelson Mound. The Davenport Jones Mound contained twenty-six burials, at least two of which included multiple individuals. Artifacts associated with most of the burials included Burke ceramic vessels, spatulate celts, shell masks and gorgets, and shell and copper beads (Thomas 1894: 338–342). The Lenoir Burial Pit contained skeletal remains of at least fifty-five individuals, many interred in large groups; included with the burials were polished celts, stone disks, shell beads, shell masks and gorgets, and Burke vessels (Thomas 1887: 68–70). In an examination of excavated materials now curated at the National Museum of Natural History, Moore (2002: 107) learned that the artifacts from the Nelson Mound and the Nelson Triangle were all designated “Nelson Mound.” He was able to confirm, though, that the published artifact lists seem to be accurate. The best-known artifacts from these sites are the metal objects. Moore (2002: 107–108) documented a small axe, an iron celt, and a portion of an iron blade. The celt is nearly identical to the one illustrated by Thomas (1887: 65, fig. 30; 1894: 337, fig. 211). It is undoubtedly

Joara in Time and Space · 49

Figure 2.8. Iron artifacts and Citico gorget from the T. F. Nelson Triangle, Caldwell County (after Thomas 1894: 337–338, figs. 211, 212, 213).

the second of the two iron celts reported as being at the right hand of the central skeleton in Group A of the Nelson Triangle; the iron blade is probably a portion of the iron blade illustrated by Thomas (1887: 65, fig. 31; 1894: 337, fig. 212). Thomas also reported copper beads and arrow points, along with four engraved shell gorgets, eight chunkey stones, a spatulate celt, and a large quantity of Burke pottery. Another feature of Burke phase mortu-

50 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

ary practice is the occurrence of Citico-style shell gorgets (see figure 2.8c). Although none has been found at the Berry site, archaeologists recovered two from a disturbed burial at 31BK56 (Ward 1980a, 1980b), and a collector found one at the Catawba Meadows site (31BK18). Several were also recovered at the Nelson Mound and the Lenoir Burial Pit on the upper Yadkin (Moore 2002: 107, 114). Citico gorgets are known from Alabama to Virginia, and most archaeologists generally consider them to have been produced from A.D. 1500 to 1625 (Smith 1987: 108–112). Although these Yadkin Valley mounds yielded a large quantity of mortuary data, the best-documented Burke phase mortuary data come from the Berry site, though only two burials were excavated there. Each was a shaft and chamber–type grave, characteristic of Pisgah phase burials in the North Carolina mountains (Dickens 1976: 102–132) and also common in the Dan River and Eno River regions from as early as the fifteenth century into the early eighteenth century (Hogue 1988; Navey 1982; Ward and Davis 1993: 407–432; Wilson 1983). Two individuals were buried within the single chamber of Burial 2 at the Berry site; such multiple interments are not common practice either during the Pisgah phase or among colonial-era Piedmont Siouans. Burial 1 was an extended adult male placed in a rectangular pit with a full-length side chamber. A burial bundle containing a turtle shell container, a clay elbow pipe, several projectile points, and stone abraders accompanied the individual. Similar bundles have been found in sixteenth-century burials at the King site in Georgia (Hally 2008) and the Toqua site in eastern Tennessee (Polhemus 1987). An iron knife—possibly a trade good from the Pardo expeditions (see chapter 10, this volume)—was placed across the individual’s upper chest and may have been worn suspended from his neck. In sum, data on mound construction and mortuary patterns suggest that (1) people in the upper Catawba and Yadkin valleys engaged in large-scale corporate labor projects, manifested in earthen mounds and mass-interment facilities; and (2) there is evidence of status differentiation in Burke phase mortuary assemblages. Both of these lines of evidence are often associated with, but are not limited to, the regional polities that archaeologists identify as chiefdoms. Additional—and we believe compelling—evidence for regional integration of the upper Catawba Valley during the Burke phase comes from settlement pattern data and documentary sources. Because of a lack of systematic survey and testing data, we are generally limited in our ability to discuss Burke phase settlement patterns within the Catawba Valley. However, Beck’s (1997a) survey of Upper Creek/Warrior

Joara in Time and Space · 51

Figure 2.9. Large Burke phase settlements near the Berry site (after Beck and Moore 2002: 200, fig. 7).

Fork near the Berry site reports a pattern that might have been more widespread during the Burke phase. Beck recorded and systematically collected data for twenty-two of the twenty-six known Burke phase sites in this watershed; all of the known Burke phase sites along Upper Creek/Warrior Fork are located in the valley’s floodplain (figure 2.9) and moreover are restricted to two specific types of floodplain soil. These sites appear to cluster into four different size classes, or modes. Of the twenty-one sites for which adequate size data exist, eleven are less than 0.5 hectares, seven are 0.8–2.0 hectares, two are 2.5–3.0 hectares, and one, the Berry site, measures 4.5 hectares. Beck refers to these four site size classes as Fourth Order, Third Order, Second Order, and First Order, respectively (figure 2.10). Berry, the only recorded First Order site, is the largest Burke phase site within the Upper Creek/Warrior Fork basin and is the only site in the basin with an earthen mound. The Second and Third Order sites are distributed at very regular intervals (1.44 kilometers, with a standard deviation of 200

52 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

Figure 2.10. Histogram of Burke phase sites along Upper Creek/Warrior Fork (after Beck and Moore 2002: 200, fig. 6).

meters) upstream and downstream from the Berry site. Six of the seven identified Third Order sites occur in pairs, while both of the Second Order sites are unpaired. There appears to be little patterning to the distribution of Fourth Order sites, which are distributed around and between the settlements of the larger three size classes. Beck suggests that this patterning represents a chiefdom settlement system characterized by three levels of integration: the household level, the community level, and the multicommunity level. The foci of householdlevel integration were domestic structures or compounds located in dispersed farmsteads (Fourth Order sites) and within nucleated villages (First, Second, and Third Order sites). The foci of the community level were the nucleated villages themselves; nucleated villages, that is, probably acted as local, community-level central places. Each First and Second Order site, with its associated farmsteads, and each pair of Third Order sites, with its associated farmsteads, probably represents a single community. The focus of the multicommunity level was Berry, the only First Order site identified within the study area and the only site in this basin with an earthen mound. Berry probably acted both as a local center for its associated community

Joara in Time and Space · 53

and as a regional center for all of the communities in the Upper Creek/ Warrior Fork basin. Given that archaeological evidence supports our identification of the Berry site as Joara (Beck et al. 2006), accounts from the Pardo expedition describing the region’s political structure are highly pertinent to our reconstruction of Burke phase chiefdoms. First, notary and interpreter Domingo de León recorded that multiple local communities were associated with the “province” of Joara. According to León, it was “populated with many towns, although small. Since they are very warlike, they are scattered” (Worth, chapter 3, this volume). Although León did not record the number of towns existing at the time of his description, his record of the expedition does suggest that the Burke phase polity visited by Juan Pardo at Joara consisted of multiple towns. Second, of two reports by Juan de la Bandera, Pardo’s notary and scribe, the second and more extensive suggests that there were at least two administrative offices in the upper Catawba Valley’s aboriginal political system: mico, or great lord (un gran señor), and orata, or minor lord (un menor señor) (Hudson 1990: 63). Bandera records that while Pardo encountered at least 120 oratas over the course of the second expedition, he dealt with only three micos (Hudson 1990: 63), one of whom resided at the town of Joara (that is, the Berry site) and was referred to as Joara Mico. Although Bandera often refers to micos and oratas as caciques, or chiefs (Hudson 1990: 61), the mico clearly had greater political authority than did the orata (Anderson 1994: 96–97). On one occasion, for example, Pardo met 19 ca­ciques at Joara, each of whom, with the exception of Joara Mico, was an orata; of the 19 chiefs, Bandera describes Joara Mico as “the most important one” and refers to 5 of the oratas as “caciques of ” Joara Mico (Bandera 1990a: 278). We believe that most of the 19 oratas who met Pardo at Joara likely traveled to Joara from other Burke phase communities along the upper Catawba and Yadkin rivers and that Joara was the most important midsixteenth-century political center in this region. Although we do suggest that there is a correlation between the geographic extent of Burke phase pottery and the homes of the 19 oratas, we do not equate the scale of Joara Mico’s administrative hierarchy with that of the entire Burke phase area. Rather, it is likely that Joara Mico’s political authority extended over a much smaller region, probably represented by the 5 local-level oratas who Bandera notes were his caciques; these 5 oratas, that is, may have represented the Burke phase communities located much closer to the Berry site on Up-

54 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

per Creek/Warrior Fork and Johns River. Documentary accounts and archaeological evidence suggest that the Burke phase chiefdom centered at Berry was marked by two levels of hierarchy: the archaeologically identified local level correlates with the historically documented office of orata, and the archaeologically identified regional level correlates with the historically documented office of mico (Beck and Moore 2002: 201). Taken together, written accounts and archaeological data from the upper Catawba Valley suggest that regional chiefdoms had emerged there by the mid-sixteenth century. Moreover, evidence of mound building and mortuary complexity suggests that multicommunity polities may likewise have appeared on the upper Yadkin during the Burke phase. The rise of Mississippian chiefdoms along the upper courses of these two rivers, located in the foothills of the Appalachians, contrasts with the lack of such polities in most other Piedmont locales north of the Fall Line. We know of only two other likely chiefdoms in the North Carolina Piedmont: the polity centered at the Town Creek Mound (31MG2) on a tributary of the lower Yadkin River and the Guatari polity described in accounts of the Pardo expedition and probably centered in the heart of the Caraway phase just north of Town Creek on the middle Yadkin. What factors contributed to the rise of chiefdoms in the Appalachian foothills but not in most adjoining locales? We suggest that two specific factors, an ecotone setting and a location along major trade routes, had particularly important roles in the success of Mississippian strategies along this southern Appalachian frontier. Hally (1994b: 159–161) has suggested that river-deposited soils located just below the Fall Line are significantly more rich in nutrients than are sediments immediately upstream or farther downstream and that these soils are deposited in greater quantities at the Fall Line because of a rapid decrease in stream gradient. He also notes that the juxtaposition of two physiographic provinces, the Piedmont and the Coastal Plain, would have provided a broader array of plant and animal resources at Fall Line locales. Hally (1994b: 162–163) extends this model to Mississippian settlement patterns on the Great Smoky Fault in northwestern Georgia, suggesting that sediments along this fault are more nutrient rich than soils just upstream or downstream, and that a similar juxtaposition of the Blue Ridge and the Valley and Ridge provinces provided an ecotone ecology similar to that located along the Fall Line. Meyers (1995) systematically tested soil fertility along the Great Smoky Fault to address the role of this variable in the clustering of large Mississippian sites just west of the fault in the Valley and Ridge province. After

Joara in Time and Space · 55

analyzing floodplain soil samples collected above, below, and at the fault along both the Etowah and the Coosawatee River, she concluded that there appears to be no significant difference in the fertility of these samples. This suggests that fertility played a less important role in the fault-line clustering of large Mississippian sites—including Etowah and Little Egypt—than the resource variability offered by this ecotone locale (Meyers 1995: 95). We suggest that the emergence of Burke phase chiefdoms along the upper Catawba and Yadkin rivers is at least partially due to the existence of an ecotone at the interface between the mountainous Blue Ridge province and the Carolina Piedmont. As with the Fall Line and Great Smoky Fault ecotones, this “foothills ecotone” would have provided access to natural resources located in two physiographic provinces. Without favoring environmental determinism, we offer that this ecotone locale provided wouldbe leaders a favorable arena in which to pursue some of those strategies that we refer to as Mississippian. Pursuing the foothills ecotone as a framing factor in the rise of Mississippian polities in this headwaters region, we may ask how this ecological factor influenced the ability of emergent leaders to harness the organizational and administrative opportunities provided by Mississippian ideology. Mississippian chiefdoms and their associated institutions were supported through the production of an agricultural surplus. Because of a lack of agricultural intensification, the leaders of Mississippian chiefdoms could increase their access to surplus only by increasing their access to human labor; much of Mississippian ideology therefore emphasized groupbuilding strategies, in which would-be leaders competed with one another to persuasively attract new followers (Beck 2003, 2006). For persuasive strategies to be effective, would-be leaders required locales with the potential to sustain an increase in producers, particularly through periods of agricultural shortfall. That is, Mississippian polities tended to emerge in areas where people had access to a variety of potential food resources (Smith 1978); resource variation served both to supplement agricultural production during times of bounty and to counter the need for fissioning in times of agricultural stress. By dispersing human labor, stress-induced fissioning undermined the political economies of persuasively organized chiefdoms (e.g., Anderson 1994), and we suggest that Mississippian strategies were most successfully pursued in areas that offered a higher degree of population permanency (Smith 1978), whether the classic meander belts and oxbow lakes of the Mississippi Valley or the southern Appalachian ecotones discussed in this chapter.

56 · Robin A. Beck, David G. Moore, and Christopher B. Rodning

Current archaeological evidence suggests that the upper Catawba and Yadkin valleys had higher population densities in the late precolonial era than did adjoining areas of the North Carolina Piedmont, and we believe that this density was attributable, in part, to the greater resource variability of the foothills ecotone. Adjacent Piedmont locales, apart from the Fall Line, were probably unable to maintain comparable population densities through periodic episodes of agricultural stress, and we suggest that these zones were less capable of sustaining Mississippian chiefdoms. Mississippian polities along the upper Catawba and Yadkin valleys were also situated at the crossroads of a network of aboriginal trails that joined this region to Mississippian peoples in northwestern Georgia, eastern Tennessee, southwestern Virginia, and the South Carolina Piedmont (Beck 1997b). Documentary accounts and archaeological evidence suggest that these polities were all engaged in the exchange of exotic raw materials. The Chiscas in southwestern Virginia were exchanging salt and copper to Indians farther south and were joined by an important north-south footpath to the Berry site along the upper Catawba River. Berry (Joara) was likewise connected to polities in central South Carolina, along the same network of trails that Soto and Pardo used to approach the Appalachians, and Indians encountered on the South Carolina coast as late as 1605 knew of Joara, which they located deep in the interior at the base of the mountains (e.g., Hann 1986). The proto-Cherokee populations of the Appalachian Summit also had exchange relations with the Joara chiefdom, by way of an important trail that ran directly from the Appalachian Summit area through Swannanoa Gap to the upper Catawba Valley; James Mooney (1900: 532) notes that the word “Swannanoa” is derived from the Cherokee term sualinunnahi, meaning “the Suwali (i.e., Xuala or Joara) trail” (Hudson 1997: 188). Finally, an important trail also connected the upper Catawba and Yadkin valleys to Mississippian polities in northwestern Georgia (e.g., Myer 1928: 748), and the presence of Etowah-style complicated-stamped ceramics at sites in the upper Catawba region (Moore 1999) suggests that these interregional relations predated the Burke phase.

Conclusions During the sixteenth century, the Berry site (that is, Joara) sat at the northeastern edge of the Mississippian cultural world and at the northwestern edge of the Spanish colonial frontier. Berry was the political center of a Lamar Mississippian chiefdom, one of many such polities across the southern

Joara in Time and Space · 57

Appalachians and Coastal Plain from A.D. 1300 to 1650. Systematic surveys north and south of Berry have identified more than two dozen sites with Lamar ceramics, and many of these sites were probably contemporaneous with Berry. We suggest that these neighboring sites are the core of Berry’s polity, which was at its height during a period of occupation on the upper Catawba and Yadkin rivers known as the Burke phase (A.D. 1400–1600). The long-term goals of our research program are to understand the process of social change during the Burke phase; to understand the role that Spanish colonial strategies had in this process and in the process of polity formation; and to learn of Joara’s role in Catawba Indian coalescence in the seventeenth and eighteenth centuries. Excavation of Fort San Juan and Cuenca thus addresses one of these key long-term aims—evaluating the nature of Spanish-native interactions at Joara—while also offering a temporal benchmark for the study of postcontact transformations across the Carolina Piedmont.

3 Recollections of the Juan Pardo Expeditions The 1584 Domingo de León Account John E. Worth

The 1566–1568 expeditions of Juan Pardo into the Appalachian Summit region from the newly founded Spanish colonial base of Santa Elena represented Spain’s third major exploratory venture into the deep northern hinterland of southeastern North America. Following the 1539–1543 expedition of Hernando de Soto’s army and the failed colonial venture of 1559–1561 under Tristán de Luna y Arellano, Juan Pardo’s dual expeditions were the last direct contact that Spain would ever have with many of the indigenous chiefdoms of the interior of modern-day South Carolina, North Carolina, and eastern Tennessee (Hudson 1990). Although the newly established Spanish colony known as Florida would experience comparatively rapid expansion over the next decades, the deep interior lay beyond the reach of the perpetually undermanned, underequipped, and unprofitable garrison town of St. Augustine, eternally dependent on external royal support and local Indian labor for survival (Bushnell 1994; Milanich 1999; Worth 1998). Only a handful of Spaniards ever visited the northern interior again, and even these minor expeditions never penetrated farther north than Cofitachequi in central South Carolina (Worth 1994). The Pardo expeditions thus effectively represent the final glimpse of aboriginal societies along the fringes of the southern Appalachian Mountains, for when English explorers, traders, and colonists finally reestablished direct contact with the Indians of these regions, their cultures had in many cases experienced drastic and irreversible changes in the century after first contact with the Spaniards (e.g., Ethridge and Hudson 2002; Hudson and Tesser 1994). The written accounts of the Pardo expeditions reveal cultures on the literal verge of transformation and thus provide invaluable information for historians and anthropologists alike. To date, however, only a few

Recollections of the Juan Pardo Expeditions · 59

such documentary accounts have been discovered in the Archivo General de Indias in Seville, Spain. Four contemporaneous accounts are generally cited, including a short account by Juan Pardo, an account written by soldier Francisco Martínez and witnessed by several other soldiers on the expedition, and two accounts by Pardo’s notary, Juan de la Bandera—one representing the day-by-day affairs of Pardo’s second expedition (the “long” Bandera relation) and a summary based on this relation (the “short” Bandera relation).1 These four accounts are additionally supplemented by the 1600 testimony of a soldier who participated in the expedition, Juan de Ribas, and two Indian women taken as young girls during the expeditions, Luisa Méndez (Ribas’s wife) and Teresa Martín (widow of Pardo soldier Juan Martín de Badajoz, who perished in the 1576 Indian rebellion around Santa Elena).2 These last accounts, however, represent recollections some thirty-two years after the fact and consequently contain many exaggerations and errors. Yet another account, postdating the Pardo expeditions by only sixteen years, remained effectively unknown by modern researchers until the mid1990s, even though it contains several important details not mentioned in any of the remaining accounts. This account, penned by soldier, interpreter, and notary Domingo González de León in 1584, forms part of a longer letter to the king of Spain in which he relates many of the difficulties experienced by the young Florida colony and proposes the establishment of a Spanish colonial city in the interior at the Indian capital of Canos (Cofitachequi).3 Perhaps in part because the letter was written in a remarkably miserable script, with numerous eccentric usages and outright grammatical errors (substandard even in comparison with typical late sixteenth-century script), this document has never been used as a primary source relative to the Pardo expeditions. Nevertheless, its importance relative to the expeditions themselves (and the course of Pardo’s route) is supplemented by the fact that León provides information on the aftermath of the expeditions, including the fate of some of the Spanish soldiers left to man the string of forts built by Pardo, and some intriguing details regarding social changes being experienced by Indians in the deep interior during the first years following the expeditions. Consequently, a full English translation of the relevant portion of León’s letter is presented in print for the first time later in this chapter, along with supplementary discussion regarding León’s service record and qualifications and some of the implications of his new information regarding the Pardo expeditions.

60 · John E. Worth

Domingo de León’s Service in Spanish Florida Domingo González de León (more commonly known as Domingo de León by his companions)4 arrived with the adelantado Pedro Menéndez de Avilés in 1565 to found the first permanent settlement in Spanish Florida at St. Augustine. He served under Bartolomé Menéndez in the expulsion of the French from the mouth of the modern St. Johns River and at the Spanish fort San Mateo there and subsequently participated in Pedro Menéndez’s 1566 establishment of Santa Elena to the north, on modern-day Parris Island along the South Carolina coast.5 According to his descriptions of his service, as well as the supporting documentation and testimony included in his 1586 service record, León was “the first soldier who learned the language of the Indians” (referring to the area around Santa Elena, within what would later be known as the Escamaçu province north of Guale).6 Indeed, if we are to accept the 1579 testimony of Gonzalo Vicente, at the time of the establishment of Santa Elena “there was no one who understood the language of the Indians apart from said Domingo de León.”7 As a consequence, he soon became an important lengua, or interpreter, of the Indian languages and was used frequently over the course of the next decade in that capacity (although he was not given an official title at the time, or an increase in his pay, for this work). In part for this very reason, León was selected as a part of the first major exploratory venture mounted by Pedro Menéndez de Avilés after the establishment of Florida. The fact that Domingo de León actually participated in the Juan Pardo expeditions of 1566–1568 is well documented. All witnesses called in 1579 and 1584 to demonstrate his merits and services to the Crown, most of whom were soldiers who apparently participated in the Pardo expeditions as well, testified that León served under Captain Pardo’s command at Santa Elena and was a part of his expeditionary force into the interior. As León himself related, he served “in the interior that Captain Juan Pardo entered, looking for food among the said Indians, and supplying himself where he had to go, by being the most experienced and ancient in the land, and understanding the said Indians and having much credit with them in both the interior and the coast.”8 This account was confirmed, for example, by the thirty-three-year-old soldier Pedro Luís, who testified in 1579 that “the said Domingo González de León served in the company of the said Juan Pardo as soldier and interpreter, both within the interior and in any other places where he was ordered, and this witness [Luís] knows because he . . . went sometimes with the said Domingo de León among the Indians by order of

Recollections of the Juan Pardo Expeditions · 61

the said captain and always saw that the said Domingo de León dealt with much truthfulness . . . with both Indians and Spaniards.”9 Other witnesses confirmed these statements, providing clear documentation that León actually witnessed the places he describes in his 1584 account. Though only one soldier among well over a hundred, Domingo de León was even mentioned once by name late in the “long” Bandera relation, further proving his participation in (and survival of) the second Pardo expedition. On February 28, 1568, in the Indian town of Orista (named Buena Esperança by the Spaniards), notary Juan de la Bandera recorded that Captain Juan Pardo “gave an order and commanded Antonio Diaz Pereyra, Portuguese, and Domingo de León, and Juan Toribio, soldiers, to go to Guando Orata which is where there is a house [full] of His Majesty’s maize, and of that maize to try to bring, with [the aid of] the Indians of the said cacique, all that they could to the town of Buena Esperança, which in the Indian language is called Orista.”10 This brief mention reveals that the soldier Domingo de León did indeed return from the interior with Pardo’s army following their second and more extensive journey during the fall and winter of 1567–1568. Internal evidence in León’s 1584 relation hints that León was also a participant in the first expedition the previous winter, as his description and naming of the Chisca towns destroyed by Sergeant Moyano implies that he likely remained in the interior between the two expeditions. Consequently, Domingo de León may well have spent more than fourteen months in the deep interior, providing him ample opportunity to learn about the social and physical geography of these distant lands. Interestingly, León provides no detailed references to any Indian groups on the other side of the Appalachian Summit, discussing these groups instead in the somewhat “legendary” context of the recollections of Indian women brought back from the interior as young girls by Pardo (see below). This may indicate that León never actually visited these regions personally, perhaps having remained behind after the Chisca raids as one of a small contingent of soldiers first garrisoned by Moyano in Fort San Juan, constructed during Pardo’s first expedition in the Indian town of Joara (now believed to have been at the Berry site), and probably remaining there during Pardo’s subsequent trip over the mountains.11 Nevertheless, León’s information regarding Indian societies on the eastern side of the Appalachian Summit was clearly based on his own personal experience during the Pardo expeditions. Following his return from the Pardo expeditions (along with the handful of soldiers who were not left to man the string of forts left in the deep

62 · John E. Worth

interior),12 León continued to serve at Santa Elena for the next eight years, serving after Pardo’s return to Spain under the command of Captains Juan de la Bandera and Alonso de Solís. Following his marriage to Maria de Arango, in 1572 he was given the official title of notary public in the fort at Santa Elena by Pedro Menéndez de Avilés.13 After the death of the adelantado late in 1574, Domingo de León was renewed as notary public by Menéndez’s son-in-law, Lieutenant Governor Diego de Velasco, on January 22, 1575, and six days later was additionally given the official title of ecclesiastical notary for the fledgling Franciscan custodia of Santa Elena by Fray Diego Moreno.14 León’s new dual offices would ultimately result in a serious conflict of interest, for within a few months, a jurisdictional struggle between Fray Moreno and the acting governor erupted, placing Domingo de León in the unenviable position of serving as the official notary for both sides of the quarrel (Lyon 1992). At the height of the dispute, León was instructed by Fray Moreno to publish his order to excommunicate Governor Velasco. In a rage, the governor ordered León to turn Moreno’s order over to him, and when he refused (thus effectively siding with the Franciscans against the governor), Velasco is said to have slapped and punched León before having him arrested, placed in stocks, and publicly stripped. The excommunication order was seized by force, León was removed as notary, and Governor Velasco seems to have suppressed any further action in the matter.15 Such political squabbling undoubtedly added to the growing disenchantment among the settlers at Santa Elena, whose dreams of life in the young colony had not materialized. Immediately following the February arrival of Pedro Menéndez’s official successor, Hernando de Miranda (the adelantado’s other son-in-law, who promptly arrested Velasco on a variety of charges relating to poor government), a group of settlers at Santa Elena petitioned the new governor for permission to abandon the troubled colony and relocate. On February 27, 1576, three days after Miranda’s arrival, Domingo de León was among the soldiers and settlers who signed a formal petition expressing their dissatisfaction with Santa Elena and indicating their desire to leave.16 On April 3, Domingo de León left Florida in the company of Governor Miranda, reportedly having been elected as a representative by other soldiers to present their case in person to the Spanish Crown. Years later, Miranda testified that “the stated soldiers who served [at Santa Elena] elected [León] as the person who best understood what was suitable, on account of what was happening in those provinces [of Florida], so that he should come

Recollections of the Juan Pardo Expeditions · 63

and give a relation to His Majesty of what was being done there, and the necessity that was suffered there.” Miranda stated, “Having seen and understood this, I gave him license to come and give an account to His Majesty of what was suitable.”17 Once in Havana, while the governor arranged for the salaries of Florida soldiers, León was given official license by Miranda in mid-May to leave for Spain as the official “procurador of the soldiers of Santa Elena and St. Augustine.”18 While León was on his way to Spain and Miranda wrapped up his business in Havana, Alonso de Solís, the governor’s appointed lieutenant during his absence, made a series of mistakes that resulted in a violent general uprising among the Indians along the coast around Santa Elena and to the south. Shortly before the governor’s return, bloodshed erupted around Santa Elena as the Indians of Orista and Guale mounted an armed rebellion against the increasing abuses of Solís and the soldiers under his command (including the execution of several important Indians).19 The uprising during the summer of 1576 resulted in the June murders of Ensign Moyano (of Pardo expedition fame) and all but one of the soldiers under his command in the nearby town of Escamaçu and the murders of all the royal officials of Florida and their entourage in the Guale town of Ospogue (on the southern end of Sapelo Island). After Miranda’s return, Solís was ambushed and killed with his men near Santa Elena. A few days later, surrounded by hostile Indians, Governor Miranda abandoned the town of Santa Elena, which was subsequently burned to the ground in view of the Spaniards (Lyon 1987). In Spain, presumably following his report on the conditions in Santa Elena, León was ordered to accompany the royal visitor Baltasar del Castillo to Florida on his official investigation of the colony. When news of the abandonment of Santa Elena arrived, however, León successfully petitioned the Council of the Indies for permission instead to relocate with his family to New Spain. With official license, upon his return to Cuba León took his wife and children, along with his mother-in-law and her other children, and moved to the city of Mexico in New Spain.20 Domingo de León’s political problems were not over, however, for early in 1579 the new governor of Florida, Pedro Menéndez Márquez, dispatched an official request that Domingo de León be apprehended and returned to Florida, indicating that he was absent without leave and owed certain debts and was urgently needed in Florida as an interpreter. According to León, Menéndez made a “false relation,” accusing him of crimes he did not commit, in order to have León forcibly taken back to Florida for service

64 · John E. Worth

in the ongoing negotiations with rebellious Indian leaders around Santa Elena. León asserted that the reason for the new governor’s actions was revenge: “I had come to give news to Your Majesty in the year of seventy-six regarding things that were done [in Florida] in harm to Your Majesty and the republic, in which [Pedro Menéndez Márquez] participated, suffering and concealing his debts, and through this route he attempted to take vengeance on me.”21 While in New Spain to collect the colony’s yearly royal subsidy, or situado, Florida treasurer Martín de Quirós (acting on behalf of Governor Menéndez Márquez) petitioned don Martín de Enrique, the viceroy of New Spain, to have León arrested. Incarcerated on February 11, 1579, Domingo de León spent nearly two months in jail before an agreement was reached in which he was permitted to return to service in Florida, where his future salary would be used to pay a fine of 350 pesos, 50 of which were to be for his transport and upkeep during the trip to Florida, and 300 of which were designated as debts to be repaid.22 León was finally released on April 7 and was transported by horse cart to the port of Vera Cruz, where he arrived just over a week later to embark for Florida.23 Domingo de León finally arrived in Santa Elena on June 23, 1579, to serve as an interpreter and mediator with the Indians of that region.24 In mid-August, following Governor Menéndez Márquez’s arrival to pay the soldiers of Santa Elena, León evidently accompanied the governor on an expedition against the province of Orista, where the governor’s force of some sixty soldiers was surrounded and attacked by several hundred Indians about 15 leagues from Santa Elena. Although fourteen soldiers were wounded (including León, according to all his witnesses), the Indians were routed, and after briefly returning to Santa Elena, Governor Menéndez Márquez moved on with a larger force to destroy the town of Coçapoy (20 leagues away) at the end of the month.25 In mid-September, Domingo de León began to assemble a record of his service for use in petitioning the Crown for a redress of grievances related to his recent imprisonment. Nevertheless, León evidently proved to be of invaluable service in Florida, for only a month later he served as the interpreter when the cacique of Coçapoy finally agreed to peace with the Spaniards, officially rendering obedience to the king of Spain.26 During these years, however, Governor Menéndez Márquez repeatedly opposed Domingo de León’s efforts to leave and present his case in Spain and seems to have held León a virtual economic captive in Florida during these years, not permitting him to leave unless he first paid the money charged to him

Recollections of the Juan Pardo Expeditions · 65

during his imprisonment in New Spain.27 Nevertheless, in 1583 Domingo de León seems to have befriended the Franciscan friar Alonso Reynoso, who had purportedly been thwarted by Governor Menéndez Márquez in his desire to send friars among the Indians. The following year, Reynoso “obtained [León a] license [for passage] down to Havana with permission of the said general [Menéndez Márquez].” León continued, From there onward I took my own license, although with more than enough trouble, because they did much so that I would not come [to Madrid], placing a fine of five hundred ducats on whoever might bring me or hide me or feed me, and a punishment of whipping, as will be found in his auto, to whichever person who might help me to Spain or New Spain, all with the goal that I should not come and make a relation of what is happening [in Florida], all of which was not sufficient, since I came, leaving my wife and children alone in order to do what I am obligated.28 León left Florida for the last time on April 6, 1584, eventually making his way to Madrid, where he penned his relation. Domingo González de León’s scathing indictment of Pedro Menéndez Márquez’s governorship formed the core of his 1584 relation, a portion of which is translated below. Although the preceding account indicates that he reached Spain without an official leave of absence from the governor of Florida, León nonetheless appears to have been remarkably successful in various petitions made before the Spanish Crown over the course of the next two years, strongly suggesting that León had indeed been simply a victim of the political turmoil in Florida. In a series of royal cedulas issued in March and April 1586, all of the previous injustices suffered by León were corrected, including the negation of the 350-ducado debt charged to León as a result of his arrest in New Spain, as well as the payment of a salary to León for his years of service as an interpreter.29 Unfortunately, this payment of salary minus rations and clothing was not actually made until five years later, on October 23, 1591, and then only posthumously to León’s widow, Maria de Arango, and his two daughters, Luisa and Isabel, then living in Havana.30 His rather extensive record of service indicates that Domingo de León was a remarkably able interpreter and furthermore knew the Indians and their customs quite well. For example, former governor Hernando de Miranda testified in 1584 that “it was very public and known among all the soldiers that he was one of the most ancient persons who were there, and

66 · John E. Worth

of the greatest benefit with his language and solicitude, and to whom was entrusted all the things that needed to be negotiated and communicated with the said Indians, by the said Domingo González understanding the languages, and knowing the lands and their manner of living, and he was likewise a person of credit and confidence.”31 León’s description of his own merits as an interpreter was equally glowing in 1579: “I made all the journeys among the Indians, and my opinion was always very accurate in everything that was dealt with among the Indians for our conservation and peace with them, and in what pertained to the Indians my opinion was always taken, and in the journeys I followed my orders and remained in accordance with what I understood from the Indians.”32 Five years later León elaborated, noting that regarding knowing the land and language of those provinces [of Florida], no disgrace has ever happened to me as has happened to other captains and interpreters, two perishing in the power of the Indians, because whoever does not know their languages does not understand the Indians well, nor can anyone who does not know the land make journeys, because without doubt they will be lost, and thus the past generals and captains have had much confidence in my person, not only entrusting me with that which was necessary with the Indians but also with the affairs of the Spaniards, naming me royal notary.33 Other witnesses further commented that León was not only an able and willing interpreter but also one who could be trusted to be truthful in his dealings with Spaniards and Indians alike. Pedro Luís noted that León “did what his captains and superiors entrusted to him well and faithfully, without there being any deceit whatsoever, and if anything else had occurred, this witness would have known it, by always being present, and having seen it.”34 Taken together, the evidence presented above indicates not only that Domingo de León was an eyewitness to the Juan Pardo expeditions but also that he was an able and trusted interpreter, with wide-ranging experience among the Indians around Santa Elena both before and after the expeditions. Consequently, the account translated below may be presumed to be largely reliable, at least to the extent of León’s own knowledge and memory as of 1584. Inasmuch as the details of his description of the deep interior did not directly figure in León’s personal agenda against Governor Pedro Menéndez Márquez, there seems no reason to suspect intentional exaggeration or equivocation within this portion of his 1584 account. Of course,

Recollections of the Juan Pardo Expeditions · 67

the degree to which León’s memory can be trusted after sixteen years is difficult to judge, but most of the details (names, places, and so forth) that can be cross-referenced with the other contemporaneous Pardo accounts seem accurate. As a result, the León account should serve modern historians and anthropologists as yet another firsthand account (although a later recollection) relative to the important Juan Pardo expeditions, filling in several details left out by other documentary sources and adding new information regarding their aftermath.

The Domingo de León Account Archivo General de Indias, Santo Domingo 231, folios 316–318 Domingo González de León to the King, October 13, 1584 (Madrid) [f. 316r] Your Royal Catholic Majesty I am a soldier who has served Your Majesty in the provinces [of Florida] twenty years as a soldier in the province [lengua] of Santa Elena, since the entry of Pedro Menéndez, who is in glory, [and] who ejected the French from there. I did not leave [Florida] until the year of seventy-six, when I came to this court to make a relation of the things that were happening in [Florida], elected by the soldiers with license of the general of the said provinces, all of which time I served in the company of Captain Juan Pardo, in the entrances of the interior, in what was necessary to negotiate with the Indians, including supplying him both with food and with Indians to carry burdens where he was to go, and where it was necessary to give presents, and having canoes where it was necessary for the service of the said company, and providing the fort with corn when it was necessary, and when there was need among the Indians, as is public among everyone, and afterwards [I served] with the rest of the following governors in the pacification of the said Indians, as I have said, conserving them with the Spaniards, as is public and well-known . . . [f. 316v] . . . [From] what I understand of this land, having walked [across] much land, both on the coast and in the interior, as I have said, and by the information that I have from the Indians, if Your Majesty would join the two forts into one on good land, like there is where Juan Pardo entered, which the Spaniards call Canos, forty leagues from Santa Elena to the north, and there make a city, even if no more soldiers than those that Your Majesty has in Santa Elena and in St. Augustine were to go, although Your

68 · John E. Worth

Majesty might not add more people, with three hundred men that there are in the fort, two hundred of which to serve in guarding the fort and the [remaining] hundred married settlers, Your Majesty helping them with the ordinary pay until they establish themselves, and once adapted, they will sustain themselves and the number of the one hundred [settlers], inserting other needy [people] until they recover themselves. In this way many people from Havana and those islands, including New Spain, could be united, Your Majesty giving them repartimientos as has been done in New Spain, because there is a good disposition of land, much corn is sown in abundance, [and there is] a good type of Indians, and in this Canos there are four Christians—a Flemish man with his wife and two daughters, who was a fife of Juan Pardo who remained from the war when the hundred and fifty soldiers were [killed],35 and I certify that if four hundred ducats were offered they would have us there now, even though there might be more danger, and there is even news that there are plenty of people in that land from those of Juan Pardo, who were fed up with the case for the discovery of this land. It is only suitable that if this settlement is to be made that Your Majesty pay extra [les aga merced] to those who know the land, and to the first who enter and take on a great deal, in order to make use of them, permitting that they may enter and leave when they wish, or at least from time to time, I assume at least that if this were done it would not be necessary to take people from Castilla if there were [free] entrance and departure, because in hiding the men away, there is no one [f. 317r] who would go there, including men and women. When I left from Florida, there [also] left a boy of eight to ten years, who was the son of an old soldier, now dead, who served Your Majesty many years. He left hidden until Havana, and in the boat [barco] he was found by a captain who was going there for Pedro Menéndez. [The captain] tied his hands to a pulley and began to move it in order to maim him, and he would have maimed him had not some soldiers who were coming there been at hand. [All this occurred] without the boy earning a salary or saving anything of that which he lacked. On account of such matters, in that land they are so denied the desire that if someone wants to send his children outside [Florida] to school or studies, they do not give permission. I certify that if Your Majesty pays them extra as I have said, giving them license to enter and leave, that there would be more settlement in one year than in twenty years that this [settlement] is being settled, apart from this river [mentioned above] being good for settlements. I understand it to be one of the most beneficial that has been discovered until now, because it ascends up to the mountain range [sierra]. I have walked more than enough of it, and

Recollections of the Juan Pardo Expeditions · 69

it is very copious. Thirty leagues from the coast two rivers join, and only one comes out at the sea. One arm goes to the northwest and the other to the west. That which goes to the west reaches up to the mountain range and returns something like southeast, and there it comes out from a very large lake [laguna] which extends [questa dendida] from north to south, very populated roundabouts. Those from New Spain say that they have reached it. They call it the River of the Shells [el río de las conchas]. From this lake to the mountain range there is very little [distance], and this I know by relation of the Indians, and they say that a man placed on top of the mountain range sees the other sea [mar] and that on this side, because due to the greatness of the lakes they call them seas. The [lake] on the other side is a very large, round, populated lagoon. In the middle of that lake comes forth a very copious river that goes to the other sea of the west. It is a very furious river of great current. The Indians say that on top of the mountain range one sees the one sea and the other. It is not very far, and this they affirm as certain. By what I have traveled with them, I find them certain and truthful, [because] I saw the very same on a map that I have traveled in person,36 and the only area where I disagree is that there will be fifty leagues from sea to sea—I [mean to] say from lake to lake— taking it by its measurement. This river that goes in the direction of the west is depopulated something like forty leagues because some Indians have war with others, and thus the district is depopulated. A cacique of that part of the mountain range descended to this [part] of this [side] because of the tremendous war that they have there with others, which would serve for whatever news that might offer itself. The other river that goes to the northwest goes through settlements [poblaciones]. Between these two rivers is a town [pueblo], [and] from it the towns ascend at half a league and a league [apart]. From this town to that which is called Canos it will be twelve leagues, all upriver. [The river] runs populated on the east bank by little settlements [poblaciones chicas] more closely joined, all in a district for the distance of twenty leagues [toda en comarca por parte veynte leguas]. There is another cacique to the northwest, and all upriver up to the mountain range. It is populated with many towns, although small. Since they are very warlike, they are scattered [andan esparcidas]. These give much news of everything. If an Indian [were able to] go to and come from New Spain in brief time, they [would] refuse even more strongly that we Spaniards should meet each other. They never reveal what there is except under secrecy, including Indian men and women, because they say that if they let us pass or advise us of what there is, those from that land would make war on them, and the others also tell them

70 · John E. Worth

that they should not let the Spaniards pass, because upon the Spaniards meeting each other, we would then have to be captives. The Indians relate this speaking in conversation nonchalantly, and the [f. 317v] women tell us the same, that we should not pass to the other side, because on that side is a great cacique [who] eats people, and that only in order to eat people does he go to make war on his enemies, and that when he has no enemies he eats his own [people], and the women tell us this feeling pity for us, because they are very pious and loving, and if it were not for them and their advice, we would have received much damage, and thus there is no [reason] to discuss this until overdoing it. Across this plain [llanada] and mountain [montaña] there are walnuts [nogales], wild grapes [ubas salbaxes], many chestnuts [castañas], plums [ceruelas] of three or four kinds, apples [mançanas] of this very land, although small, many acorns [vellotas] in abundance, deer [venados], humpbacked cows [bison—bacas corcobadas] on the other side of the mountain range, many doves [palomas], the number of which cannot be counted, such that even on the coast a group of them takes an hour to go to the ground in search of acorns,37 hens [gallinas], many geese [gansos], ducks [patos] of all kinds, snow geese [ansares], cranes [grullas], and many, many other strange birds, many bears [osos], wolves [lobos], light brown lions [cougars—leones pardos], tigers [tigres], bobcats [gatos cervales],38 [and] birds of prey [abes de rapiña], which are what is most noted in that land—large, small, of all kinds. Next to that town which is called Canos, twelve leagues from it, is a bald hill [cerro pelado]39 where they found many large, white, pointed stones of great hardness, which, first enduring a hammer on an anvil without breaking, were placed under the steel hammer, which would break anything. Some say they are diamonds. [There have also been found] many other colored stones, and large pieces.40 The soldiers of Juan Pardo brought some back, such as one Sergeant Moyano, who was in the interior one year and brought back many, among which an interpreter brought back one that was given to him by a cacique off the main route [desbiado] where the company went, which in whichever lodging [aposento] that they took, at night it gave off as much light as was needed for all of them. This Indian carries it to see himself and shave. [This interpreter] said that near there, there was a hill that was [made] of them, save that they were large pieces.41 In what Juan Pardo traveled, mines of silver and gold were discovered, because the adelantado sent seven miners who found better mines than those of Guadalcanal, but because of the interests of some with others, and envy, they never did anything, because they passed many mines. Between Juan Pardo and Estebano de las Alas

Recollections of the Juan Pardo Expeditions · 71

and the adelantado, it would be very tedious to make a relation of them, only it is said that being ready to settle and discover, it was their laziness, or that God does not hold it for them, since they lost so many good occasions. They find the gold in a river next to some towns that Moyano, the sergeant of Juan Pardo, destroyed, which are called Maniatique and Guapere, and from these two towns there are Indian women in the fort of St. Augustine, now Christians married with Spaniards, who relate that there is gold in the river in great pieces and in many forms—long, extended [tiendos], and squared—and from some little pieces they make chagualas, although they say the gold is of lesser [quality], of what has been seen. These [towns] are at the foot of the mountain range. They have towns encircled with a wall of cane [caña], grass [yerba], mud [barro], and very strong wood, and some make their houses below the ground. They are very indoctrinable [and] sharp with reason. There was an Indian in a town that is called Juada [Joara] who in four months spoke all our language and served as [an interpreter] with the Indians. If Juan Pardo had not thrown him out, a cleric who entered with Juan Pardo, who was called Sebastian Montero, would have alone converted all those Indians, because in a town that was called Guatari he stayed six months alone with a nephew of his, understanding us until some Indians took on our language. The cacique was a woman cacica, [and] she had two sons—very strong men—who did not veer one point from what the Father ordered them. They only lacked the action of the Holy Spirit. They honored the holidays of fasting [fiestas vexilias] like the cleric himself, knowing and understanding the days [f. 318r] of the week as if they had grown up all their lives with Spaniards. Since they were mining so much in that time, Juan Pardo [became] envious, seeing that they respected the cleric as much as him, because they carried Juan Pardo on a litter [en unas andas] when he traveled, and they did the same to the cleric. He said that he should not be equal with him, and seeing that all of the Indians loved and praised [Montero] and that they said that he was an interpreter of the affairs of heaven and held him in great reverence, Juan Pardo therefore sent a letter that if he found [Montero] there, he would hang him [le abia de colgar], and thus the said cleric came to the forts and left the Indians. This envy has thrown this land to the ground with the trees that Your Majesty sent there, as well as the people. If Your Majesty were to settle [Spaniards] there in that town as I have said, or with more people, spending in a short time what is spent, Your Majesty would save it in a great hurry. I advise Your Majesty of one thing—what needs to be spent should be in brief time, putting in livestock all at once. It seems to me that it would be less

72 · John E. Worth

expensive to send them through New Spain, in order to be a much larger number, like horses and some cows, until discovering the road, and afterwards [the settlers] should look for them in the land, feeding them a certain time, and then they would sustain themselves. [Your Majesty should] do them favors in the land, and not give them money. I say repartimientos, cattle, and liberties, and not things that are valued to place in the pocket and carry it to Castilla, depopulating Your Majesty’s land and exhausting it, and [they should] settle in their [lands], making perpetual ownership [juros] in them, buying many incomes [rentas].

León’s Account and the Route of the Pardo Expeditions The exact route followed by Juan Pardo was the subject of speculation for many years, in part because Pardo’s soldiers revisited several of the same towns initially contacted by the more well-known Hernando de Soto expedition in 1540. As León noted regarding Soto’s earlier route in this region, “Captain Juan Pardo went by the same road.”42 Indeed, the reinterpretation of Pardo’s route during the 1980s by Charles Hudson and his colleagues using the previously little-known long Bandera relation provided researchers with the clues that led to a similar reevaluation of Soto’s route (DePratter et al. 1983; Hudson 1990, 1997; Hudson et al. 1984).43 What makes the route followed by Juan Pardo particularly important for any reconstruction of the important Hernando de Soto expedition is the fact that this section of the 1540 route represents the northeastern “corner” of Soto’s explorations, marking the point at which Soto’s army turned to the west after having pushed northeastward to reach the famed province of Cofitachequi. In this sense, then, the locations of specific towns along Juan Pardo’s route provide a sort of benchmark for the entire Hernando de Soto route. Perhaps the most important and controversial town location for this section of both routes is that of Cofitachequi, or Canos (as it was known during the Pardo expeditions), which was located along a well-populated river corridor leading to the mountains. A review of literature on the subject reveals that two major but widely differing conclusions emerged during the past century. Early scholars, including anthropologist John R. Swanton and the United States De Soto Expedition Commission of the 1930s, located Cofitachequi on the middle Savannah River at the South Carolina/Georgia border, at a site known as Silver Bluff (Swanton 1939: 180–185). Half a century later, Charles Hudson and his colleagues, using considerably updated archaeological information, proposed a new location along the Wateree

Recollections of the Juan Pardo Expeditions · 73

River in central South Carolina, at the Mulberry Mound site [after Baker 1974]. The details of the vigorous debate regarding this town’s location are beyond the scope of this chapter (see chapter 1); however, the 1980s De Soto Trail Commission adopted the newer reconstruction of the Soto/Pardo routes, in part because of an emerging consensus among the majority of practicing archaeologists in these regions (National Park Service 1990). In this connection, the Domingo de León account provides substantial support for the Wateree River location of Cofitachequi, and thus also for Charles Hudson’s overall reconstruction of Juan Pardo’s route. While the account does not provide the level of geographic detail necessary to trace the precise route followed by the Pardo expeditions, León’s comparatively detailed overview of the physical and social geography of the deep interior of Spanish Florida upholds the conclusion that Pardo’s route followed the Santee/Wateree/Catawba River drainage, not that of the Savannah River. A comparison of the León account with modern maps and with the distribution of sixteenth-century aboriginal occupation in South Carolina and Georgia leaves little doubt that Hudson’s route is correct. Specifically, using León’s written account, we can reconstruct his “mental map” of the rivers and their directions with respect to one another, as well as place Indian towns and chiefdoms (some named) along these rivers. As is evident from the text of the account above (portrayed in figure 3.1), León perceived the river valley occupied by Canos/Cofitachequi as one with at least two major branches. His descriptions indicate that two rivers joined some 30 leagues from the coast (just over 100 miles, following Hudson’s use of the 3.45 long league for the Pardo expedition) and that one of these two branches flowed from the west and the other flowed from the northwest. Both of these rivers were said to extend to the mountains. Furthermore, the western branch was evidently unpopulated from its junction with the other branch to a distance of 40 leagues upriver toward its source at the foot of the mountains, whereas the northwestern branch appears to have been well populated with towns, under the rule of two major Indian caciques. One of these was Canos, situated 12 leagues (or just over 40 miles) upriver from a town at the junction of these two branches, and the other was simply noted to be located far upriver to the northwest, at the foot of the mountains. Importantly, León furthermore states that Canos was “forty leagues from Santa Elena to the north,” fixing the location of the entire river valley described by León (the upper branches of which are noted to drain from the northwest and west) north from Santa Elena, apparently emptying into the Atlantic somewhat to the northeast of Santa Elena along the coast (based

74 · John E. Worth

Figure 3.1. Domingo de León’s “mental map” of La Florida’s physical and social geography.

on the distances mentioned by León). This notation by León represents an effective demonstration that the Pardo expeditions did indeed depart in a northerly direction from Santa Elena, not to the southwest, as argued by Lewis Larson (1990). A quick comparison of figure 3.1 with the modern map of South Carolina in figure 1.3 reveals that there is only one major river valley that fits all of León’s geographic descriptions: the Santee River drainage (including tributary rivers such as the Wateree, Congaree, Saluda, and Catawba). The Savannah River valley can be ruled out immediately, since it has no division between two major branches some 80 miles inland and since only this main channel reaches to the mountains. In contrast, the Santee River is formed when the substantial Wateree and Congaree rivers join just under

Recollections of the Juan Pardo Expeditions · 75

100 miles inland, and both of these rivers ultimately drain from the Appalachian highlands. Indeed, apart from the mythical “lake” said to be the source of the western branch (which neither León nor any other Spaniard ever actually claimed to have seen), the visual similarity between León’s “mental map” and the actual course of the rivers making up the Santee drainage is quite striking, particularly when compared with other nearby rivers, including the Savannah. In addition to these geographic similarities, León’s account fits quite well with the social landscape of the Santee River drainage. Archaeological evidence for the distribution of late sixteenth-century aboriginal occupation in central South Carolina indicates that the lower Wateree River valley was well populated during the late prehistoric and early contact period, whereas the Congaree, Broad, and Saluda rivers lacked the comparatively substantial occupation witnessed to their east, evidently having been largely abandoned during the fifteenth century. Furthermore, the cluster of prehistoric mound centers situated along the Fall Line zone of the Wateree River certainly reflects a major and long-lasting sociopolitical entity focused a short distance upriver from the junction of the Congaree and the Wateree (see below), corresponding well to the description of Canos/Cofitachequi. In addition, this location is just under 150 miles, or roughly 40 long leagues, and almost precisely due north of Santa Elena on Parris Island, fitting Domingo de León’s account exactly. In sum, the Domingo de León account provides convincing support for the Juan Pardo route posited by Charles Hudson and colleagues and thus for their reconstruction of the Hernando de Soto route through the Carolinas and Tennessee (León specifically notes that Pardo followed the same route). As discussed above, it seems quite clear that León’s detailed overview of the river valley occupied by Canos/Cofitachequi simply cannot be applied to the Savannah River (previously thought to be the location of Cofitachequi) but instead describes the rivers constituting the upper Santee watershed. Furthermore, he was clearly in a position to know about Pardo’s route through the interior by both firsthand and secondhand information (not only did he participate in the expeditions, but also he had a notable understanding of Indian languages and customs and communicated with the Indians on a regular basis). Consequently, unless the Domingo de León account is rejected outright, there should remain little doubt regarding the general course of Pardo and Soto through the Carolinas. While the León account provides few details regarding the specific locations of individual Indian towns along this course, and does not even mention Pardo’s excur-

76 · John E. Worth

sion across the Appalachians toward Coosa, what is contained within the account is sufficient to fix the location of the elusive Canos/Cofitachequi near the Fall Line of the Wateree River, exactly as concluded by Charles Hudson and his colleagues. The more recent archaeological discovery and identification of the Berry site as the location of Fort San Juan at the pivotal town of Joara provides further confirmation of this conclusion.

Notes 1. All of these sources have been published in both transcribed and translated form by Charles Hudson (1990) as The Juan Pardo Expeditions. Although I have utilized Paul Hoffman’s translations (pp. 205–321) for passages quoted in this chapter, their original sources are as follows: Juan Pardo, relation, no date, Archivo General de Indias (hereinafter abbreviated as “AGI”) Patronato (hereinafter abbreviated as “PAT”) 19, Ramo 22, no. 1; Francisco Martínez, relation, July 11, 1567, Santa Elena, AGI PAT 19, Ramo 22, no. 2; Juan de la Bandera, “Memorial of the Villages and What [Kind] of Land Is Each Village of Those in the Provinces of Florida, through Which Captain Juan Pardo Entered to Discover the Road for New Spain from the Point of Santa Elena of the Said Provinces, the Years of Fifteen Sixty-Six and Fifteen Sixty-Seven,” January 23, 1569, Santa Elena, AGI PAT 19, Ramo 20 (hereinafter Bandera, “Memorial,” 1569); Juan de la Bandera, “Autos for the Relation That Juan Pardo Gave of the Entrance He Made in the Land of Florida,” April 1, 1569, Santa Elena, AGI Santo Domingo (hereinafter abbreviated as “SD”) 224. 2. The testimonies of these three witnesses appear in the transcript of an investigation by Gonzalo Méndez de Canço, “Relation of La Tama and Its Land and of the Settlement of Englishmen,” February 4–6, 1600, St. Augustine, AGI SD 224, Ramo 5, no. 36. Hudson (1990) discusses this testimony, although complete translations are not provided. Teresa Martín and her daughter, Inés Martín, received Juan’s final pay as a squad leader in Santa Elena on August 17, 1577, following his death on July 22, 1576, when “the Indians killed him,” as related in the accounts of Tomás Bernaldo de Quirós, AGI Contaduría (hereinafter abbreviated as “CD”) 944. 3. Domingo González de León, letter to the Crown, October 13, 1584, Madrid, AGI SD 231 (a selection of this letter [hereinafter León, relation, 1584] has been translated in this chapter). 4. Although the name “Domingo González de León” would normally be abbreviated as “Domingo González,” or simply “González,” in the Spanish naming system, Domingo was evidently known as “Domingo de León” by the soldiers serving with him (all the witnesses he called for his 1579 and 1584 service summaries [see later notes] referred to him as such in their testimony, even though the formal petition was signed “Domingo González, native of León”), and he was also mentioned as such in the “long” Bandera relation. Most importantly, the original 1584 letter translated in this chapter was signed “Domingo de León.” Consequently, I have chosen to use the abbreviated form “León” when referring to the subject of this chapter.

Recollections of the Juan Pardo Expeditions · 77

5. This and subsequent details of Domingo de León’s service are principally drawn from documentation in his 1586 petition for a captain’s title, which included transcripts of earlier testimony dating to 1579 and 1584. Domingo González de León, petition to the Crown, April 22, 1586, Madrid, AGI SD 14, no. 33 (hereinafter León, petition, 1586). Details of León’s service record that are generally repeated throughout this document reference the 1586 service record, but specific points brought up in individual letters, certifications, or testimony reference the particular document involved. For more information relative to the establishment of St. Augustine and Santa Elena and the earliest years of the Spanish colony of Florida, see Lyon 1976. 6. León, petition, 1586. 7. Gonzalo Vicente, testimony, September 13, 1579, Santa Elena, transcribed in León, petition, 1586. Vicente was in 1579 a thirty-two-year-old soldier in St. Augustine and would later lead Pedro Menéndez Márquez’s exploratory venture to the Chesapeake Bay in 1587. 8. León, petition, 1586. León also described his service during the Pardo expedition in his 1584 relation (translated in this chapter). 9. Pedro Luís, testimony, September 12, 1579, Santa Elena, transcribed in León, petition, 1586. 10. Bandera, “Memorial,” 1569 (translation by Paul Hoffman in Hudson 1990: 293). León mentioned such activities in his 1584 relation (see this chapter’s translation); in 1586, León’s only specific reference to his duties during the Pardo expeditions was that he was sent “looking for food among the said Indians,” confirming that he probably served as an interpreter during food-gathering activities along the route of the expedition. León, petition, 1586. 11. The fact that León does not appear on the list of soldiers who received sandals and shoes in Chiaha/Olameco before turning back toward Santa Elena might support this conclusion (although not all soldiers who accompanied Pardo across the mountains were included in this list). See Juan de la Bandera, list of soldiers who received sandals and shoes in Olameco/Chiaha, April 21, 1568, Santa Elena, AGI Contratación, R. 7, no. 2, translated by Paul Hoffman in Hudson 1990: 334–336. 12. Hudson (1990: 152–153) suggests that fewer than twenty men may have returned with Juan Pardo to Santa Elena in March of 1568. How many of the more than one hundred soldiers left in the interior survived the general uprising that followed almost immediately is unclear. The number seems to have been quite small, and thus Domingo de León may have been one of only a handful of soldiers who literally lived to tell about the Pardo expeditions. This number, of course, does not count those soldiers who are known or suspected to have remained living in the interior after the uprising. 13. Pedro Menéndez de Avilés, title granted to Domingo González de León, August 4, 1572, Santa Elena, transcribed in León, petition, 1586. Two days previously, León’s name appeared on a list of married soldiers in Santa Elena compiled by notary Domingo Hernández de la Havana. List of married people in Santa Elena, August 2, 1572, Santa Elena, AGI Escribanía de Cámara 1024A. I am grateful to Eugene Lyon and Dian Hartley of the Center for Historic Research at Flagler College for providing a summary of this last reference, along with several other related references from their computer database. The name of León’s wife is found in the record for a posthumous pay disbursement to her on October 28, 1590, St. Augustine, AGI CD 942.

78 · John E. Worth

14. Diego de Velasco, title granted to Domingo González de León, January 22, 1575, Santa Elena, and Diego Moreno, title granted to Domingo González de León, January 28, 1575, Santa Elena, both transcribed in León, petition, 1586. 15. In his service record, León stated that Velasco had him forcibly replaced as notary public by Gonzalo Pérez in 1575. Domingo de León, questions for the witnesses (hereinafter León, questions, 1579), September 12, 1579, Santa Elena, in León, petition, 1586. 16. Hernando de Miranda, investigation relative to the petition of the settlers, February–March 1576, Santa Elena, AGI SD 231 (transcription, Connor 1925: 146). León later hinted that his trip to Spain was prompted in part by having been “fired.” León, questions, 1579. 17. Hernando de Miranda, certification of the service of Domingo González de León (hereinafter Miranda, certification, 1584), December 4, 1584, Madrid, transcribed in León, petition, 1586. 18. Hernando de Miranda, license to Domingo González de León, May 14, 1576, Havana, transcribed in León, petition, 1586. 19. León later implied that the rebellion might not have occurred had he been in Santa Elena at the time. In 1584, León commented that the internal political squabbling among the Spaniards gave the Indians the opportunity they needed for rebellion, noting that “regarding this subject the past [officials] were lost, and they wished to kill each other, as was proposed when Don Diego Belasco and Pedro Men[én]dez and Otalora [were in power]. From here began the discord, and upon the Indians understanding the division that there was among [the Spaniards], they took advantage of it.” León, relation, 1584. 20. León, relation, 1584. This account of León’s 1576 trip to Spain, along with his numerous complaints relative to problems under the governorship of Pedro Menéndez Márquez, appears in portions of his 1584 relation that were not translated for this article. 21. León, petition, 1586. 22. Alcaldes of the Royal Audiencia of New Spain, provision respecting Domingo González de León, April 1, 1579, Mexico, transcribed in León, petition, 1586. 23. Alcaldes of the Royal Audiencia of New Spain, auto and certifications regarding the release of Domingo González de León, April 7–15, 1579, Mexico, transcribed in León, petition, 1586. 24. Juan Mel, certification regarding the arrival of Domingo González de León, September 15, 1579, Santa Elena, transcribed in León, petition, 1586. 25. León, petition, 1586; see also Pedro Menéndez Márquez, letter to the Crown, January 3, 1580, St. Augustine, AGI SD 224, R. 2, no. 14; Antonio Martínez Carvajal, letter to the Crown, November 3, 1579, Havana, AGI SD 125. 26. Juan Miel (?), certification regarding peace with the cacique of Coçapue, October 15, 1579, Santa Elena, transcribed in Tomás Bernaldo de Quirós, petition, June 20, 1584 (date of ruling by Council of the Indies), Madrid, AGI SD 125, translation by Eugene Lyon. At that time (1579), Domingo de León seems to have been one of the two most important interpreters in Santa Elena. León evidently served for the Indians of the province of Orista (around Santa Elena), while Alonso Díaz de Sevilla interpreted for the Guale province to the south. 27. Domingo González de León, petition to Pedro Menéndez Márquez, September 15, 1579, Santa Elena, and Domingo González de León, petition to Pedro Menéndez Márquez,

Recollections of the Juan Pardo Expeditions · 79

January 23, 1581, Santa Elena, both transcribed in León, petition, 1586. León’s debt of 350 pesos was nearly eight times his yearly salary as a soldier in Florida (only about 44 pesos). 28. León, relation, 1584. 29. King Philip II, cedula to royal officials of Florida, March 16, 1586, Aranjuez, AGI SD 2528; King Philip II, cedula to viceroy and audiencia of New Spain, March 31, 1586, San Lorenzo, AGI SD 2528; King Philip II, cedula to governor of Florida, March 31, 1586, San Lorenzo, AGI SD 2528; King Philip II, cedula to the Casa de Contratación, April 13, 1586, San Lorenzo; King Philip II, cedula to royal officials of Florida, April 13, 1586, San Lorenzo, AGI SD 2528. Summaries provided by Eugene Lyon. 30. On October 28, 1590, a voucher was issued in St. Augustine for 117,667 maravedís (equivalent to 411 pesos, or 299 ducados), amounting to just under five years’ salary (April 10, 1579–April 6, 1584) at 200 pesos annually, minus daily rations and supplementary clothing valued at 588 pesos. The salary was calculated based on the standard pay in the audiencia of Mexico for a naguatato, or interpreter, and was comparable to that known to have been paid to other interpreters in Santa Elena during this period. AGI CD 942. 31. Miranda, certification, 1584. 32. León, questions, 1579. 33. Domingo de León, letter to the Crown, March 19, 1586, Madrid, transcribed in León, petition, 1586. 34. Pedro Luís, testimony, September 12, 1579, Santa Elena, transcribed in León, petition, 1586. 35. The word in the manuscript here is virtually illegible, but it appears to be le mato, based on the similarity of the final two letters to their subsequent appearance in ciento. The entire passage reads “pifano de Ju.o pardo que quedo de la gerra de quando le mato los ciento y cincuenta sol.os.” Although the grammar is incorrect (as in other parts of León’s account), the context of this passage seems clear based on what is known about the loss of Pardo’s interior forts. 36. This passage is extremely confusing, and could be transcribed and translated in several ways. The original manuscript appears to read “por lo que yo e andado con ellos allo los ciertos y vdaderos lo propio bi de nunapa lo que yo andube al natural.” Because of the difficulty of León’s handwriting and writing style in this relation, the adjacent letters bidenunapa may be his attempt to write bi de una mapa using abbreviated words. The context makes this the most likely solution, especially given his reference to measurements below. 37. The original text of this extremely difficult passage appears to read “muchas palomas que no se puede encimerar que aconte aun en la costa tardar una manade dellas una hora lapasar a tierra calien y en busca de bellota.” 38. The 1726 Diccionario de autoridades defines the adjective cerval as “a thing that pertains to the deer [ciervo] (from which it derives) or is similar to it in something.” Since León already mentioned “lions” above, and since this probably refers to the long-tailed cougar (the reference to tigers is presumably based on rumor or the misidentification of cougars), it seems possible that this other reference to cats using the “deer-like” adjective defined above is his description of the bobcat, which does indeed possess a short tail like that of the deer. 39. The location of León’s “bald hill” is not clear from his account, but he may have been referring to the crystal mine registered on December 10, 1567, downriver from Ysa, which

80 · John E. Worth

Charles Hudson locates just over 20 leagues north of Canos/Cofitachequi. This mine was said to contain unusual crystals, which were “sharp and faceted.” Bandera, Autos, translation by Paul Hoffman in Hudson 1990: 282–283. 40. Importantly, although León cites others as believing that these stones were diamonds, he makes no conclusion. His descriptions indicate that these stones were quartz crystals or perhaps corundum. Hudson (1990: 156–164) discusses the other documentary references to these “crystal mines” discovered by Juan Pardo and his soldiers. 41. Citing his Indian informant, León here appears to be describing a massive outcrop of the same crystalline material identified at the crystal “mines” actually visited by Pardo’s soldiers (presumably quartz). The Spaniards do not seem to have visited this “hill” of solid crystal, but this almost certainly represents the source of the later legend of Los Diamantes, a mountain of diamonds said to be located deep in the northern interior. León seems to make a clear distinction between this legendary crystalline hill (which was described to him by the same Indian informant who claimed that he owned a “glowing” crystal) and the crystal mines that were actually visited by the members of the Pardo expeditions. The fact that these two sources of information (first- and secondhand) were later combined into the legend of a single, mythical mountain called Los Diamantes reveals the limitations of oral tradition for precise historical interpretation. 42. In fact, Indians along Pardo’s route remembered the earlier Soto expedition, although their information was frequently exaggerated or erroneous. 43. Beck (1997b) later adjusted portions of the Hudson route to reflect the identification of the Berry site as Joara.

III Where They Lived Household Archaeology at Fort San Juan

December 27, the day of Saint John the Evangelist, was a day of celebration at Joara, a place that the Christians would thereafter know as Cuenca, and on this late afternoon their camp at the edge of the town was filled with laughter, talk, and commotion. A fort, called San Juan in honor of the day, would be laid out the next morning, as would structures for housing those men left behind to hold this ground for Spain. While native women prepared great pots of strong-smelling stew over low fires in the Christians’ camp, the Indian men roasted chunks of bear meat nearby. Guillaume Rouffi—Guillermo Ruffin now that he served the Spanish Crown—leaned against an oak where the plain dropped down to the stream, content to watch and listen. His skill as a translator had all but ensured him a place on the expedition, and for most of this day he had proved his worth while the captain and the mico held council for many long hours. From where he stood by the oak he could see the captain showing one of his officers (was it Escudero?) how to arrange the gifts he planned to present during the feast: swaths of red taffeta, a few chisels, and strings of deep blue, clear, and turquoise glass beads. Just then the mico and his principal men came in view of the camp, and all of those within—Indians and Christians alike—grew quiet as they approached. Ruffin started back to the gathering. This day, he thought, this victory for God, would not soon be forgotten.

*

*

*

Colonial encounters are profoundly spatial processes; the actions that shaped, constrained, and transformed social life during such entanglements were in-

82 · Part III. Where They Lived: Household Archaeology at Fort San Juan

extricably bound to the specificities of place. As Henri Lefebvre (1991: 73) aptly writes, “Itself the outcome of past actions, social space is what permits fresh actions to occur, while suggesting others and prohibiting yet others.” We devote the second part of the book, “Where They Lived,” to a close study of the Spanish compound and its different architectural, spatial contexts. In chapter 4, we discuss excavations in two of the five burned buildings identified within the compound, Structures 1 and 5. Surprisingly, our work revealed that these two buildings were constructed to different architectural grammars. Structure 1 was built in a style typical of Late Mississippian buildings across the southern Appalachians. This was not unexpected, since the Spanish accounts specify that the people of Joara built such houses for Pardo’s men. Structure 5, in contrast, appears to have been a much more expedient—if not flimsy—construction. A comparison of these two buildings, along with stratigraphic data from associated pit features and analyses of foodways, suggests temporal variation within the architecture of this encounter. Some of the buildings, including Structure 1, were constructed at the beginning of the fort’s occupation, when relations between the Spaniards and their native hosts were most amicable. Later, perhaps before the second winter, more buildings were added, one of which was Structure 5. The expedience of its construction indicates that it was put up quickly, and by people with much less knowledge and skill than the builders of Structure 1. In addition to the structures, we have excavated more than a dozen features here, many of which contain small quantities of Spanish artifacts along with native-made Burke ceramics, lithic debris, and plant and animal food remains. The stratigraphic relationships between these features and the burned buildings, together with other lines of evidence, allow us to identify two temporal stages in the use history of Fort San Juan. Specifically, Structure 5 and at least one other building on the west side of the compound were built atop several features that were probably used for the preparation of daub for Structure 1 and other buildings on the east side of the compound. We can thus identify an early fort component associated with the features under the western buildings and a later component associated with the group of features just west of Structure 5. The significance of this interpretation is especially clear in the analysis of food preparation and consumption in the Spanish compound, to which we turn in part 4 of this volume. In chapter 5, “Wood Selection and Technology in Structures 1 and 5,” Lee Ann Newsom offers a detailed study of the remarkably well-preserved woods recovered from Structures 1 and 5. Her careful analysis sheds unique light on the selection of different woods used in these buildings, on the technology

Part III. Where They Lived: Household Archaeology at Fort San Juan · 83

and seasonality of construction, and on the interactions between native peoples and soldiers in the cultural context of house construction. Among the most intriguing of her findings is that many of the wooden architectural elements in both buildings were procured and prepared with metal tools such as axes, chisels, and saws, corroborating the documentary accounts that specifically mention these carpentry tools among the expedition’s equipment. Moreover, her comparison of wood and other organic remains recovered from these buildings both confirms the expedient construction of Structure 5 and suggests that some details of its architectural grammar—the use of wood shakes or shingles as wall siding, for example—derive from a European carpentry tradition.

This page intentionally left blank

4 The Built Environment of the Berry Site Spanish Compound Robin A. Beck, David G. Moore, Christopher B. Rodning, Sarah Sherwood, and Elizabeth T. Horton

Written accounts of the Juan Pardo expeditions, as essential as they are for reconstructing both Pardo’s route across the Carolina Piedmont and the political geography of its native peoples, tell us almost nothing about daily life at Fort San Juan. To understand how daily practice shaped the contours of this colonial encounter, as well as how these contours changed over the course of eighteen months, we must turn to archaeology. Fortunately, the rich and extraordinarily well-preserved archaeological contexts at the Berry site allow us to develop a nuanced account of the interactions between Pardo’s men and the people of Joara. Throughout this book, we draw on a diverse range of material evidence to develop our account, but we start here with a detailed analysis of the built environment, focusing on that part of the Berry site that we identify as the Spanish compound. It was here that men and women from Joara and other native communities engaged in daily, routine interactions with Pardo’s garrison, and it is this place that anchors our story. The Spanish compound at the Berry site covers approximately 0.3 hectares (0.75 acres), of which we have excavated more than 1,200 square meters to date. The compound (figure 4.1) contains five burned buildings— designated Structures 1 through 5 (figure 4.2)—dozens of pit features, and hundreds of postholes. Our excavations of Structures 1 and 5 revealed that all of the burned buildings are square, single-set post constructions with rounded corners and central hearths; four of the five, Structures 1 through 4, were built in semisubterranean basins. Structure 1, excavated in 2008, covers about 56 square meters and measures 7.5 meters on a side (excluding the basin). A well-defined entryway with exterior wall trenches was at the

86 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.1. Plan map of the Berry site Spanish compound, with all excavated features numbered.

west corner of the building. Structure 5 was slightly smaller, about 7 meters on a side (with a total area of 49 square meters), and was built either in a very shallow basin or directly on the ground surface. All five buildings were in use for a relatively short period of time, and none has evidence of any rebuilding after the blaze that consumed them all. Most of our fieldwork at Berry since 2001 has focused on the remains of the Spanish compound, and in this chapter we offer a descriptive and interpretive synthesis of its spatial organization, tracking changes to its spatial layout from the time of its initial construction in late 1566 to that of its fiery end.

The Built Environment of the Berry Site Spanish Compound · 87

Figure 4.2. Structure 3 after plow zone removal, 2003 (looking north).

Excavations in the Spanish Compound Excavations of household contexts in the Spanish compound have yielded a diverse range of archaeological evidence about the construction of its buildings, the kinds of activities that took place inside them, and the kinds of activities that took place in the spaces between and around the buildings. Our excavation methods constitute a modified form of archaeological practice that has proven successful during research conducted by the University of North Carolina’s Cherokee and Siouan Projects, based, respectively, in western North Carolina and the central North Carolina and Virginia Piedmonts (e.g., Dickens 1976, 1978; Dickens et al. 1987; Keel 1976; Ward and Davis 1993, 1999). One critical component of our approach to excavations at Berry is broad horizontal exposure of large areas to reveal the site layout by stripping the plow zone with shovels and dry-screening sediments through quarter-inch hardware mesh. Another key component of our field strategy is the hand-tool excavation of intact architectural debris and floor deposits, as well as the hand-tool excavation of pit features inside and outside of buildings, and the water-screening of all sediment deposits from

88 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

architectural and feature contexts. Four teams of archaeologists, working simultaneously, conducted excavations inside structures. Excavation teams collected at least one 10 liter bulk flotation soil sample for every excavation locus in the structures and features; wood and plant fiber specimens were also collected for analysis. We excavated within natural stratigraphic zones in all structures and features using 1 × 1 meter units in the structures and sections in features to control for provenience and mapping. In parts of the compound between the structures, we made broad horizontal exposures in contiguous 2 × 2 meter and 3 × 3 meter units, dry-screening the plow zone in these squares and mapping the location of all features and postholes. We made detailed plan maps, profiles, and photographs of all excavated structures and features, and standardized forms were completed for all contexts. These forms included data on soil characteristics, elevation, plan maps, artifact density and types, and related variables. One of our essential recording techniques within the burned buildings has been creating photomosaics of architectural remains as excavation proceeds; photomosaics were compiled by taking overhead photographs of all 1 × 1 meter units in each stratigraphic zone, then stitching the photographs together. We then made detailed, digital plan maps of all organic materials in situ in the buildings by drawing directly from the photomosaics. Given the unusual preservation of organic remains in the burned buildings, we conducted on-site microexcavation of particular architectural contexts. At the close of our fieldwork, we backfilled and stabilized open areas. All artifacts were bagged by context in the field, then cleaned, sorted, catalogued by type, and permanently curated at the Warren Wilson College Archaeology Laboratory. All organic remains recovered from excavation contexts in the compound—including carbonized and uncarbonized wood, cane, and flotation samples—have been curated in a similar manner prior to processing. Excavation of Structures 1 and 5 In 1997, Tom Hargrove conducted a gradiometer survey of the northern 0.5 hectares of the Berry site, where olive jar, majolica, and wrought iron nails had been recovered during previous surface reconnaissance. Hargrove identified several large, prominent magnetic anomalies that seemed to represent burned structures, as well as many smaller anomalies that seemed to have the shape and size of pit features. Subsequent auger testing of four of the larger anomalies demonstrated that all were burned, semisubterranean buildings with intact deposits up to 30 centimeters deep. These anomalies

The Built Environment of the Berry Site Spanish Compound · 89

were labeled Structures 1 through 4. Excavations in 2002 located a fifth burned building, labeled Structure 5, in an area not surveyed by the 1997 gradiometry. There was little to suggest that this structure was built in a semisubterranean basin, and its architectural deposits appeared to be much less substantial than those of the other four buildings. We conducted test excavations in Structure 1 in 2003 and 2004, both to learn the extent of its intact deposits and to devise a strategy for more extensive work in this and the other burned buildings. These exposures (figure 4.3) revealed that the Structure 1 basin preserved a stunning array of carbonized architectural elements, including upright wall and roof sup-

Figure 4.3. Excavation units in Structure 1, 2003, with preserved organic materials exposed (looking north).

Figure 4.4. Excavations in Structure 1, 2003: a, overhead view of notched timbers and chestnut plank, OG#2003a; b, detail of chestnut plank, OG#2003a.

The Built Environment of the Berry Site Spanish Compound · 91

Figure 4.5. Structure 5 after plow zone removal, 2007 (looking north).

port posts, timbers, thatch, and strips of bark. All of these organic remains had been sealed in the basin by sediments from the walls and roof when the house burned and collapsed. The aim of these exploratory trenches was to expose organic materials in situ rather than to excavate through the house floor to subsoil. Nevertheless, we did completely excavate a single 1 × 1 meter unit (Unit 39) near the entry (see figure 4.15) and remove specific architectural elements from two other units (figure 4.15, Units 18 and 28). These elements include a short piece of a notched timber (figure 4.4a) and part of a substantial plank (figure 4.4a,b). Intact deposits ranged in depth from 5–10 centimeters in the north corner of Structure 1 to nearly 30 centimeters in the south part of the building. Test excavations in Structure 1 also offered us important insights into the stratigraphy of the building (see later discussion). Using the limited test excavations in Structure 1 as a guide, we conducted full data recovery of Structures 1 and 5 in 2007 and 2008. Yet rather than start with the deep and complex deposits of Structure 1, we decided to begin with Structure 5, one of two buildings on the western side of the compound. This building’s slightly smaller size and its comparatively shallow deposits made it an intriguing contrast with the other buildings (which coring has confirmed were all built in basins), as well as a less complex feature in which to further develop our excavation methods and recording techniques. We began by exposing Structure 5 in its entirety (figure 4.5)

92 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.6. Structure 5 excavation units.

and then establishing an excavation grid aligned to the sides of the building itself (figure 4.6). Plow scars made it difficult in some cases to differentiate plow-disturbed dirt from intact debris (figure 4.7), but exposing broad areas enhanced our ability to follow plow scars while troweling the surface of the undisturbed architectural deposits. Excavations began along the southern side of the structure and continued north in rows of 1 × 1 meter units across the extent of the building. We covered structure excavation areas with large canopy tents and, when necessary, with sheets draped from the sides of these pole-frame tents. This covering protected organic materials from exposure to direct sunlight and created favorable lighting conditions both for differentiating color and tex-

The Built Environment of the Berry Site Spanish Compound · 93

Figure 4.7. Excavating plow scars over Structure 5, 2007.

ture during the excavation and for consistency in photography. Exposing contiguous units enhanced our ability to detect and to interpret spatial patterns and stratigraphic profiles. On the basis of their size, specimen type, and structural integrity, we assigned particular timbers and other organic remains a unique number for which we used the designation “OG.” Specimens given OG numbers were removed, wrapped in archival-quality foam, and packed in cardboard boxes for transport to the laboratory. We removed many timbers, posts, cane bundles, and other specimens intact, but in several cases we needed to remove samples in sections. Structure 5 There were five main zone designations for Structure 5. The Zone 1 designation was used only at the beginning of excavations on the south edge of the building, at which time we believed that Structure 5 might have been built in a shallow basin. Previous excavations in Structure 1 had identified a zone of relatively clean, homogenous fill between the edge of the building’s basin cut and the edge of its carbonized architectural remains (see figures 4.16 and 4.17). Designated Zone 1, these sediments represent constructionrelated fill thrown into the basin after Structure 1 was built, filling the narrow space between its completed outer walls and its basin cut. Once we recognized that no such zone existed along the outer edge of Structure 5

94 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

(that is, that it was not built in a basin), we discontinued use of the Zone 1 designation in this building. Zone 2 is the layer of burned and highly fired sediment overlying the architectural remains in Structure 5 (see plate 2). This sediment likely derived from the structure walls and perhaps the roof and is the uppermost intact deposit in the building. Zone 3 (see plates 3 and 4) refers to the carbonized architectural elements of Structure 5, including all timbers and posts, that had collapsed atop Zone 4, the structure’s floor. This floor was largely composed of unmodified subsoil with little evidence of prepared floor coverings such as mats (Sherwood 2014). Finally, Zone 5 is a deposit of highly oxidized orange-red sediment on the south side of the structure (see figure 4.5 and plate 2). Removal of Zone 2 was often complicated both by the presence of large timbers and posts that extended across adjacent units and by complex concentrations of wood and rivercane directly on the floor of the structure. In units with the highest concentrations of cane, particularly in the southeastern corner of the structure, these materials were exposed, mapped, and photographed (figure 4.8). In the next chapter, Newsom offers a highly detailed and comprehensive analysis of wood and other organic architectural debris from both burned buildings, but here it is pertinent to discuss the cane from Structure 5 analyzed by Elizabeth Horton. The rivercane (Arundinaria sp.) specimens recovered from Structure 5 were dominated by culms, or stems. Moreover, whole cane culms and cane fragments were extremely common in the fill of this building, particularly along its eastern wall. During the careful excavation of two areas with large numbers of intact, whole culms in the southeastern corner of the building (see plate 4), several “straps”—thin, flexible strips of cane used to bind whole stems together—were observed in and around the culms. This raised the possibility that the culms were the cane surfaces of interior benches, bound together with some sort of straps that interlaced between the culms and held them in place. Most of the specimens given OG numbers in the field were recovered from the adjacent Units 15 and 23. These two units had an extraordinary amount of intact cane culms. Sorting and analysis of the fragments, however, indicate an unusual profile that would not be expected either for cane artifacts or for benches. The culm diameter measurements were taken at the internode, at least 1 centimeter either below or above a node (cane nodes are always slightly larger in diameter than the internode culm). While average diameter was 1.1 centimeters, the range was from 0.5 centimeters to 2 centimeters. Because cane tapers in thickness from the ground to the upper branched and leafy

Figure 4.8. Rivercane (Arundinaria sp.) culms in situ in Structure 5: a, Unit 23; b, Unit 14.

96 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.9. Comparison of primary culm nodes in rivercane: a, primary culm nodes from Berry site Structure 5, showing unprocessed culms with sharp spikes; b, c, wellprocessed culm nodes from Ozark bluff shelters.

portions, this likely indicates the presence of entire culms, from the lower, thicker portions of the culms to the upper, thinner portions. Moreover, the small-diameter culm fragments with evidence of the distal portion of the branching complement (that is, of the leaf culms that grow out from nodes of the primary culm) may indicate the presence of minimal processing. Even more pointed evidence for the minimal processing of these culms is the frequency of the basal portion of the branching complement (figure 4.9a) at the primary culm nodes and the presence of primary culm node buds. These are not soft structures in the living plant. When leaf culms are torn off the primary culm, a hard, sharp spike will often be left behind at the basal area. If this spike is not removed, then running one’s hands down

The Built Environment of the Berry Site Spanish Compound · 97

the primary culm from top to bottom may cause torn and cut flesh on the hands. Thus, processing of the node is frequently observed in cane culm artifacts such as basketry, arrow and dart shafts, flutes, and more. This processing can include cutting off or shaving down the distal portion of the branching complement or smoothing, shaving, and sanding the bottom portion of the node and the internode lip below the node (figure 4.9b,c). In fact, Horton’s (2010) study of desiccated plant remains recovered from Ozark bluff shelters found that the removal of these sharp spikes distinguished rivercane that was debris from specimens that were fragmentary artifacts or in the process of production. The frequency of culms with this sharp spike present makes it clear that cane from Structure 5 was neither intensively nor extensively processed prior to use. Indeed, the frequency of these unaltered nodal attributes, beyond the removal of the branching complement, points to no more treatment than the leafy culms having been snapped off before the cane was used. The Structure 5 rivercane specimens are also noteworthy for the presence of what seem to be cane splint lashes. These curled and twisted cane splint pieces range in width from 5 to 8 millimeters and consist of the entire thickness of the culm wall, from the lignified outer epidermal layer to the interior of the culm wall, indicating a splint and not an intensively processed peel similar to those that are used for basketry (e.g., Horton 2010). This average width indicates the splitting of whole culms into eighths, roughly the last split performed in basketry production as well (see Hill 1997; Horton 2010). Several of these splints were recorded during excavation as having been around or over fragments of whole culms. Cane splint fragments were noted in seven of the thirteen specimens taken as OG samples. While most of the splints from Structure 5 were Arundinaria sp., a few appear to have been some kind of other plant material, possibly wood strips. In sum, the evidence of minimal processing, of entire—or nearly entire—lengths of primary culms with a range in diameters of 0.5–2 centimeters, of interlaced splint elements, and of the masses of cane concentrated in the southeastern corner of the building all suggest that the cane in Structure 5 represents some kind of internal architectural element for the building. While our initial thoughts in the field were that these were the remains of wall benches, the presence of such small-diameter cane suggests otherwise. These culms would likely not have been strong enough to support much substantial weight. Moreover, because nearly the entire lengths of culms seem to have been used, any such benches either would have been extremely wide—reaching nearly 5 feet into the interior of the building—or

98 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

would have had culms orientated lengthwise along the wall. Either circumstance would have created impractical and weak benches. Another possibility raised was the use of the lashed canes as some form of flooring in one or more areas along the southeastern corner and eastern wall. Again, though, the presence of the smaller-diameter portions of the primary culms, as well as the basal elements of the branching leaf culms, may have made this impractical. Crushing and cracking from the weight of a person walking over such a floor would have presented a very real problem, considering how thin some portions of the culms were. Walking barefooted or even in thin-soled shoes or sandals (such as we know Pardo’s men wore) across cane culms with the hard, sharp basal elements intact would not only have been painful but might also have caused injury. Finally, the cane culms would need to have been tightly secured to keep them from rolling in place. While the interworked splint elements do suggest that the culms were secured, this may not have produced a stable floor surface. Some of the cane from the in situ and larger samples have culms situated above and intermixed with small-diameter wood fragments, perhaps small poles; this would seem more likely the result of a lashed cane structure having fallen with wood elements of the building rather than being the surface onto which the burning building was falling. The most likely possibility is that the cane was upright in this corner, being used as some sort of internal screen or partition. In fact, the cane masses in this part of the building are divided into two discrete concentrations running parallel to each other and with a gap in between (see plate 4). These paired masses may represent the walls of a narrow vestibule or passage to the interior of Structure 5, creating a corner entry that opened onto the central plaza just like the entryway in the west corner of Structure 1. One final possibility is that the cane was simply being stored in Structure 5 for future use. We recovered very few artifacts from in situ contexts inside Structure 5—whether from the collapse deposits or from the floor itself—suggesting that the building was stripped of its contents before being set ablaze. The largest and most important artifacts recovered from the debris above the floor are two pieces of badly deteriorated iron that we have interpreted as fragments of a scale and its metal pan (figure 4.10; see figure 8.8). Each of these artifacts was visible at the top of Zone 2 in Unit 54 and was removed in a large soil block for laboratory conservation; they are discussed in greater detail in chapter 8. Because both objects were part of the collapse deposits (they were not resting on the structure floor), they may have been left suspended from a timber or overhead rafter at the time the structure

The Built Environment of the Berry Site Spanish Compound · 99

Figure 4.10. Steelyard scale “hook” fragment from Structure 5, Unit 54.

burned. They might thus have been overlooked during the stripping of the structure’s contents. Another artifact, a small quartz crystal discovered only a few centimeters from the iron scale (figure 4.11), is also important for its potential connection to Fort San Juan. According to Juan de Ribas, a member of Pardo’s second expedition, many of the Spaniards stationed at Joara collected “tiny diamonds” from veins of crystal discovered on a high hill

100 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.11. Small quartz crystal from Structure 5, Unit 54.

near the town, later gambling for them back at the fort (e.g., Hudson 1990: 163). There are no known sources of diamonds in the Carolina Piedmont, though outcrops of quartz crystal are quite common, and this is almost certainly the type of mineral that Ribas reported. The quartz crystal from Structure 5 is the only quartz crystal that we recovered in the excavation of either structure, and this may be one of the “tiny diamonds” collected by Pardo’s men. Structure 5 had a single floor, referred to as Zone 4 (see plate 5). The Zone 4 sediments were essentially unmodified subsoil composed of inconsistent organic sand. There is little evidence, as noted, that any parts of the Structure 5 floor had formal coverings such as organic mattings. Near the center of the building was a shallow and roughly circular hearth, Feature 91, that measured 0.9 meters in diameter and 14 centimeters deep (figure 4.12). Feature 91 was a simple basin lacking evidence of a more formal preparation, though it did contain abundant carbonized wood, at least some of which was probably fuelwood associated with the final use of the hearth (see chapter 5). Like the Zone 4 floor surface, the Structure 5 hearth contained only a single construction/use stage, which (as we discuss later) is a departure from Structure 1. Structure 5 had forty-six identified exterior wall posts and six interior supports (figure 4.13). A sample of seven excavated exterior posts had an average posthole depth of 31 centimeters (19–45 centimeters) and an average diameter of 27.4 centimeters (17–51 centimeters); four excavated interior posts had an average posthole depth of just 13 centimeters (6–18 centimeters) and an average diameter of 23 centimeters (13–32 centimeters). The shallow depth of the interior supports was surprising and would probably

Figure 4.12. Feature 91, central hearth in Structure 5.

102 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.13. Plan map of Structure 5 at top of subsoil, showing posthole distribution.

have made this a relatively unstable or even flimsy building. In fact, the four central posts appear to have been insufficient to support the building’s roof and superstructure, such that two additional posts were eventually placed between existing supports on the southern and western sides of the central post array. Although wall posts were placed much more deeply than central support posts (the most shallow of the excavated wall posts was deeper than the most deeply placed central post), the method of wall post emplacement, too, would have contributed to the building’s lack of structural stability. Many of the exterior wall posts were placed in postholes much wider than the girths of the posts (figure 4.13). These holes may have been dug with shovels and could therefore have resulted in posts that wobbled in their respective postholes. In one post, for example, the diameter of the light-colored posthole is more than twice that of the carbonized post within it (figure 4.14). A large piece of wrought iron—identified as a jack-plate

Figure 4.14. Detail of posthole, post mold, and carbonized post in Structure 5, Unit 2: a, overhead view with iron jack-plate fragment indicated; b, side view.

104 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

armor fragment (see chapter 8)—was wedged in between the edge of this posthole and the side of the post. While the posthole might have offered a convenient place to curate the object, there is another intriguing possibility. In medieval European folklore, iron was known for its protective properties against witches and evil spirits (Frazer 1890: 175; Merryfield 1987: 162), and iron objects were often buried in the foundations of buildings to prevent witches from entering (Manning 2012: 305–309). The iron jack-plate fragment jammed into this posthole may thus have served to protect or literally armor the building against dangerous supernatural forces. Other European artifacts from Structure 5 include six glass beads, an unidentified sheet copper or copper alloy fragment, a large piece of an unidentified iron object, and a small unidentified iron fragment. To summarize, Structure 5 seems to have been built quickly. Its internal support posts are shallow, and many of the outer wall posts were planted in postholes that may have been dug with metal shovels, affecting the structural integrity of the building. Moreover, analysis of cane culms from Structure 5 indicates that these were likewise prepared rapidly, with sharp spikes left on the culm; because these spikes can cause painful injuries, they were typically shaved down as part of fully processing the culms. Unlike the other burned structures, Structure 5 was not built within a basin of any kind. It thus has little internal stratigraphy save for the debris from a single building stage that burned and collapsed directly on top of the floor, with unprepared subsoil serving as the floor itself. There is a simple hearth at the center of the building, and there is no evidence for any post replacements or other major architectural renovations, other than the addition of at least two posts to bolster the weak central supports. Structure 5 was stripped of its contents before it burned, and we recovered very few artifacts on its floor surface. The presence of two large iron objects in the northeastern corner of the structure suggests that they were left by accident or perhaps were missed while hanging in a dark corner of the building. Structure 1 We began the 2008 excavations of Structure 1 by stripping all plow-zone soil from the building (see plate 6), revealing its overall form and the paired entry trenches located at the west corner of the building. Evident too are the excavations from 2003 and 2004 that exposed burned timbers in the western and northern parts of the structure. As with Structure 5, we divided Structure 1 into 1 × 1 meter units (figure 4.15). While the Structure 5 grid was oriented on the sides of the structure, with “structure grid north”

The Built Environment of the Berry Site Spanish Compound · 105

Figure 4.15. Structure 1 excavation units.

slightly west of “site grid north,” the 2008 Structure 1 grid maintained an orientation to the site grid established during 2003 preliminary testing. All of the field techniques used in Structure 1 followed those developed in Structure 5. There were six primary zone designations in Structure 1 (figures 4.16 and 4.17). Zone 1, which we usually referred to as a berm, is the clean fill deposited between the edge of the basin cut and the outer wall of the building after its construction was completed. We encountered Zone 1 in all edge

106 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

units except those in the entry, where the stratigraphy was significantly different. Three zones constitute the sediments directly atop the collapsed and carbonized architectural remains of Structure 1. Zones 2 and 3 refer, respectively, to highly oxidized red and unfired medium-brown sediments that filled the basin after the building’s collapse. Zone 6 refers to the pockets of highly oxidized red sediments of the same color and texture as Zone 2 that occur in Zone 3. These three zones, in sum, consisted of thoroughly mixed fired (red) and unfired (brown) sediments and were very difficult to differentiate as distinct deposits in the process of excavation. On the two profiles (figures 4.16 and 4.17), Zones 2 and 3 are combined, while pockets of Zone 6 are indicated separately. Some of these sediments likely derived from the walls and roof of Structure 1 and thus constitute part of the collapse deposit; other sediments in these zones, particularly the uppermost unfired sediments, may have been dumped over the smoldering remains of the building following its destruction. Zone 4 (see plates 7–9) is the lens of burned architectural debris in Structure 1 and corresponds to Zone 3 in Structure 5. What was initially among the more puzzling issues relevant to Zones 2 and 3 in Structure 1 and to the collapse deposits from Structure 5 is the nature of daub in these two buildings. Here we draw on thin section analysis of sediment micromorphology blocks removed from Structure 1. Daub typically makes up a significant proportion of the artifact assemblage where archaeologists excavate burned wattle and daub buildings. Daub is conspicuously absent from the macroartifact assemblage collected from the plow zone and from Structures 1 and 5, despite the bright red sediments revealed beneath the plow zone marking the outline of the burned buildings (see figure 4.5 and plate 2). Usually recovered as hardened red or yellow clayey aggregates, daub can also be identified archaeologically on the basis of frequent plant impressions (typically grasses) and occasionally other artifact inclusions such as pottery sherds, burned bone, and lithic debitage. Daub can often reveal important architectural details about a structure, including the building design, the types of wood used in construction, the nature of the smoke hole, and the presence of pigments still adhering to the walls. The combustion of a building exposed to temperatures exceeding 500°C transforms the structure of the clay minerals by forcing off water—removing the hydroxyl group—which causes the clay to permanently harden (Gibson and Woods 1990; Sinopoli 1991). Such temperatures also cause the ferrous iron oxide (Fe+2) to turn into ferric oxide (Fe+3) and re-

Figure 4.16. East-west profile across south corner of Structure 1, showing excavation zones.

Figure 4.17. North-south profile across south corner of Structure 1, showing excavation zones.

108 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

sult in the typical bright orange and red colors of daub (Schwertmann and Taylor 1989). The absence of traditional hardened daub at the Berry site is likely attributable to the composition of this important building material. Daub was clearly visible during excavation as localized areas of red sandy clays with plant pseudomorphs—leaf and stem impressions—observed. This material, however, was not hardened in the conflagration of the structure and is therefore part of the matrix of the site rather than an artifact that can be easily retrieved in the screening process. The analysis of daub from Structures 1 and 2 is therefore limited to field observations and mapping and details observed in the soil/sediment micromorphology thin sections (Sherwood 2014). Clays are not widely available in the immediate vicinity of the Berry site, either in the soil parent material (alluvium) or pedogenically (weathered and accumulating in the B horizon). The thin sections (see plate 11) suggest rather that clay was added to the daub. Micromorphology reveals relatively pure clay pieces—papules—that were produced by the process of levigation, wherein the coarse fraction of a sediment is separated from the fine fraction by adding water, mixing it into a suspension, and letting it settle and dry, thereby grading the sediments with the clay deposited on top of the column; the clay is then removed and crushed or ground. This process is typically carried out with containers for the levigation, after which troughs or pits are used to mix the clay with the nearby soils. The large pit features in the courtyard area of the compound (discussed later in this chapter) may have been used for this purpose. The origin of the clay could have been the upland residual soils located approximately 100 meters northwest of Berry. The added silt and sand-sized clay papules are visible as bright orange and red pieces in thin sections (see plate 11). The technique of levigation—or separating clays—was probably already in use for local pottery manufacture and could have been easily modified and adapted for such daub preparation. Limited amounts of clay would have been necessary to get the local sand and silt to adhere to the wattle. However, in the end, this limited amount of clay did not harden the daub when fired, but what clay was present in the daub matrix oxidized, changing the color but not the consistency. In several areas within Structure 1, there are thin black lenses in Zone 4. It is tempting to identify these as mats or as charred mat material from a bench. The thin section analysis, though, revealed that while there is a significant amount of monocot or cane among the microcharcoal, it is both

The Built Environment of the Berry Site Spanish Compound · 109

ephemeral and chaotic. One such lens in Zone 4 contained—besides the monocot—conifer wood, probable cattail, and an herbaceous stem with its surrounding sheath of leaves, as well as other unidentified charred plant material. Whatever this lens may be, the combination of materials and the absence of structure do not support its identification as a mat or suggest any one specific activity. The plants might have been used for floor covering, but they are known to have had many other purposes as well. Another thin black lens observed in the field in Structure 1 was also identified as potential matting. This lens (similar to the one described above) contained a mix of materials including a possible bark fragment, a conifer wood fragment, and stem tissue of horsetail or scouring rush. In sum, Sherwood’s (2014) micromorphological observations offer no consistent and definitive evidence for floor mats. If such mats were not preserved in a charred form, then the phytoliths should have been present since, in general, silica accumulation is higher in monocot than in nonmonocot species. If present, then, such phytoliths would likely have been preserved (e.g., Piperno 2006). Concentrated anthropogenic sediments—beyond building material— are also conspicuously lacking in thin sections from Structure 1 floor sediments. Microartifacts often become embedded in a floor and are far more likely to be left behind when macroartifacts are collected in a confined space. Sweeping of such sandy substrates as found in Structures 1 and 5 might prove difficult, so that the microartifacts should simply have worked their way into the floor. In domestic structures, abundant microscale debris relating to everything from storage in the rafters and on the floor to food consumption and preparation in and around the hearth should have been left behind. Such sediments often contain charred wood, seeds and other plant remains, burned and unburned bone, and pottery—not to mention wood ash and other secondary products. Only charred wood is ubiquitous in the Structure 1 deposits, and the wood types observed in thin sections correspond to the architectural identifications (Newsom, chapter 5, this volume). A single Burke phase sherd was identified in thin section, but this was from Zone 3 fill rather than from the floors. Perhaps most significant is the relative absence of burned bone, usually common in and around cooking hearths. Although there was little unburned bone recovered from the macroartifact assemblage, one should expect to see it preserved in the thin sections. At the center of Structure 1 we identified two superimposed hearths (figure 4.18). Before the excavation of the hearths, there was stratigraphic

110 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.18. Feature 119, central hearth in Structure 1: a, excavation profile showing superimposed Stage 1 and Stage 2 basins; b, overhead view showing ring of shallow postholes around Stage 1 basin.

evidence that the building had distinct, superimposed floors. The earliest of these surfaces was associated with the original basin cut; the other, Zone 5, was in use when the building burned (figures 4.16 and 4.17). The excavation of the two hearths confirmed our stratigraphic observations. The lower hearth, which we referred to as Feature 119, Stage 1, was a shal-

The Built Environment of the Berry Site Spanish Compound · 111

low basin dug through the original floor surface (that is, the basin cut). In this it was very similar to Feature 91, the single-stage hearth in Structure 5. This Stage 1 hearth had a ring of shallow postholes around it (figure 4.18b), and the heavy, greasy fill in the hearth suggests that some kind of rack was used for cooking or smoking a fatty substance such as bear meat. The upper hearth, designated Feature 119, Stage 2, contained a similar, greasy fill. The only evidence of floor preparation within Structure 1 occurs as patches of prepared sand beneath and surrounding each hearth basin, visible as red or rubified areas located at the center of the structure (see plate 10). While the preparation was minimal, clay papules were visible in thin section (see plate 12), indicating the use of clay levigation, as in the preparation of the daub. The clay must have been present for oxidation to have occurred, and the presence of such papules indicates that the clay was introduced. This kind of preparation is typical of Late Mississippian hearths but is usually more substantial, presumably to retain heat and to confine the fire. Zone 10 was a daub concentration surrounding the upper hearth in Units 20, 41, 42, and 65. This zone was within the collapse debris and was probably associated with a smoke hole in the ceiling. Like Structure 5, Structure 1 appears to have been stripped of its interior contents before it was burned. As a result—and also as a result of the limited kinds of activities that occurred within the building—few lithic and ceramic artifacts were recovered from in situ contexts on either of the floor surfaces. Combined, these floors contained just 61 fragments of lithic debitage, and a pitted quartzite hammerstone was found in Zone 5 (Moody 2013; Moore et al., chapter 9, this volume). These same floor surfaces contained 212 potsherds, however, suggesting that the absence of lithic debitage resulted not from sweepings of the structure floors but from the lack of lithic production inside Structure 1 (Moody 2013). The same interpretation holds for Structure 5, where 169 lithic debitage fragments weighing 154.4 grams (5.44 ounces) were recovered from the entire structure and Zone 4, the floor surface, contained only 2 of these artifacts. The presence of more than 30 potsherds in Zone 4 suggests that sweeping was not a factor in the small quantity of debitage in Structure 5. Large sections of at least two pottery vessels were recovered from the collapse deposits in Structure 1, Zones 2 and 3. These vessels were not resting on the last floor, Zone 5, at the time of the building’s destruction but may have been suspended from a timber or rafter, much like the scale recovered from the collapse deposits in Structure 5. We recovered fewer European artifacts in Structure 1 than we did in Structure 5, but these include four glass beads, two unidentified iron

Figure 4.19. Excavations in the entryway of Structure 1: a, entry trenches, entry step, and doorsill in situ (looking northwest); b, detail of wooden doorsill (OG#84), with sediment matrix removed.

The Built Environment of the Berry Site Spanish Compound · 113

fragments, and two pieces of iron wire that are likely fragments of chain mail. The entryway of Structure 1 was one of the most complex parts of the excavation (figure 4.19a). Just inside the entryway trenches was a place in the original floor surface (the base of the basin cut) where the subsoil was not cut to as deep a level as in surrounding areas, thus creating a step from the entryway down to the original, subsoil floor surface. There too was a piece of wood that was likely part of the doorsill or threshold; this timber, OG#84 (figure 4.19b), was clearly a distinct architectural component of the building and was not simply a part of the collapsed debris. We removed the sill as a block to facilitate its transfer to the laboratory, where Newsom has since identified it as a red oak plank (see figure 5.13). A similar wooden doorsill was recovered at Santa Catalina de Guale (1686–1702) on Amelia Island (Saunders 1991: 130). Each of the entryway trenches was excavated as a distinct feature; the northern trench was designated Feature 104, while we labeled the southern trench Feature 105. Both of these trenches terminated in double posts beyond the structure walls, and both measured approximately 1.5 meters in length (see plate 8). Each trench was shallowest at its terminal, exterior end, where it measured 4–6 centimeters deep, but Feature 105 reached a depth of 40 centimeters near the point where it met the structure’s outer wall; Feature 104 was not as deep, reaching a depth of 28 centimeters near its junction with the wall (figure 4.20a,b). The space between the trenches—the passage itself—was approximately 50 centimeters across, forming a narrow entry that would probably have required an adult man to walk into the structure with his shoulders turned sideways. The part of the entry area bounded by these two trenches, the doorsill, and the corner of the basin cut was a gently sloping surface or “step” down to the level of the structure’s original floor surface (figure 4.19a). This area was within the basin cut yet outside the structure walls, such that its sediments contained almost no carbonized organic debris. The uppermost zone of fill in this area had the consistency and texture of Zone 3 and may have similarly derived from the collapsed walls and roof of the entryway passage. Beneath these sediments was a compact, mottled deposit identified in field notes as a trampled area and referred to as Zone 7. Zone 8 was a deposit of dark, organic sediments in Zone 1 along the southern edge of Unit 50, while Zone 9 was a modern disturbance in the same unit. The building’s footprint was fully apparent at the base of Zone 5, especially the entryway trenches and the edges of the basin cut (see plate 10).

114 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

40 cm

40 cm

Figure 4.20. Profile and cross section of Structure 1 entry trenches: a, Feature 104, northern entry trench; b, Feature 105, southern entry trench.

Also clear was the distribution of postholes on the southeastern and southwestern walls and the line of posts that formed a partitioning wall or screen inside the entry (figure 4.21). The southeastern and southwestern wall postholes were spaced about 50 centimeters apart, while upright posts on the same walls were spaced a bit less than 1 meter apart (see plate 9). There were seven upright posts and twelve postholes on the southwestern wall, and

The Built Environment of the Berry Site Spanish Compound · 115

Figure 4.21. Plan map of Structure 1 at top of subsoil, showing posthole distribution.

four upright posts and six postholes on the southeastern wall. This pattern of posts and postholes may indicate the differential preservation of posts along the two walls, though it may also suggest a pattern of post replacement. That is, posts may either have been spaced quite close together but were differentially preserved as carbonized uprights, or else they may have been spaced farther apart at any given time during the building’s period of use and were periodically replaced. In excavating approximately 60 percent of Structure 1, we recorded twenty-seven exterior wall posts and four central supports, not including the screen or partitioning wall inside the entry (figure 4.21). A sample of fifteen excavated exterior wall posts had an average posthole depth of 35 centimeters (16–59 centimeters) and an average diameter of 19 centimeters (14–34 centimeters); the two excavated supports had an average depth of 59 centimeters (55–63 centimeters) and an average diameter of 15.5 centimeters (14–17 centimeters). The western support post, OG#21/61 (figure 4.22), was first observed, though not removed, in 2003. After removing

Figure 4.22. Carbonized base of western central support post, OG#21.

Figure 4.23. Excavation of OG#21/61 posthole: a, preserved portion of post in posthole; b, Elizabeth Horton removing preserved portion of post (OG#61).

The Built Environment of the Berry Site Spanish Compound · 117

the carbonized, upright part of this post (designated OG#21), we began to excavate its associated posthole. We were soon surprised to find that much of the belowground portion of the post was intact—neither rotted nor burned (figure 4.23). The intact portion of the post was 68 centimeters in length and 12.5 centimeters in diameter; this part of the post was designated OG#61. Notice too how closely the diameter of the posthole corresponds to the girth of the post, suggesting that it was rammed into place. The southern support post (OG#60/72) was similarly preserved in its posthole, though the intact portion was less substantial than that recovered from the western support. Feature Excavations A feature designation was assigned to all pits, hearths, midden deposits, building entrance trenches, and large posts not associated with the burned buildings. Since 1986 we have identified more than two hundred features at the Berry site, and of these we have excavated more than sixty. The vast majority of features we have identified and excavated thus far are located in the Spanish compound (see figures 1.5 and 4.1). Our aim here is not to describe all excavated features but rather to concentrate on that sample of features containing European artifacts or copper fragments that likely derive from nonaboriginal sources (figure 4.24). To date, we have excavated sixteen such features in the Spanish compound. (For in-depth perspectives on feature contents—including botanical and faunal remains, European artifacts, and artifacts of native manufacture—see chapters 7 through 9.) For reasons explained in the following section, we have subdivided these features into two different clusters that we refer to as the central features (21, 23, 25, 38, 48, 64, 68, 69, 83, 92, and 112) and the western features (66, 71, 103, 106, and 108). The descriptions that follow are sorted accordingly. Central Features Feature 21 This feature was a shallow and irregularly shaped pit located just outside the southwest wall of Structure 1; the southern portion of the pit was longer than the northern portion and was filled with fire-reddened cobbles. Feature 21 measured only 10 centimeters deep and was 1.28 meters in length by 94 centimeters in width. Besides the fire-reddened stones, this feature contained few artifacts (one potsherd, one soapstone fragment, and two flakes). However, two very small fragments of possible copper or copper alloy were recovered during water-screening of Feature 21 fill.

118 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.24. Plan map of the Berry site Spanish compound indicating all features with European artifacts.

Feature 23 Feature 23 consisted of two overlapping pits (figure 4.25a), one of which— the north lobe—was oval in shape (figure 4.25b), while the other, the south lobe, was circular (figure 4.25c). Excavations revealed that the northern edge of the south lobe was intrusive into the north lobe. Both pits had flat bottoms and several zones of fill (figure 4.26), though the north lobe was

Figure 4.25. Feature 23: a, base of plow zone; b, north lobe excavated; c, south lobe excavated.

120 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.26. Feature 23 profiles: a, north lobe profile; b, south lobe profile.

larger and had the more complicated stratigraphy. The north lobe pit had a maximum depth of 50 centimeters and measured 2.2 meters long by 1.48 meters wide. The south lobe pit had a maximum depth of 30 centimeters and measured 1.66 meters long by 1.68 meters wide. Zone 1, a dark gray to black deposit, extended across the tops of both pits. It contained substantial quantities of potsherds, animal bone, lithic debitage, charcoal, and shell, along with a scattering of European artifacts. It was likely a midden or refuse deposit that formed after the lower zones of the north and south lobes had slumped and left a pair of depressions in the ground surface. Feature 23 contained more European artifacts than any other excavated feature. Zone 1 (in both lobes) yielded a sherd of Spanish Coarse Redware, three glass beads, one copper bead, and four sheet copper or copper alloy fragments. Below Zone 1, all European artifacts were concentrated in the north lobe: Zone 5 contained one copper tinkling cone or aglet, four sheet cop-

The Built Environment of the Berry Site Spanish Compound · 121

per or copper alloy fragments, and one unidentified iron fragment; Zone 6 yielded one glass bead and two copper or alloy fragments. Feature 25 Feature 25 was remarkably similar to Feature 23 in that it was also composed of two overlapping, roughly circular pits (figure 4.27). Of these, the north lobe measured 48 centimeters deep, 1.3 meters long, and 1.1 meters

Figure 4.27. Feature 25: a, base of subsoil; b, after excavation.

122 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.28. Feature 25 profiles: a, north lobe profile; b, south lobe profile.

wide; the south lobe was broader but more shallow, with a maximum depth of 35 centimeters and measuring 1.95 meters in length by 1.35 meters in width. As with Feature 23, some of the Feature 25 deposits—in this case Zones 2 and 3—extended across the two lobes, indicating that both were open at the same time (figure 4.28). Zone 2, comparable to Zone 1 in Feature 23, was a very dark brown, charcoal-flecked deposit containing substantial quantities of potsherds, animal bone, and other artifacts. It, too, is likely a midden or refuse deposit. Six European artifacts were recovered from Feature 25, evenly distributed across both pits. Zones 2 and 3 yielded two copper tinkling cones or aglets, one copper or alloy fragment, and one unidentified iron fragment; Zone 6 contained one copper tinkling cone or aglet. Feature 25 also contained one sherd of Mexican Red Painted ware, but its stratigraphic context is undetermined.

The Built Environment of the Berry Site Spanish Compound · 123

Feature 38 This roughly circular pit was located just west of Feature 25. It measured 41 centimeters deep, 1.83 centimeters long, and 1.65 centimeters wide (figure 4.29a). Four substantial zones were identified, together with several thin lenses between zones (figure 4.29b). As in Features 23 and 25, Zone 1 contained quantities of artifacts and other materials, including potsherds, animal bone, lithics, and charcoal. Two European artifacts were recovered

Figure 4.29. Feature 38: a, base of plow zone; b, plan and profile.

124 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

from Feature 38, including an unidentified iron fragment from Zone 1 and a sherd of Spanish Fine Thin Redware from Zone 4. Feature 48 This feature was a shallow and amorphous stain located outside the northeast wall of Structure 5. It measured 1.1 meters in length by 74 centimeters in width, and its depth varied from 2 centimeters to 7 centimeters. Feature 48 contained little cultural material, but a single glass bead was recovered from the heavy fraction of its 5 liter flotation sample. Features 64/68/69/83 Near the center of the Spanish compound were four large, round to oval pit features with edges that touched or overlapped only slightly (figure 4.30, see also figure 4.24). All were very similar in size and shape both to one another and to Features 23, 25, and 38, and like these others, all had multiple zones of fill. Feature 64 was the smallest and shallowest of its group and contained somewhat less cultural material than its neighbors (see table 9.1). It was 28 centimeters deep, 1.37 meters long, and 1.2 meters wide. Most of its associated artifacts were concentrated in Zone 1 (figure 4.31a), a dark brown fill that may be the same sort of midden deposit identified in each of the large pits previously described; a single olive jar sherd was recovered from Zone 5. The southeastern edge of Feature 64 overlapped just slightly with Feature 68, which measured 35 centimeters in depth, 1.65 centimeters in length, and 1.65 centimeters in width. Most of the cultural material in Feature 68 was evenly distributed between Zones 1 and 2, which together made up 63 percent of the fill volume of the feature (figure 4.31b). Zones 3 through 10 were thin lenses that made up 37 percent of the remaining fill. An unidentified iron fragment was recovered during the water-screening of deposits removed while troweling for a photograph, and it cannot be assigned to a specific zone. Feature 69 was the largest feature of this cluster, measuring 53 centimeters deep, 2 meters long, and 1.8 meters wide, and it contained more artifacts than any of the other three pits in its cluster. It had a very complicated stratigraphy (figure 4.32a), holding ten distinct zones (like Feature 68), and—as with most of the large central pits—we recovered most of its cultural materials from Zone 1. It also had four European artifacts, more than any other pit in this cluster, including a glass bead recovered during profile trimming and three unidentified sheet copper or copper alloy fragments, all recovered from Zone 1. Feature 83 measured 46 centimeters deep, 1.6

Figure 4.30. Feature 64, 68, 69, and 83 cluster at base of plow zone.

Figure 4.31. Features 64 and 68 profiles: a, Feature 64 profile; b, Feature 68 profile.

126 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.32. Features 69 and 83 profiles: a, Feature 69 profile; b, Feature 83 profile.

meters long, and approximately 2.5 meters wide; its precise width is not yet known, as its eastern edge falls into an unexcavated unit (see figure 4.30). It contained nine stratigraphic zones (figure 4.32b), and though it yielded fewer artifacts than Features 68 and 69, with about the same density of material as Feature 64, it did yield a European artifact, a sherd of Spanish Coarse Redware recovered from Zone 3. Feature 92 Feature 92 was a large pit located beneath the southwestern corner of Structure 5 and predating the building; the structure’s southwest support post was intrusive into the fill of the pit (see figures 4.13 and 4.24). The feature’s form was unusual in that it consisted of two distinct cuts: a relatively shallow “shelf ” measuring 2 meters in width, 2 meters in length, and no more than 20 centimeters in depth; and a deeper, circular pit measuring 80 cen-

The Built Environment of the Berry Site Spanish Compound · 127

timeters in diameter and extending 32 centimeters below the base of the shelf (figure 4.33a). The feature’s fill was quite homogenous, however, as a single deposit designated Zone 2 filled both the pit and the shelf (figure 4.33b), suggesting that Feature 92 had been rapidly sealed in one episode. The fill seems to have slumped over the central pit area, and Zone 1—a very dark brown midden accumulation—then filled the slump. Artifacts were distributed through Zones 1 and 2, and the feature contained one European artifact, an iron butted chain mail ring recovered from Zone 2. Feature 112 This feature was a large pit located immediately outside the southwestern corner of Structure 5 and just south of Feature 92. Feature 112 was roughly circular and of nearly identical dimensions as Feature 92, measuring 54

Figure 4.33. Feature 92: a, after excavation; b, profile.

128 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.34. Feature 112: a, excavating feature; b, profile.

centimeters deep, 2 meters long, and 2 meters wide (figure 4.34a). It also had fewer fill zones than the comparatively smaller features to the east, particularly those making up the Feature 64/68/69/83 cluster. Feature 112 contained more fragments of nutshell—especially hickory shell—than any other excavation context within the Spanish compound. Gayle Fritz views this frequency as evidence for the preparation of kunuchee, the traditional Cherokee hickory nut soup (see chapter 6). While most of its cultural materials were recovered from Zone 1, the underlying Zone 2 (figure 4.34b)

The Built Environment of the Berry Site Spanish Compound · 129

revealed an unusual assemblage of sherds from several ceramic jars. Most of these vessels were tempered with soapstone, which is typical of Burke phase ceramics, but had surface treatments more commonly found on Pisgah ceramics from the Appalachian Summit area to the west of the Berry site (see chapter 9). Zone 2 consisted of a highly organic, greasy fill that might be explained by its high concentration of animal bone. Fritz, though, also suggests that the greasy texture might have resulted from hickory nut oil or milk production (see chapter 6). Zone 2 yielded the feature’s only European artifact, an unidentified copper or alloy fragment. Western Features Feature 66 This relatively shallow feature was at the northern end of the mass of pits and midden deposits that we identify as the western features; all of these numbered contexts appear to have been associated with Structure 5, since all are located immediately behind this structure and run north to south along its western wall (figure 4.35). This feature cluster includes four circular pits (designated Features 66, 71, 111, and 122) and four somewhat arbitrarily defined sections of an extensive but shallow midden (Features 103,

Figure 4.35. Western feature and midden complex, looking northwest.

Figure 4.36. Feature 66: a, after excavation; b, plan and profile.

The Built Environment of the Berry Site Spanish Compound · 131

106, 109, and 110). The midden deposit appears to have overlain two of the associated pit features, Features 111 and 122, while the other two pits, Features 66 and 71, may have been intrusive into the midden. Feature 66 was 1.5 meters long, 1.25 meters wide, and only 18 centimeters deep (figure 4.36a). It consisted of two stratigraphic zones (figure 4.36b), and although both zones contained few artifacts, we did recover an unglazed olive jar sherd while shovel skimming across the surface of this pit or basin. The shallow depth of its deposits may indicate that this feature is part of the midden rather than a separate pit. Feature 71 Feature 71 was a circular pit at the southern end of the western feature group; it was 1.4 meters long, 1.35 meters wide, and 48 centimeters deep (figure 4.37a). Of all the features in the western group, it was most similar in size, shape, and stratigraphy to those in the central group, especially those in the 64/68/69/83 cluster. It had at least six distinct stratigraphic zones (figure 4.37b), but unlike the central pits, its artifacts were distributed evenly throughout the deposits. Only Feature 23 contained more European artifacts than Feature 71, which contained four glass beads, one iron butted chain mail ring, and two unidentified iron fragments. Features 103/106 Features 103 and 106 were arbitrarily defined sections of the extensive but shallow midden deposit west of Structure 5 that also included Features 109 and 110; the latter two sections contained no European artifacts. Features 103 and 106 were located in the north and central areas of the midden, respectively, and each comprised two stratigraphic zones that together reached a maximum depth of 10 to 13 centimeters. The Feature 103 section measured 1.95 meters in length and 1.5 meters in width, while the Feature 106 section measured approximately 3.5 meters in length and 3 meters in width. Feature 103 contained one European artifact, a glass bead from Zone 2; Feature 106 yielded four glass beads, a wrought iron nail, and two fragments of a copper bell, all recovered from Zone 1. Feature 108 This feature is located just south of the primary mass of western features but is considered as part of the western features (along with Feature 107) because of its clear spatial association with this group. Feature 108 is ap-

132 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.37. Feature 71: a, base of plow zone; b, plan and profile.

proximately 1.5 meters long and 2.5 meters wide, but because it has not yet been excavated, we have no data on its internal stratigraphy or depth. We include it here because we recovered a small iron fragment from its surface during trowelling.

The Built Environment of the Berry Site Spanish Compound · 133

A History of the Spanish Compound Juan Pardo founded the Spanish compound at the Berry site—the domestic context of Fort San Juan—in late December 1566 or early January 1567. By the spring of 1568, it lay smoldering in ruins. In itself, this window of less than eighteen months provides a remarkably tight chronology for studying the archaeology of colonial encounters, particularly so in the American Southeast, where archaeologists often measure time in decades or even centuries. Yet the Spanish compound gives us more than a snapshot of the encounter. The archaeology of its built environment, the buildings and features, allows us to see how the encounter changed over time; we have been able, that is, to divide the brief history of the Spanish compound into two even briefer phases of occupation, each of which may have lasted no more than several months. In the rest of this chapter, we discuss our understanding of how the spatial organization of Fort San Juan’s domestic contexts changed from the time of its founding to that of its fiery destruction. There are two dimensions to our analysis of the compound: comparison of architectural attributes in Structures 1 and 5 and comparison of the stratigraphic associations between these buildings and the large features just described, from which we can extrapolate in a limited way to the other, as-yet-unexcavated buildings. First, the most obvious difference between the structures is the amount of time and labor exhibited in each building’s archaeological footprint. Comparing the base of excavation photomosaics from Structures 1 and 5 (plates 10 and 5, respectively) clearly shows that Structure 1 has a well-defined basin, deep entry trenches, and four large, central support posts. Structure 5, having neither a basin pit nor entry trenches, appears to have been more ephemeral. The nature of the posts is another point of difference. As discussed earlier, many of the Structure 5 posts display an unusual pattern in which the posts are significantly smaller than the holes in which they were placed, suggesting that the postholes were dug out with metal tools such as the shovels that Pardo left at Fort San Juan (see table 8.4 for a list of supplies and other goods that Pardo issued to each interior fort). In Structure 1, though, postholes correspond to the size of their respective posts (see, for example, figure 4.23). Lee Newsom’s wood analysis (chapter 5, this volume), also reports dramatic distinctions in the age and size of wood used in each structure. Roundwood used in Structure 1 was twice as old, on average, as roundwood from Structure 5 (the diameter of Structure 1 samples likewise averaged double the width of samples from Structure 5). Moreover, the majority of nonpost roundwood

134 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

from Structure 1 consisted of entire boles, while the majority of Structure 5 samples derived from halved boles. Structure 5, in sum, was a far less durable and substantial building than its counterpart, Structure 1. Many of the architectural attributes associated with Structure 1—especially its basin pit, its entry trenches, and the placement of its posts—are quite typical of Late Mississippian structures in the southern Appalachians (see Newsom, chapter 5, this volume, for an in-depth comparison of the Berry site structures with Late Mississippian architecture at other sites). What is atypical of indigenous buildings at other sites is the use of metal tools in wood preparation and construction. Both of these Berry site buildings have clear evidence for the use of such tools, and in Structure 1 these include saws, axes, and iron nails or spikes. Structure 1, that is, displays a combination of European and Native American carpentry practices and techniques. Yet were it not for the exceptional preservation of wood and other organics in this building, its footprint would be all but indistinguishable from that of buildings at contemporaneous sites that lack a Spanish component. This strongly suggests that the construction of the building was planned and accomplished by native workers using wood cut and prepared by Europeans. Structure 1 would thus seem to have required a substantial degree of cooperation between the soldiers of Fort San Juan and their native hosts. Structure 5 is distinct, as its architectural style and form have less in common with Native American construction practices in this time and place. Besides its lack of a basin pit, Structure 5 has no clearly defined entryway, unless the paired cane concentrations in the eastern corner of the building created a vestibule-like passage to the structure’s interior. Such a feature, though, would itself be unusual in a Late Mississippian building, and Horton’s analysis of the cane indicates that it was likely processed by people unfamiliar with this material (for example, the extremely sharp spikes at the nodal ends of the culms were not removed before use, as is typical on samples recovered from wholly indigenous contexts). Many of this building’s posts, as noted, were set in large postholes dug with shovels, and the central support posts were placed in extremely shallow holes averaging only 13 centimeters deep. This appears to have created enough instability that two additional center posts were added on the south and west sides of the structure’s interior. Finally, Newsom (chapter 5, this volume) reports likely wooden roof shakes or shingles among the organic specimens recovered from this building, one of which has an apparent nail hole (see figure 5.8). While the use of wooden shakes was a common practice in six-

The Built Environment of the Berry Site Spanish Compound · 135

teenth-century European architecture, it was not a traditional construction practice in the Native South. In sum, Structure 5 seems to have been built according to an architectural grammar distinct from that of Structure 1 and of contemporaneous, native-built structures in the southern Appalachians. Along with this variability in carpentry practices and construction techniques, there is also clear evidence of a temporal distinction between Structure 1 and Structure 5, suggesting that each was built at a different time in the compound’s history. Recall that Structure 1 had superimposed hearths at the center of the building and two floor surfaces—one associated with the original basin cut and the other with the final burning event. Structure 5, in contrast, had only the original floor surface, and its hearth had only a single construction stage. We are confident, then, in concluding that Structure 1 saw a longer period of use and occupation than Structure 5. We should point out, however, that neither building appears to have been in use for an extended period of time—neither structure contains evidence of extensive remodeling, for example (aside from the addition of two support posts in Structure 5 and the new floor surface and accompanying hearth in Structure 1), and neither was rebuilt after the fire that consumed them both. By comparing the spatial and stratigraphic relations between each building and the central pit features, we can further deduce that Structure 1 predates Structure 5. Structure 5 intrudes into no fewer than five of the large pits, yet none of these features was dug in close proximity to Structure 1 (see figure 4.24). Because all but one of the pits we have excavated in the central feature area held European artifacts (Feature 65 was a large but shallow feature that had little in common with the deeper pits), we assume that all were associated with the Spanish occupation of Fort San Juan. This observation allows us to make limited inferences about Structures 2, 3, and 4, as well. Like Structure 1, Structure 3 does not appear to be intrusive into any earlier pits, and the east wall of Structure 4 similarly shows no evidence of intruded features. Structure 2, though, intrudes into a cluster of three large features, and while none of these has yet been excavated, all are located in the central pit area. We suggest that Structures 1, 3, and 4 were built first and that the central pits were dug in association with the construction and use of these three buildings. Structures 2 and 5 were built at a later point in the compound’s history and so intrude into the central pits. We propose two phases of primary construction activity within the Spanish compound. In the first phase of occupation, three houses—Structures 1, 3, and 4—were built together in a slightly curved arc at the northernmost edge of what is now the Berry site (figure 4.38). The start of this

136 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.38. Plan map of Spanish compound, first-phase buildings and features with possible kitchen indicated.

phase coincides with Pardo’s arrival at Joara in late December 1566, and we expect that this early period of the garrison’s occupation would have been characterized by close, cooperative relations between the soldiers and their Indian hosts. This would explain the blending of distinct carpentry and construction practices observed in Structure 1. Indeed, when Pardo returned to Joara with his second expedition in September 1567, having left thirty men to garrison the fort the previous winter, scribe Juan de la Bandera (1990: 265) reported that “he found built a new house of wood with a

The Built Environment of the Berry Site Spanish Compound · 137

large elevated room full of maize, which the cacique of the village . . . had built by the command of the captain for the service of His Majesty.” This might well have been Structure 1. Yet Bandera’s testimony also poses a challenge to our analysis of the compound’s history. Specifically, his account could be interpreted to mean that only one structure had been built at the time of Pardo’s September return. We find this to be an improbable scenario. Even though many of the thirty men left at Joara under Hernando Moyano would spend much of the next few months in the Appalachians, we think it unlikely that they would all have expected to live in the same house while they were at the compound. Pedro Menéndez de Avilés organized his soldiers into ten-man mess groups who lived together and received all rations and supplies as a unit (Manucy 1997: 62). If Pardo used a similar plan, then the three structures we identify as early could each have housed ten of the thirty men left to garrison Fort San Juan. In this case, Bandera’s reference to “a new house of wood” may have been to a structure built by Joara Mico and his townspeople between the time of the two expeditions. The two other structures in the compound’s early phase might have been built either by Indians and Spaniards working together while the first expedition was at Joara or by the soldiers working alone after Pardo’s departure. Only additional excavation of these structures can further resolve such tensions between archaeology and text. The central pits were probably dug at the same time that Structures 1, 3, and 4 were being built and may have initially served as levigation pits for processing daub. After this first phase of building construction was complete, these features became the primary locus of refuse disposal in the newly established compound. This pattern, with trash disposal in discrete pit features located some distance away from houses, is typical of Spanish colonial settlements and contrasts with the English colonial (or Brunswick) pattern of sheet midden accumulations near house doorways and around the buildings (Deagan 1983: 76–78, 107–108; South 1977: 47). Moreover, daub processing pits at other Spanish settlements—including Santa Elena (1566–1587) and Santa Catalina de Guale (1602–1680)—were likewise converted to refuse pits (e.g., South 1982: 26–28; South and DePratter 1996: 57–64; Thomas 2010: 36, 39). At a house site in eighteenth-century St. Augustine, a primary sheet midden deposit did occur near a structure believed to have served as a kitchen (Deagan 1983: 106, 111), a point to which we return when detailing the second phase constructions. A circular hearth (figure 4.39) measuring 66 centimeters in diameter and about 10 centi-

138 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.39. Feature 76 after excavation, showing fuelwood in situ.

meters deep was situated at the eastern edge of the central pits and may have functioned as an open-air cooking area or kitchen during the first phase of the compound’s use. Archaeologists have identified such open-air hearths—though more formally designed than Feature 76—at both the de Hita site in eighteenth-century St. Augustine (Deagan 1983: 75) and behind the officers’ barracks at Presidio Santa Maria de Galve, occupied from 1698 to 1719 in Pensacola, Florida (Bense 2003: 119–120). Like Feature 76, these cooking hearths were located near circular pits that contained large quantities of refuse. We cannot be certain when the second phase of construction activity began, though it may have commenced with Pardo’s three-week sojourn at Joara in mid-November 1567. For much of the time between Pardo’s first and second expeditions, there were perhaps as few as ten soldiers in residence at Fort San Juan, since on at least two occasions Hernando Moyano took twenty men across the Appalachians to attack Joara’s enemies. When Pardo returned to Joara in September 1567, on his way to relieve Moyano at Chiaha, he may have discovered the compound in a relative state of neglect. According to Francisco Martínez, Moyano had been in the mountains since midsummer with two-thirds of the garrison, leaving but a few men

The Built Environment of the Berry Site Spanish Compound · 139

to hold the fort. It is likely no coincidence that when Pardo left Joara for what would be the last time, he replaced Sergeant Moyano with his ensign, Alberto Escudero de Villamar; afterward, Bandera (1990: 286) identifies Moyano as “former sergeant of the company.” Pardo conferred on Escudero the legal authority to govern the other forts from Fort San Juan, to travel among them in order to do so, and to preserve the friendship of all the chiefs (Bandera 1990: 278). Clearly, Pardo meant for Fort San Juan and its adjoining town of Cuenca to be the Spanish capital in the interior (Hudson 1990: 150). It is reasonable to suggest that Escudero, taking his charge seriously, soon began to formalize the community. In this second phase of occupation, two buildings—Structures 2 and 5—were built together along the same axis 15 to 20 meters west of the first phase structures (figure 4.40). Each of the new buildings intrudes into the central pit features, establishing the temporal priority of the latter. We can say little beyond this about Structure 2, since as yet we have exposed only its southern corner and a portion of its eastern side. Auger testing conducted in 1997, however—when the structures were all first discovered—indicates that Structure 2 is the same overall size and shape as the other buildings in the compound. Moreover, these tests revealed that deposits in Structure 2 are similar in depth to those of Structures 1, 3, and 4; that is, Structure 2 appears to have been built in a basin of some kind. Structure 5 is thus an anomaly. It is to this building and to the western features, an accumulation of circular pits and sheet midden deposits immediately behind it—that we must turn for a better appreciation of both this second phase of construction and the kinds of changes it may have introduced to daily life and practice in the Spanish compound. Although we argue that Structures 1, 3, and 4 (and likely Structure 2) were built as houses to lodge the Spanish soldiers at Fort San Juan, the specific peculiarities of Structure 5 suggest a different purpose to its construction. The phrase Spanish compound conveys the idea that this was a place— an arrangement of buildings, features, and the spaces in between—where Spanish soldiers lived; it should also convey the idea that the place itself was perceived as Spanish, at least by some of its inhabitants, and that we must therefore consider Spanish practices for understanding aspects of its layout and design. Archaeologists have excavated admittedly few Spanish contexts that are directly comparable to the early frontier situation of Fort San Juan, but after reviewing an array of data from the Spanish colonial towns and missions of La Florida, we suggest that Structure 5 was a formal cocina, or kitchen, for the second phase of the compound’s use.

140 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.40. Plan map of Spanish compound, second-phase buildings and features with possible kitchen indicated.

Across La Florida, from the sixteenth century through the eighteenth, Spanish kitchens were usually stand-alone structures located in close proximity to refuse pits, middens, and wells. On house lots in towns such as Santa Elena and St. Augustine, kitchens were located at the back of the lot (Manucy 1997: 62–64). No obvious kitchens (or even well-defined house lots) have been identified in sixteenth-century St. Augustine, though Kathleen Deagan (1985b: 13) has suggested that a wattle and daub structure at the Trinity Episcopal site may have served as a kitchen, as indicated by the

The Built Environment of the Berry Site Spanish Compound · 141

presence of a nearby well and the large quantity of food-processing debris. At Santa Elena, too, no definite kitchens have yet been excavated. Stanley South and Chester DePratter (1996: 105–107), however, may have located the edge of a kitchen in a high-status lot dating to the second occupation of the town (1577–1587). Some of the most detailed data from kitchens on town sites comes from the de la Cruz site in eighteenth-century St. Augustine. While this site is far removed in time from the Spanish compound at the Berry site, it does offer some suggestive comparative data. The de la Cruz site consisted of three buildings, two of which—both built along the street—have been identified as houses (Deagan 1983: 109). The third building, placed in back of the other two, was considerably smaller than the houses and was the only area of the lot with primary sheet midden; there was also a concentration of trash-filled pits around the structure (Deagan 1983: 107–108). In addition, it was close to the de la Cruz well, it was built on top of earlier pits associated with cooking activity, and all of the lot’s nonceramic food-processing wares were found in its immediate vicinity (Deagan 1983: 109–111). In general, Deagan (1983: 77–78) links kitchens with the St. Augustine pattern of refuse disposal, observing that trash pits were located “away from the house structure . . . with the heaviest concentrations of refuse in the vicinity of wells and the kitchen.” If we turn to the Franciscan missions for another example, a royal ordinance required mission complexes to have a church, a convento that housed its priests, and a self-contained cocina, all of which were arranged around a central plaza (Jones and Shapiro 1990: 504; Weisman 1992: 6). The best documented of the mission cocinas is perhaps the one associated with Santa Catalina de Guale on St. Catherine’s Island (e.g., Thomas 1991: 117). This Santa Catalina kitchen was located about 20 meters north of the convento (with which it was aligned) and measured 4.5 meters by 6 meters. The structure was enclosed on three sides by wattle and daub walls; the fourth side, facing the well, remained open. Squared-off pine posts set in pits supported the walls, and a hearth was located near the center of the structure. Midden deposits had accumulated in large features close by, which may have originally served as daub processing pits. A probable kitchen has also been excavated at the Harrison Homestead site on Amelia Island, Florida, where Santa Catalina de Guale moved in 1680 after the St. Catherine’s Island mission site was abandoned (Saunders 1991: 126). This rectangular building measured 6 meters by 3 meters and was located to the northwest of the convento, with which it shared an alignment of 21 degrees east of

142 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

north (Saunders 1991: 128, 132). As with the St. Catherine’s mission kitchen, this building was open on one or more sides, forming a ramada. Rebecca Saunders (1991: 132) observes that “such a light structure associated with a midden was consistent with a kitchen.” Located just to the south of the structure was a feature that contained burned wood and seeds, probably a hearth, and near it were concentrations of animal bone and shell. None of these structures—whether those in town or in mission settings—is an exact match for Structure 5 at the Berry site, particularly in the details of the architecture. This is not unexpected, given the different cultural and temporal contexts that separate an early frontier garrison such as Fort San Juan from later towns and missions. We do note, though, that there is substantial variation in the architecture of later town and mission cocinas, too. The point of this comparison, then, is not to match the architectural details of Structure 5 to these identified cocinas but rather to recognize the broader trends and patterns in these buildings such that we might identify kitchens in different kinds of colonial settings, and here we think that the exercise is illuminating. We have yet to find evidence of a well near Structure 5—or anywhere else at the Berry site—but otherwise this building had much in common with those described as cocinas elsewhere. It was unsturdy or lightly built in relation to other buildings in the compound, and while it was enclosed on all sides, an open plan like that of the two Santa Catalina de Guale structures would not have been practical given the colder winter temperatures so close to the Appalachians. Alone among the structures in the compound, it was immediately adjacent to active refuse pits and a large midden, the only such deposit identified in the compound area. Microartifact data also suggest that meat preparation took place inside Structure 5. While bone was poorly preserved in macroartifact samples (see Lapham, chapter 7, this volume), several units in the building—Units 21, 30, 38, 43, 46, and 47—exhibited high bone counts (>100) in the microartifact samples; significantly, all of the bone in these samples was calcined. This burned bone cannot be explained by the fire that destroyed the building; rather, it suggests that cooking was an important activity inside Structure 5 (Sherwood 2014). Moreover, the five units with the most bone (21, 30, 38, 46, and 47) form a contiguous area just east of the hearth (see figure 4.6), suggesting that this zone was associated with meat processing activities. Identifying Structure 5 as a possible kitchen also explains the presence of the steelyard scale. Spanish soldiers received rations of food and other supplies in measured quantities, often as part of a mess or com-

The Built Environment of the Berry Site Spanish Compound · 143

rade group. This scale might thus have been a necessary piece of equipment for the compound’s cocina, guaranteeing that each mess group garrisoned at the fort received its allotted amount of food. Finally, we can return to the iron jack-plate fragment recovered from one of the structure’s external wall posts. If this building was indeed a kitchen where native women regularly prepared dishes for Spanish soldiers, then the iron fragment may have ensured that no witches entered the building with the intent of poisoning the soldiers’ food—a threat that must have seemed all too real for medieval Catholics stationed among non-Christians. As we widen our view to the Spanish compound as a whole, we find another clear parallel between its second construction phase and most other Spanish settlements in La Florida. All of the places just described—towns and missions alike—were tightly gridded communities, with structures, plazas, and other major features having the same orientation. While particular layouts might vary from one settlement to the next, often following the natural topography, all exhibit clear evidence of having been highly planned (e.g., Deagan 1983: 9–10, 2009: 325; DePratter and South 1995: 84; Saunders 1991: 128; Thomas 1991: 109). This town planning was formalized in the 1573 Laws of the Indies, Ordinance 110, which stipulates that “a plan for the site is to be made, dividing it into squares, streets, and building lots, using cord and ruler, beginning with the main square from which streets are to be run to the gates and principal roads and leaving sufficient open space so that even if the town grows, it can always spread in the same manner” (Mundigo and Crouch 1977: 254). Moreover, this very act of town planning was meant to have a profound effect on how indigenous communities perceived the Spaniards among them. According to Ordinance 137, While the town is being completed, the settlers should try, inasmuch as this is possible, to avoid communication and traffic with the Indians, or going to their towns . . . nor [should the settlers] allow the Indians to enter within the confines of the town until it is built and its defenses ready and houses built so that when the Indians see them they will be struck with admiration and will understand that the Spaniards are there to settle permanently and not temporarily. [The Spaniards] should be so feared that [the Indians] will not dare offend them, but . . . will respect them and desire their friendship. (Mundigo and Crouch 1977: 258) Although these ordinances of Philip II were not established until 1573, by which time Pardo’s Fort San Juan lay ruined, we should note that Philip’s

144 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

Figure 4.41. Geometry linking Structure 1, Structure 2, Structure 5, and Feature 76.

new regulations generally expanded and revised previous decrees and codified established colonial practice (Mundigo and Crouch 1977: 248). We suggest that such practices as stipulated here may have played a role in both the reorganization of the Berry site Spanish compound and its ultimate destruction. Like most other Spanish settlements in La Florida, this compound— Pardo’s Cuenca—appears to have been reorganized on a tight grid during its second phase of construction. Clearly, Structures 2 and 5 share the same alignment, but as illustrated in figure 4.41, this alignment seems to have been based on a preexisting relationship between Structure 1 and Feature 76, the outdoor hearth. The geometry of the grid probably began with a line drawn from just north of Structure 1’s entryway trenches to Feature 76; this line is 16.5 meters in length, or almost exactly 20 Castilian varas. Turning 90 degrees west on this line from Feature 76 and running another line the same distance gives us the location of Feature 91, the hearth at the center

The Built Environment of the Berry Site Spanish Compound · 145

of Structure 5. Turning 90 degrees north from this position and running another line the same distance again takes us to what should be, based on 1997 auger tests, the midpoint of Structure 2 and its presumed hearth. Feature 76, Feature 91, and the center of Structure 2 thus create an isosceles right triangle or half square. The corners of Structures 2 and 5 fall close to these same lines and were probably set in at the same time. Using such simple geometry—the same geometry that we archaeologists use to lay out our excavation units—the soldiers of Fort San Juan were able to orient Structures 2 and 5 both to each other and to Structure 1, creating a rectangular plaza that incorporated all three of these structures as well as Structure 3 (but not Structure 4; figure 4.42). This plaza is oriented 58 degrees west of north and measures 19 meters wide by 23 meters long. Additional evidence in support of our identification of this space as a plaza comes from the patterning of square postholes in the compound area. The

Figure 4.42. Plan map of second-phase Spanish compound with rectangular plaza indicated.

146 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

entire compound is marked by a proliferation of postholes (see plate 1), and while some of these may constitute fences or screens, it is difficult to identify clear patterns (Johgart 2011). One exception is the tightly grouped set of postholes along the eastern edge of the compound between Structures 1 and 3, which almost certainly constitutes some kind of structure. We cannot say whether this is a fragment of a larger building or is complete in itself (perhaps as a defensive work), but its shape and architectural grammar— the triple rows of posts, in particular—are unlike any other building in the compound. Another pattern, one that becomes key to our understanding of the plaza, is the way that square postholes cluster into distinct groups. The shape of a posthole can often be a good indicator of who created it. Indians likely did not dig square postholes, for example, because native peoples lacked square-bladed tools. Square postholes are uncommon in the Spanish compound (only twenty-three have been identified there), and many of these were reported as possible shovel test pits when first recorded,

Figure 4.43. Square postholes 1109 and 1176, before and after excavation (after Johgart 2011: 34, fig. 11).

The Built Environment of the Berry Site Spanish Compound · 147

Figure 4.44. Plan map of second-phase Spanish compound with plaza and square posthole clusters indicated.

because of their shape and their very mottled fill (Johgart 2011: 20). Five of these have since been excavated, however, with three of those five revealing round or square post molds (figure 4.43). Yet it was still difficult to know whether the square postholes as a class were associated with the Spanish compound or with much later (eighteenth- and nineteenth-century) farming and husbandry activities. Our identification of the plaza appears to clarify this problem. Abra Johgart’s analysis (2011: 36–37) shows that nineteen of the twenty-three square postholes cluster into three spatially distinct groups. When we add these clusters to a plan of the compound with the plaza indicated (figure 4.44), we find that all three posthole groups fall

148 · R. A. Beck, D. G. Moore, C. B. Rodning, S. Sherwood, and E. T. Horton

precisely along the western or southern sides of the plaza and help to define its boundaries. Because the middle of these groups has a post that intrudes on Feature 23—and because the intrusive post was visible at the top of the feature—we may conclude that this group postdates the feature, which was filled before the post was placed (Johgart 2011: 37). Indeed, it seems likely that all three posthole groups are associated with the second construction phase, and while the proliferation of postholes in this area hinders us from marking clearly defined structures, these may represent storage rooms or other less-durable buildings associated with the reorganized compound. If the reconfigured Spanish colony at the Berry site was based on a tightly plotted grid—as the Laws of the Indies would stipulate just six years later—then is there evidence for the exclusion of Indian peoples from the settlement during its reorganization, as would also come to be codified in the 1573 Laws? We have already suggested that the builders of Structure 5 drew on a different architectural grammar from those who built Structure 1. Although the form and style of Structure 1 are consistent with Late Mississippian practices across the southern Appalachians, save for the use of metal tools, the grammar of Structure 5 is not; indeed, many features of Structure 5 seem more consistent with European architectural grammars. The same is true of the probable storage rooms with square postholes identified on the western and southern sides of the plaza. These architectural data therefore suggest that native builders had a less direct role in the compound’s second construction phase; but does this also mean that Indians spent less time in the compound? Food remains provide one of our most compelling lines of evidence for investigating how Indians contributed to daily life in the Spanish compound, as well as how their contributions may have changed over time. Faunal data from the central pit features—that is, from the first period of occupation—indicate that complete deer carcasses were brought into the compound for processing (Lapham, chapter 7, this volume). Much of this butchering work would have been performed by native men and women, so that there must have been considerable daily traffic between the native and Spanish communities. During this phase, the soldiers consumed deer and bear meat in nearly equal quantities. Southern Indians often reserved bear meat for special or honored guests, and the central features contain few remains of bear processing waste; rather, Indian hunters regularly provisioned the soldiers with choice cuts of bear meat prepared elsewhere. Both of these patterns would change after the compound’s reconfiguration, when disposal areas west of Structure 5 were in use. Lapham’s work

The Built Environment of the Berry Site Spanish Compound · 149

shows that then deer meat, too, entered the compound in a prepared state. No longer, that is, were Indian hunters processing food outside the homes of their guests. Simultaneously, the quantity of bear meat entering the compound dropped sharply. It is interesting to note that Gayle Fritz’s analysis of plant remains (chapter 6, this volume) finds a simultaneous dietary change, as hickory nut overtook acorn as the most common nut in the western features. In this case, though, Indian women seem to have acted on a food preference of the soldiers rather than denying them a preferred food, as in the withholding of bear meat. We might conclude, then, that as Pardo’s soldiers began to reorganize their colony, they also began to limit access to the compound, specifically for Indian men. Indian men, in turn, began to withhold bear meat, a preferred food. Meanwhile, Indian women—whether from Joara or taken down from the mountains as hostages—changed plant dishes according to the soldiers’ preferences. Although the first phase of colonization under Moyano’s watch encouraged close links between the soldiers and the men of Joara, including participation in joint military ventures, the later occupation under Escudero instituted a more formal order that ultimately proved ill-suited to the circumstances and may have helped to bring about the destruction of Cuenca and Fort San Juan.

5 Wood Selection and Technology in Structures 1 and 5 Lee Ann Newsom

The Juan Pardo expedition left Santa Elena on the first of December, 1566, with the intent of penetrating the interior of the continent. Spanish goals were twofold: to establish viable relationships with inland groups of Native Americans and to reach the silver mines at Zacatecas, Mexico, having conceived that an efficient route across the Appalachians would be beneficial and was readily obtainable (Hudson 2005: 23). First and foremost, however, and by order of the marquis Pedro Menéndez de Avilés, captain-general of the Indies, this mission was intended to establish a greater presence in the region, both for the sake of controlling the Indian population and for creating a network of forts and way stops for secure passage. Had it been successful, this plan would have allowed regular traffic back and forth along the proposed route, an especially important goal considering what were at the time very mixed receptions and intermittent hostilities on the part of some of the indigenous peoples. To facilitate this operation, the expedition pressed pacified Indians and others encountered along the trek into service as guides and as builders of some of the infrastructure, particularly the safe houses intended to serve the European contingent and any others who would sooner or later travel the route (Hudson 2005: 141–143). Spanish accounts also suggest that Pardo and members of his company were expected to build forts at various points along the route as the principal nodes for protection, reprovisioning, and information transfer (Hudson 2005: 146). The Indian town of Joara proved pivotal to the Pardo expedition because of its location along established Indian trails leading into the mountains, as well as inherent demographic and sociopolitical dynamics among the local indigenous peoples. For these reasons, and as Pardo’s westward advance was halted in early 1567 because mountain passes were snowed in,

Wood Selection and Technology in Structures 1 and 5 · 151

Joara became the first place where Pardo built a garrison—christened Fort San Juan—in his planned chain of forts (Hudson 2005: 25, 146). There is no documented description of this fort, though it likely contained several buildings and defensive works. Moreover, because Fort San Juan was occupied for eighteen months, its internal structure and organization may have changed over the course of its occupation (see chapter 4). Excavations at the Berry site (31BK22) in Burke County, North Carolina, appear to confirm that this location was the site of Joara and Pardo’s Fort San Juan (Beck et al. 2006; see chapter 1, this volume). In this chapter, I focus on the wood and other construction materials that were used to build a pair of fairly substantial buildings at the settlement, the remains of which have been excavated and documented in the course of work at the site over several years. One central question concerns the nature and function of the architecture, specifically, whether the structures had previously been built for Native American purposes and uses and then later were given over to the Spanish visitors for their exclusive use, or whether one or both structures were built upon the arrival of the expedition for its exclusive use. This latter possibility presents three intriguing scenarios that I consider with respect to how the buildings were constructed: (1) they were built exclusively by the Joarans on behalf of the Spaniards; (2) they were built as a cooperative effort between the two groups; or (3) they were constructed solely by Pardo’s men using their own equipment and architectural knowledge. Distinguishing among these scenarios is important because it can help us to elucidate some of the initial social dynamics between the two groups, including perhaps some sense of the degree of coercion or cooperation between the Spaniards and their indigenous hosts. To this end, analysis of the materials used in these two buildings and close scrutiny of their construction details have the potential to illuminate culturally specific woodworking and architectural practices. These can provide insights both into how the structures were built and into the builders themselves. Analysis of the structures also has the potential to determine the actual time of construction, thus providing some ground-truthing of the ethnohistoric documents that recount the progress of the expedition. At least two courses of action and actors may thus be posited in considering construction of the two buildings. If one or both of the structures were constructed exclusively by the native Joarans, either before or after the arrival of the Spaniards, then the architecture should reflect traditional (that is, native) construction practices. In this case, the buildings should be consistent with the types of indigenous construction that occurred at any

152 · Lee Ann Newsom

time of the year but would more than likely reflect cool-season single- or multifamily dwellings. Many such buildings appear to have had circular (Hudson 2005: 42, 143) or roughly square (Hally 1994a: 236–237; Hally 2008) floor plans, with wattle and daub constructions framing the sides and sometimes the roof; they would also have been well insulated with bark, moss, and other organic packing materials. If, instead, these buildings represent a cooperative effort between the Spaniards and the Joarans or else an effort achieved exclusively by members of the expedition, then such buildings may reasonably be expected to deviate in particular details of form and construction practice from strictly indigenous styles. Some implications of the above three scenarios that are relevant to this study are clear. If the buildings were constructed exclusively by the Joarans using their traditional knowledge, tools, and practices, then the particular woods used and how they were deployed and conformed should be consistent with evidence from prehistoric or otherwise exclusively Native American buildings; this should also extend to secondary and tertiary construction elements (for example, wall/wattle structure and organic insulation materials, respectively). In this case, all tool marks and the reduction of raw material such as wood and bark would be consistent with indigenous technology, particularly with the stone axes, adzes, and other woodworking implements that were part of the traditional tool kit. To this end, I draw upon a body of knowledge about Native American construction practices for comparative purposes that was previously generated by William Bartram (1995 [1791]), David Hally (2008), Cameron Lacquement and colleagues (Lacquement, ed. 2007), and others. Alternatively, if the building effort was cooperative, with the native Joarans working in concert with the Spaniards, then evidence of hybrid technology and practices—including at least some signs of the use of European carpentry tools—might be expected, particularly implements with metal blades or bits. Finally, if they were solely constructed by expedition members, then the buildings should deviate from and contrast (at least to some degree and perhaps considerably) with native constructions. Hypothetically, this could take the form of material selection, material reduction, joining, tool marks, and other carpentry details consistent with European practices of the period for which there exists a large body of literature. Especially relevant for the latter two scenarios is mention in the Pardo documents of European carpentry tools such as metal axes, chisels, nails, and saws that outfitted the expedition (Hudson 2005: 45, 149–150, 152). The length of time that the buildings were meant to last is another ques-

Wood Selection and Technology in Structures 1 and 5 · 153

tion, one that would have affected the quality and durability of construction if such was a concern for the actual builders. This is a difficult problem to approach while focusing exclusively on the remnants of the two buildings that are examined here, and indeed answering it may be beyond the reach of the subject data. Nevertheless, it does have the potential to shed important light on the mindsets of the builders and perhaps on the nature of cultural interactions at the site. It is therefore a question that I consider here. It seems reasonable to assume that the buildings were intended to be sufficiently durable to serve for at least a couple of years, given Pardo’s overall plans to explore the region and to establish a chain of garrisons. Moreover, Menéndez’s master plans for an overland link to the distant silver mines in Zacatecas implied an evolving infrastructure of inns, forts, and way stations that would have been in place for an indefinite but presumably lengthy period of time. Yet Pardo’s soldiers, as well as the Joarans themselves, whether building separately or together, may have had other ideas and concerns about the longevity of the buildings at the time of their construction. Excavations have revealed that the first of the two structures, designated Structure 1, had a relatively deep and well-defined floor basin and that the building itself was correspondingly a fairly substantial structure (see chapter 4). About 60 percent of Structure 1 was excavated during the 2003, 2004, and 2008 field seasons, and its dimensions on the ground are determined to be approximately 56 square meters (see plate 7). The building’s footprint indicates that it was square in floor plan, with three large central support posts (presumably three out of four total, the fourth as yet unexcavated) and a clearly defined entryway. These characteristics are all consistent with what is known of traditional Native American structures at other Lamar towns such as the King site (Hally 2008). Beck et al. (chapter 4, this volume) infer that Structure 1 may have been among the first built in the Spanish compound, and they suggest that its construction may have involved the full cooperation and labor of native Joarans. Structure 1 was thus hypothetically built at some time in the winter or spring of 1567, along with perhaps two other buildings, Structures 3 and 4. In contrast to Structure 1, the field evidence for Structure 5 (see plate 2) suggests that it was less sturdy and was more expediently constructed (see chapter 4). This building has been completely excavated and encompasses about 50 square meters of floor space, which means that Structure 5 was the smallest of the buildings that appear to have constituted the Spanish compound. Structure 5 is further distinguished by the lack of a clear floor basin and large, deep central posts. Indeed, at least two and perhaps three of the

154 · Lee Ann Newsom

central support posts were evidently insubstantial enough to have required a second support post. The perimeter posts either surrounded a very shallow basin or else the floor was simply the unprepared ground surface. One hypothetical explanation for Structure 5’s distinguishing characteristics is that it was built later in the year, during the warmer months, when lighter structures were traditionally built. Another possibility is that this building was created exclusively by Pardo’s men when the expedition returned in the fall of 1567. The construction was done perhaps under some duress and possibly rather quickly. Regardless of time or season, the construction details for such a structure would theoretically bear little relationship to traditional native southeastern practices. Structure 5 therefore has the potential to provide an evidential data set for a European structure, however expediently and/or expeditiously done, in contrast to a building constructed exclusively by Native Americans or as a cooperative effort between the two groups. In addition to the buildings with their structural components, one additional set of materials that was incorporated in this analysis is fuelwood remains from hearth features located within the confines of the two structures. These remains have relevance as a separate assemblage of woody materials that represents a distinct purpose (such as cooking and heating) and thus may have distinguishing characteristics that further serve to highlight wood selection and use. This, too, has interesting alternative possibilities and—like building construction—may reflect a dynamic situation. Initially, at least, it seems likely that the Joarans, as good hosts who welcomed the Spaniards, would have daily supplied their European visitors with both food and fuelwood, since they were most familiar with the landscape, its array and abundance of wild food resources, and its standing stock of woods suitable for construction raw materials and fuel. As time passed, and as the Europeans began to overstay their welcome, these foreigners might well have had to forage on their own for wood to burn or materials for repairs. It is reasonable to hypothesize that in this case selection would have been less specific and more skewed toward whatever was most readily available and easily gathered. This would be especially likely if relations had begun to sour and it was unsafe for the Spaniards to be out at a distance pursuing wood for daily needs. If such was the case, then charcoal residues in the Structure 5 hearth would likely be dominated by low-value woods and those from secondary forest regrowth and disturbed habitats—particularly ash, cherry, and small-diameter shrubs—rather than high-quality fuels such as oak, hickory, and chestnut.

Plate 1. Plan map of the Berry site Spanish compound, showing all buildings, features, and postholes.

Plate 2. Photomosaic of Structure 5, base of plow zone.

Plate 3. Photomosaic of Structure 5, top of Zone 3.

Plate 4. Plan map of all organic remains in Zone 3, Structure 5. Color code: blue, in situ posts; brown, timbers; yellow, rivercane; gray, small wood.

Plate 5. Photomosaic of Structure 5, top of Zone 4/subsoil.

Plate 6. Removing plow zone above Structure 1, 2008.

Plate 7. Photomosaic of Structure 1, top of Zone 4.

Plate 8. Plan map of all organic remains in Zone 4, Structure 1. Color code: blue, in situ posts; brown, timbers; yellow, rivercane; gray, small wood; green, thatch; purple, bark; orange, plank removed in 2003 (OG#2003a); red, outline of basin cut.

Plate 9. Plan map of organic remains in Zone 4, Structure 1, with small wood removed.

Plate 10. Photomosaic of Structure 1, top of subsoil.

Plate 11. Thin section of sediments from Structure 1, showing clay papules added through levigation: a, plane polarized light; b, cross-polarized light.

Plate 12. Thin section of sediments from hearth area in Structure 1, showing clay papules added through levigation: a, plane polarized light; b, reflected polarized light.

Wood Selection and Technology in Structures 1 and 5 · 155

Laboratory Analysis and Data Collection This research involved two sets of samples consisting of wood and related materials recovered from the two buildings. The field-designated building elements include upright posts, other larger timbers, and a variety of what were probably secondary (for example, rafters and minor wall supports) or tertiary (for example, cane wattle siding/framing, bark siding or roofing material, and packing and insulation between roofing or wall timbers) elements employed in building construction. The main sample set consists of the remnants of the timber frames and related construction debris from the two structures, including 74 samples from Structure 1 (figure 5.1) and 122 from Structure 5 (figure 5.2). In addition, eight samples of ter-

Figure 5.1. Structure 1 floor plan with all individual organic samples (OGs) plotted (see plate 8, table 5.2, and table 5.4 for identification of sample types; OG#87 and OG#88 were not assigned).

156 · Lee Ann Newsom

Figure 5.2. Structure 5 floor plan with all individual organic samples (OGs) plotted (see plate 4, table 5.3, and table 5.5 for identification of sample types).

tiary construction materials were analyzed, including one from Structure 1 and seven from Structure 5. The second major sample set represents other wood uses, specifically fuelwood. The latter is based on charcoal remnants recovered from hearth features inside the two buildings. This includes three samples from Feature 119 of Structure 1 and five samples from Feature 91 (OG#60 series) of Structure 5. The excavation teams very expertly exposed and mapped, then quite successfully wrapped and removed, a variety of individual timbers and other elements. Their careful efforts have made it possible for me to work with relatively intact construction members and thus enabled my docu-

Wood Selection and Technology in Structures 1 and 5 · 157

mentation of complete or reconstructed sizes and forms, as well as my examination of individual specimens for retention of bark, terminal growth increment—and thus potentially the season of harvest—tool marks, and other details. For the purposes of this analysis, the individual wood items recovered were classified according to five basic form categories (partly in conjunction with the field designations): upright posts, roundwood, lumber, secondary elements, and tertiary elements. Roundwood is a descriptor borrowed from European archaeology that is used to describe radial (that is, cylindrical) segments of branch or stem from trees and shrubs; the individual segments can be of any length and diameter. All of the upright posts are also encompassed within the roundwood category, as are many other timbers from the site. Each roundwood specimen was further classified according to its relative completeness with respect to the original circumference, whether whole (full circumference), half, or finer segments. Whenever possible I sought to refit or match separated fragments of original roundwood segments by comparing the broken ends and/or the growth ring patterns, and for reasons specified later in the chapter I analyzed the posts both separately and together with the rest of the roundwood. Likewise, some analysis focuses exclusively on the nonpost roundwood assemblages. Lumber technically refers to a timber that has been shaped into a plank, or, more precisely, into “an elongated, rectangular, solid piece of wood that has been separated from the log by sawing” (Hoadley 2000: 213). Formal lumber classifications include three main categories: board, dimension, and timber (Hoadley 2000: 218). Boards equate with lumber less than 5 centimeters in nominal thickness and 5 centimeters or more in nominal width; those less than 15 centimeters wide may be classified as strips. The dimension category includes lumber from 5 centimeters up to (but not including) 12.5 centimeters in nominal thickness and 5 centimeters or more in nominal width; such lumber may be designated as framing, joists, planks, rafters, studs, or small timbers. Finally, the timber group encompasses lumber nominally 12.5 centimeters or more in least dimension and may be designated as beams, stringers, posts, or girders. In this analysis I use plank-like in referring to a set of items that would technically fit the board category, although one timber closely approaches the dimension category. I use slat-like for other specimens that would fall into the board category, specifically as strips. The difference is a matter of scale, with the slats or strips being similar in form (that is, flat and rectangular) but of much finer size than the other board or plank-like items.

158 · Lee Ann Newsom

Secondary construction elements consist of the smaller-diameter roundwood that served to complete the structural framework, including roof supports, rafters, secondary wall supports, and framing. Tertiary construction elements include all finishing materials, such as wattle frames, bark siding, roofing, and moss or other materials used to insulate between timbers and seal spaces between elements. Wattle can overlap with roundwood in the secondary construction category, but otherwise it includes the stems of robust grasses, specifically, the native cane Arundinaria gigantea. Any additional siding, roofing, or insulation may consist of a variety of materials potentially used to shield against wind and weather. Ethnographic records suggest that this may include bark and moss. The overall dimensions of each recovered item were recorded, along with the form category as indicated above and any other directly observable features such as the preservation state (some specimens were uncarbonized). The retention of bark and any other features that may signify the presence of the terminal growth increment (such as bark beetle channels) was noted, and whenever possible the number of growth increments was counted to arrive at an age estimate. When both pith and bark are present— providing the complete transverse profile—it is possible to determine or at least to estimate closely the age at harvest, taking care to discriminate missing and false growth increments that tend to correlate with advanced age (senescent growth) and/or particular taxa (for example, pine). Obviously those specimens with such indications of the terminal stage of growth provide for a more reliable age estimate, and they may also be used to specify the season of harvest given sufficient preservation of the cell structure. For roundwood in the Berry site assemblage that lacked bark or that had clearly truncated rings, I did not generally attempt to estimate the number of any missing sapwood rings for a rougher estimate of age; for the one type of conifer wood that is present in the assemblage, I took care to distinguish false from true growth increments. Each item was also carefully scrutinized for the presence of tool marks and any other suggestions of the specific reduction/carpentry techniques used in its procurement and preparation. Two basic types of tool marks were recognized. Facets are prismatic-shaped marks, typically longer than wide, that generally slope upward toward the end of the specimen. Depending on the tool bit used, they can be relatively flat surfaced or concave in form. Bevels are sloping surfaces or edges that generally result from straight angular cuts and are commonly wider than long; these likewise may be quite flat or slightly concave depending on the tool used.

Wood Selection and Technology in Structures 1 and 5 · 159

All specimens were identified according to three-dimensional anatomy, with carbonized wood specimens examined under magnification using incident light and uncarbonized material being thin sectioned and then compared using transmitted light microscopy. Each specimen was assigned to the lowest possible taxonomic rank using a variety of resources, including published wood identification manuals (Panshin and de Zeeuw 1980), pertinent floras (Radford et al. 1968), online resources (InsideWood 2004–; Wheeler et al. 1986), and a comparative collection housed in the Environmental Archaeology Laboratory at Pennsylvania State University.

The Composition of Structures 1 and 5 Wood Types Altogether, eight types of wood were identified in the course of this analysis; all of them are listed in table 5.1. Among the woods identified is a single conifer or “softwood,” a pine (Pinus sp.) subgeneric belonging to the Taeda anatomical group (table 5.1). This group includes all of the southern “hard” or “yellow” pines that collectively are common constituents of the Appalachian and Blue Ridge physiographic provinces, particularly the forests of the subtending Piedmont and Coastal Plain. Examples include the scrub pine (P. virginiana) and the short-leaf pine (P. echinata) of the western portions of the Carolinas (Radford et al. 1968: 37–38), as well as the longleaf pine (P. palustrus) and slash pine (P. elliottii) that dominate much of the lower Southeast. The Taeda group does not include the eastern white pine, P. strobus, which eventually became very popular with European and Table 5.1. Archaeological wood taxa from Structures 1 and 5 Common name

Taxonomic assignment

Family

Yellow or “hard” pine American chestnut Ash Black locust Hickory, pecan group Hickory, true group Oak, red/black group Oak, white group Ring porous hardwood

Pinus sp., subgenus Pinus, section Pinus Castanea dentata Fraxinus sp. Robinia pseudoacacia Carya sp. Carya sp. Quercus sp. Quercus sp. catawba tree (Catalpa sp.) or Osage-orange (Maclura pomifera)

Pinaceae Fagaceae Oleaceae Fabaceae Juglandaceae Juglandaceae Fagaceae Fagaceae Bignoniaceae or Moraceae

160 · Lee Ann Newsom

American shipbuilders (Record and Hess 1943). The pine wood in Structures 1 and 5 was recovered in both carbonized and fully uncarbonized degraded condition, the latter form occurring in the centers and/or bottom portions of upright posts, especially the largest ones. Pine wood was also identified as fine carbon particulate microdebris found among floor deposits in conjunction with Sherwood’s analyses (chapter 4, this volume) and may represent microdebris from hearths. Seven hardwood taxa were identified among the Berry samples analyzed: American chestnut (Castanea dentata), ash (Fraxinus sp.), black locust (Robinia pseudoacacia), two types of hickory (Carya spp.), and two types of oak (Quercus spp.).1 Hickory wood may be assigned to either of two anatomical groups (Panshin and de Zeeuw 1980), both of which are represented in this assemblage. These are the “pecan” and “true” anatomical groups, and in the tables and sections below I refer to them as pecan and true hickory, respectively (or “Carya pecan” and “Carya true”). All but two species of Carya in this region belong to the latter group, including the pignut (C. glabra). The pecan group encompasses the other two species, specifically, water hickory (Carya aquatica) and pecan (C. illinoinensis). The water hickory, as the name implies, is a wetland species that is most common in swamps and floodplains in the eastern portion of the state today, and its nuts are bitter (Radford et al. 1968: 364–365). Pecan’s natural range is likewise not generally understood to encompass western North Carolina, though Albert Radford and colleagues (1968: 363) indicate that it may occur in “low woods” as an escape in North and South Carolina; otherwise, it is (or was) found “scattered” throughout the Piedmont and Coastal Plain. Radford and colleagues’ (1968: 364) maps indicate that records exist for the presence of pecan at least as far west as central North Carolina. In any event, this anatomical group is present among the wood samples from the Berry site. Interestingly, Gayle Fritz (chapter 6, this volume) reports the presence of relatively thin Carya nutshell fragments from Feature 112, alluding to the correspondence with the pecan hickory group, but cautiously suggests that these nutshells may represent one of the other taxa from the true hickory group. One possibility to consider is that both the thin nutshell and the pecan wood group indeed represent pecan, which in turn suggests that either the species had a more expansive natural range in the past or else pecan trees were planted and cultivated beyond their original range under human management. Oak wood classification also involves separation into anatomical groups,

Wood Selection and Technology in Structures 1 and 5 · 161

in this case ones that roughly approximate the finer rank of section within the subgenus Quercus, culminating in three oak taxa from the standpoint of wood anatomy. The many species of the so-called red and black oaks, all of which belong to section Lobatae, together constitute the “red oak” wood anatomical group. Similarly, section Quercus equates largely with the “white oak” wood anatomical group, which comprises multiple species; among these, the live oaks (such as Q. virginiana) may be further distinguished based on their characteristic anatomy. Both the red and white oak groups are represented in the Berry site assemblage, yet the live oak subgroup is absent. In the sections that follow, the oak taxa are simply termed red oak and white oak, and in the tables the two groups are indicated as “ROG” and “WOG.” Building Components A large assemblage of primary support members and various other wood construction elements was analyzed in this study, including 74 samples from Structure 1 and 122 samples from Structure 5. Among these were the remnants of upright posts, other larger roundwood timbers, and various smaller-diameter roundwood. At least 10 items in the lumber category were also recognized. Seven wood taxa were identified in Structure 1, including American chestnut, black locust, true hickory, pecan hickory, pine, red oak, and white oak. The same set of woods was identified in Structure 5, with the addition of a single specimen of another hardwood, ash. Red oak proved to be the most commonly used wood overall for the construction of both buildings, constituting 36 percent of the identified wood in Structure 1 and 47 percent of the identified wood in Structure 5. Pine was the second most abundant wood type, constituting 24 percent of the Structure 1 assemblage and 15 percent of that from Structure 5. Third in frequency for both structures was white oak (22 percent in Structure 1, 13 percent in Structure 5), followed by chestnut for Structure 1 and true hickory for Structure 5, both at 10 percent of their respective sample sets. Chestnut also constituted 7 percent of the Structure 5 wood assemblage. Black locust and pecan hickory were present in minor amounts in the samples from both structures, each constituting 3 percent of identified woods, which is likewise the proportionate representation of true hickory in Structure 1. Ash was identified only in Structure 5, from a sample designated “cane and wood mass” (OG#33); it may thus have originated as fuelwood or else represents something other than a construction element.

Table 5.2. Upright posts from Structure 1 OG# 6 10

Taxona

Accn. #

Unit

Zone

Field designation

Bark

048-1031 048-1584

7 7

4 4

SW upright post NE upright post

no no

WOG Castanea

11

048-894

2

4

upright post

no

Robinia

21

048-1553

40

4

center post

no

Pinus

Radius (cm)

Diameter (cm)

4.3 2.5

8.6b 5.0

10.0

17.0

3.9

7.7b

Ring count

5.4 56

29

048-1279

17

5

upright post

yes

Castanea

32b

048-1101

5

5

upright post

no

Pinus

13.6

41

048-1572

50

4

upright post

no

Castanea

11.8

42

048-1596

50

4

upright post

no

ROG

7.2

47

048-1253

14

4

upright post

no

Pinus

53

048-1527

38

5

upright post

no

ROG

54

048-1528

38

5

upright post

no

Castanea

9.2

60

048-1361

20

5

upright post

no

Pinus

12.3

61

048-1561

40

5

post 16

no

Pinus

12.5

63

048-1569

19

5

upright post

no

ROG

72

048-1575

20

5

center post 18

no

Pinus

14.0

82

048-1580

26

5

upright post

no

Castanea

8.5

16

85

048-1597

72

4

upright post

no

Robinia

7.3

15

93

048-1592

65

5

center post

no

Pinus

16.0

73

7.2

13.0

86

100

4.1

6.7

5.6

13.4b

18 samples total Average diameter of 18 samples:c

a

ROG=red oak anatomical group; WOG=white oak anatomical group. Data estimated from radius measurements. c Standard deviation: 3.86. b

10.4 17.0

max.

4.1

min.

Table 5.3. Upright posts from Structure 5 Radius (cm)

Diameter (cm)

Bark

Taxona

upright post

no

Robinia

9.3

timber, wood from posthole

no

ROG

9.6

2

upright post

no

WOG

2

4

upright post #2

no

Pinus

46

PH#7

upright post

no

Maclura

4

OG#

Accn. #

Unit

Zone

Field designation

52

042-1756

29

F.94

53

042-1683

F.92

58

042-1383

51

59

042-1756

61

042-1501

62

042-1680

6

63

042-953

15

64

042-965

16

65

042-966

16

5.7

11.4b

6.7

10.4

upright post

no

Pinus

upright post

no

Castanea

10.4

4

upright post

no

Pinus

8.0

4

upright post

no

Pinus

66

042-1499

4

4

67

042-1494

43

PH#5

76

048-1593

wood from PH#6

no

Pinus

upright post

no

Robinia

burned wood from PH#27 N½

no

Castanea

8.5

17.0b

Ring count

47

12 samples total Average diameter of 7 samples:c

a

ROG=red oak anatomical group; WOG=white oak anatomical group. Data estimated from radius measurements. c Standard deviation: 2.90. b

10.87 17.0

max.

8.0

min.

164 · Lee Ann Newsom

With the exception of the hickories and ash, all of the above-mentioned taxa were variously used as upright posts for the two buildings (see tables 5.2 and 5.3). The eighteen analyzed upright posts from Structure 1 indicate an emphasis on pine, which makes up nearly 40 percent of the post assemblage. This includes the three recovered center posts, which were the most substantial such elements (in terms of diameter) used for this building (table 5.2). Pine is followed by chestnut (28 percent) and red oak (17 percent) in the assemblage of upright posts. Essentially the same pattern of species selection for posts is true of Structure 5, with pine being the majority (42 percent) of upright members (n=12), followed by chestnut (17 percent) (table 5.3). One pine post from Structure 5 (OG#65) is of a similar diameter class as the center posts for Structure 1. To the extent that the data are representative, this shared pattern of wood selection for the major support elements—emphasis on pine and chestnut with only minor use of the other taxa—is also reflected in the age classes of the material selected. Age estimates for six of the posts from Structure 1 span 15–100 years; four pine posts were the oldest, individually aged to 56, 73, 86, and 100 years (OG#21, OG#93, OG#32b, and OG#47, respectively [table 5.2]). The 73-year-old post (OG#93) was harvested from a leaning tree or large branch, based on the presence of compression wood anatomy (see Barnett and Jeronimidis 2003), but this would not have adversely affected its employment as one of several major support posts. Only one upright from Structure 5 could be reliably age estimated, this being the aforementioned OG#65, a 47-year-old pine post from Unit 16. In general, post diameters are fairly consistent between the two structures, ranging from around 4 centimeters to 17 centimeters, with an average of 10–11 centimeters (tables 5.2 and 5.3). To reiterate, the largest posts—including the center posts for Structure 1—are pine, which thus served as the primary structural support for both buildings. These largest posts include the four older ones mentioned for Structure 1, with diameters ranging from 13 centimeters to 17 centimeters. The pine post from Structure 5, Unit 16, has the largest diameter (17 centimeters) among those recorded for that structure, with the girth of other pine, black locust, and chestnut posts ranging from 8 centimeters to 11 centimeters (table 5.3). In contrast, nonpost wood from the two structures—mostly roundwood that presumably represents secondary construction elements such as roofing and minor wall members (see tables 5.4 and 5.5)—is dominated by oak, especially red oak. More than half (52 percent) of nonpost roundwood from Structure 5 (n=110), derived from trees of the red oak anatomical

Table 5.4. Nonpost wood samples from Structure 1 OG#

Field Unit Zone designation

Bark

Taxona

Formb

Segmentc

Radius (cm)

Diameter (cm)

Length (cm)

Width (cm)

Comments

1

7

4

wood

no

WOG

RW

1/2

multiple fragments in linear alignment, all suggestive of one original; appearing cut or split down length

2

7

4

wood

no

WOG

RW

1/2

appearing cut or split down midline (2 photos)

3

7

4

wood

no

WOG

RW

1/2

4

7

4

wood

yes

WOG

RW

1/2

3.4

25.0

multiple fragments in linear alignment, all suggestive of one original; appears cut or split down midline; one end is clearly faceted, shaped into a point (drawn)

23.5

roundwood in 5+ aligned segments; flat on one side, appears cut or split down midline (continued)

Table 5.4—Continued

OG#

Field Unit Zone designation

Bark

Taxona

Formb

Segmentc

Radius (cm)

Diameter (cm)

Length (cm)

Width (cm)

12.0

3.5

Comments

7

7

4

wood

no

WOG

RW

1

15

13

6

partly burned timber, E½

no

Pinus

RW

1

16

13

6

partially burned timber, W½

no

Pinus

RW

1

14.0

27.0

partially carbonized; unburned material set aside for sectioning radial; much of center rotted away

19a

40

4

wood mass sample

yes

ROG

RW

1/2

3.5

4.0

OG#19 sample 1; roundwood broken into 3 fragments (now 4)

19b

40

4

wood mass sample

yes

ROG

RW

1/8

1.6

OG#19 sample 2; partial section with branch base (knot), bark attached

19c

40

4

wood mass sample

no

WOG

RW

1/8

2.3

OG#19 sample 3, #1

19d

40

4

wood mass sample

no

WOG

RW

1/4

2.4

OG#19 sample 3, #2

blocky charcoal in linear alignment; 2 sides are very flat, appears shaped or split

19e

40

4

wood mass sample

yes

ROG

RW

1/8

19f

40

4

wood mass sample

no

WOG

RW

1/16

19g

40

4

wood mass sample

no

WOG

RW

1/8

19h

40

4

wood mass sample

no

WOG

RW

19i

40

4

wood mass sample

no

Carya, pecan

RW

20a

40

4

small wood

no

ROG

RW

1.8

OG#19 sample 3, #3 OG#19 sample 3, #4, probably goes with #1 and/or #2 above

2.1

OG#19 sample 3, #5 OG#19 sample 3, #6 2.4

1

3.2

3.4

OG#19 sample 3, #7; partial roundwood segment with tool marks (vertical facets exterior) (10 photos) OG#20 sample 1, roundwood in 3 vertically aligned segments (central set in group photo [5 group photos]), plus additional fragments, n=26 total; 1 fragment appears to have clear bevel cut (2 photos)

(continued)

Table 5.4—Continued

OG#

Field Unit Zone designation

Bark

Taxona

Formb

Segmentc

20b

40

4

small wood

no

ROG

RW

1/2

20c

40

4

small wood

no

ROG

RW

1/2

21b

40

4

wood packed around center post

yes

ROG

RW

1

Radius (cm)

Diameter (cm)

Length (cm)

5.7

Width (cm)

Comments OG#20 sample 2, roundwood left of above set (or below in group photos), may all equal one original; appears to have been cut down center/length, passing through pith & lateral knot also appears to have been cut (trimming secondary branches?) OG#20 sample 3, 2nd large roundwood to right (above in group photos) of above two and may go with above; 10+ additional fragments go with this (and/or above)

1.8

1.8

OG#21 bag 1 sample 1

21c

40

4

wood packed around center post

no

ROG

RW

21d

40

4

wood packed around center post

no

Carya, pecan

RW

24

12

4

small-diameter wood with bark

yes

Pinus

RW

1

10.5

13.0

27

12

4

small-diameter wood

yes

WOG

RW

1/8

2.4

20.0

roundwood broken into 7+ segments

31

13

4

large timber

no

Pinus

RW

1/2

2.6

4.7

60.0

roundwood broken into several segments

32a

10

4

large timber

no

Castanea

RW

1

9.8

13.3

pith present

1

1.5

3.8

3.1

OG#21 bag 1 sample 2

2.1

OG#21 bag 1 sample 3

(continued)

Table 5.4—Continued

OG#

Field Unit Zone designation

Bark

Taxona

Formb

Segmentc

Radius (cm)

Diameter (cm)

Length (cm)

Width (cm)

Comments

35

10

4

timber

no

WOG

RW

1

11.0

31.0

nice large timber, fully burned, breaking up into large segments; lengths of 3 largest segments, 63.02 mm, 139.40 mm, 110.38 mm = 312.8 mm or ~31.3 cm long

38

6

3–4

horizontal timber

no

Pinus

RW

1

10.5

35.0

at least 27 additional fragments, all pine and probably all part of original timber

43

50

4

timber

yes

ROG

planklike

—

28.0

13.0

narrowed at one end (8 cm wide); squarish/angular form overall due to loss or removal of wood along one side

44

50

4

timber

yes

ROG

planklike

—

30.0

12.0

large timber similar to above, likewise possibly shaped and narrowing toward one end (6.5 cm wide) (same tree?)

51

21

3

timber

no

Pinus

RW

1

5.9

19.0

roundwood in 2 parts, central core and a branch trace are uncarbonized (2 photos)

55

38

4

timber

yes

ROG

RW

1

4.4

40.0

multiple fragments of large timber in linear alignment

58

20

4

small wood

no

WOG

RW

1

59

20

4

small wood

no

ROG

RW

1/2

3.0d

10.2

2 roundwood segments that refit (2 photos), possibly cut or split down centerline

62

19

4

bark (wood)

no

Pinus

RW

1

5.4

15.0

roundwood, bark absent

64

19

4

timber

no

ROG

RW

1

7.5

24.0

large timber with possible facets/shaping at both ends and possibly also along tangential surfaces (bark removal)

20+ small WOG roundwood fragments in linear alignment and probably from one original

(continued)

Table 5.4—Continued

OG#

Field Unit Zone designation

Bark

Taxona

Formb

Segmentc

Radius (cm)

Diameter (cm)

Length (cm)

Width (cm)

Comments

65

19

4

notched timber

yes

Pinus

RW

1

7.6d

40.0

roundwood cut/ trimmed on both ends: one with narrowing facets, the other truncated with bevel cuts; flares at both ends, diameter is average of 3 along length

66

19

4

timber

yes

Pinus

RW

1

10.2

26.0

timber in two parts/ sections; one surface appears cut, trimmed

67

19

4

small wood

yes

WOG

RW

1/3

3.2

7.0

roundwood possibly cut/split down length

68

19

4

small wood

no

WOG

RW

1

2.4

10.3

73

41

4

timber

yes

ROG

RW

1/2

5.5

5.5

roundwood possibly cut/split down length

74

41

4

timber

yes

ROG

RW

1/2

6.9

22.0

roundwood possibly cut/split down length

75

41

4

timber

no

Carya, true

RW

1

9.0

28.0

roundwood, narrowing toward one end (6.5 cm diameter)

76a

41

4

small wood

no

ROG

RW

1/4

2.5

2.8

OG#76 sample 1

76b

41

4

small wood

no

ROG

RW

1/4

76c

41

4

small wood

no

Carya, true

RW

1/2

76d

41

4

small wood

no

ROG

RW

78

38

4

timber

yes

ROG

RW

1/2

4.5

32.0

large roundwood timber in several fragments; one end appears sharply cut on diagonal across full circumference (4 photos)

79

38 & 50

4

timber

yes

ROG

RW

1

4.0

49.0

large radial, partly lacking heart

2.8

OG#76 sample 2 3.9

OG#76 sample 3 OG#76 sample 4, additional ROG charcoal fragments, probably all goes together with samples 1 and 2

(continued)

Table 5.4—Continued

OG#

Field Unit Zone designation

Bark

Taxona

Formb

Segmentc

81

26

3

timber

no

ROG

RW

1

84

38

4

“doorway wood”

no

ROG

planklike

—

90

50

4

small wood

no

ROG

RW

1/4

91

72

4

timber

no

Pinus

RW

1

92

26

5

wood

no

Pinus

RW

1/8

Radius (cm)

Diameter (cm)

Length (cm)

11.0

17.0

3.3

Comments large roundwood with sharply angled, apparently cut end, the cut extending fully across circumference, and there is also a long narrow apparent tool mark extending across the central portion that suggests one or more saw marks (several photos)

roundwood broken into 10 segments, plus additional smaller fragments; may be cut or split down length 7.6

5.4

Width (cm)

36.5

interior unburned, thin sectioned

2003a

18

4

massive timber

no

Castanea

planklike

—

2003b

18

4

timber goes w/above

no

Castanea

planklike

—

76.2

24

thickness ~3–4 cm; tangential plank, growth rings parallel with wide axis (tangential surface exposed on broad face), indicates plane sawn timber (saw marks, if present, obscured) thickness ~3 cm (2 photos); sample overlain with mass of cane

56 samples total Average diameter of 25 samples:e

a

6.6 14.0

max.

2.4

min.

ROG=red oak anatomical group; WOG=white oak anatomical group; two anatomical groups of hickory include “true” and “pecan”; HW=hardwood indeterminate. RW=roundwood (radial, branch or stem); “plank-like” refers to a rectangular, thin, narrow piece. c Proportion of full circle/circumference (1=whole, 1/2=half, etc.). d Data estimated from radius measurements. e Standard deviation: 3.12. b

Table 5.5. Nonpost wood samples from Structure 5 OG#

Unit

Zone

Field designation

Bark

Taxona

Formb Segmentc

Radius Diameter (cm) (cm)

Length (cm)

1

2

2

wood

yes

WOG

RW

1/2

1.3

9.0

7

11

2

notched wood

no

WOG

RW

1/2

2.3

6.5

8

11

2

no

WOG

RW

9

11

2

yes

Pinus

RW

12

10

2

no

ROG

RW

13

10

2

no

ROG

RW

14

10

2

yes

ROG

RW

1/2

15

10

2

no

ROG

RW

1/8

19

14

2

burned wood possible bark burned wood burned wood burned wood burned wood small wood

yes

WOG

RW

21a

10

F.80

timber

yes

ROG

RW

Width (cm)

notch in one end just above a knot; uncertain whether deliberate or superficial (3 photos)

12.0

1/2

1/3

2.4

3

Comments

6.5

bark plus very thin layer of wood associated w/OG#13

5.0

associated w/OG#12

5.0

associated w/OG#15 associated w/OG#14

1.8

3.6d

6.0

1.8

3.6d

17.0

4.5

OG#21 sample 1, largest specimen (timber); may be cut/split down length along one or both inner (radial) surfaces but gives the impression of having been cut down the length (10 photos)

21b

10

F.80

small wood

yes

ROG

RW

1

21c

10

F.80

small wood

no

ROG

RW

1/3

21d

10

F.80

small wood

yes

ROG

RW

21e

10

F.80

small wood

yes

ROG

RW

2.2

2.4

9.5

OG#21 sample 2, breaks at both ends across grain could be natural OG#21 sample 2, 19 MNI, 8 larger fragments with relatively sharp angles as Lab Sample #1, suggesting cut or split surfaces; 1 specimen (illustrated, measured; in 2 fragments, cross mended) appears to have a beveled end & facets on outer surfaces (but uncertain due to charring) OG#21 sample 3, roundwood fragments, n=3 (plus additional smaller fragments), plus loose bark (n=15, probably oak) OG#21 sample 4, apparent roundwood, n=4 plus 5 fragments of bark, some still covering, loosely adhering to wood surfaces

(continued)

Table 5.5—Continued

Radius Diameter (cm) (cm)

Length (cm)

Unit

Zone

Field designation

Bark

Taxona

21f

10

F.80

small wood

yes

ROG

RW

21f

10

F.80

small wood

yes

ROG

RW

21f

10

F.80

small wood

yes

ROG

RW

1

21g

10

F.80

small wood

yes

ROG

RW

1

21h

10

F.80

small wood

yes

ROG

RW

1

1.7

6.8

21h

10

F.80

small wood

no

ROG

RW

1/2

2.3

10.7

OG#

Formb Segmentc 1

2.3

4.1

Width (cm)

Comments OG#21 sample 5, bag w/3 pouches, 1st is roundwood segment w/bark (loosely associated); drawn, one end may be cut OG#21 sample 5, pouch 2, 10 charcoal fragments plus detaching bark OG#21 sample 5, pouch 3, 2 charcoal fragments plus detaching bark OG#21 sample 6, roundwood in fragments, including detaching bark OG#21 sample 7, 2 roundwood segments, specimen 1 OG#21 sample 7, 2 roundwood segments, specimen 2, strong indication of sharp cut down entire length exposing pith, essentially cutting/splitting piece in half, drawn

21i

10

F.80

small wood

yes

ROG

RW

1/2

21j

10

F.80

small wood

no

ROG

RW

1/4

21k

10

F.80

small wood

yes

ROG

RW

21l

10

F.80

small wood

no

ROG

21l

10

F.80

small wood

yes

WOG

2.4

10.2

OG#21 sample 8, roundwood broken into 4 fragments, one surface appears split or cut down middle, down entire length bits of bark also present (possibly the other ½ of OG#21h)

2.6d

4.5

1/2

1.8

5.6

OG#21 sample 9, cut/split or just fractured along/down ray? OG#21 sample 10, appears cut/split down entire length & also a facet on outer (tangential) surface at one end, which is 14.34 mm long × 12.88 mm wide

RW

1/2

2.2

3.7

RW

3/4

2.1

8.1

2.4

3.3

21m

10

F.80

small wood

no

ROG

RW

1/8

21n

10

F.80

small wood

yes

ROG

RW

1/2

1.3

4.8

OG#21 sample 11, appears cut/split along two radii, drawn; part of OG#21h? OG#21 sample 11, no clear evidence of cuts 1.2

OG#21 sample 12 OG#21 sample 13, 9 roundwood fragments; this fragment appears cut or split down length (continued)

Table 5.5—Continued

Radius Diameter (cm) (cm)

Length (cm)

OG#

Unit

Zone

Field designation

Bark

Taxona

21n

10

F.80

small wood

yes

ROG

RW

3/4

21n

10

F.80

small wood

no

ROG

RW

>1/4

21o

10

F.80

small wood

yes

ROG

RW

1

21o

10

F.80

small wood

no

ROG

RW

21o

10

F.80

small wood

no

Carya, true

21p

10

F.80

small wood

yes

ROG

RW

1

1.5

5.2

21q

10

F.80

small wood

yes

ROG

RW

3/4

1.5

5.6

21r

10

F.80

small wood

no

ROG

RW

1/2

1.5

4.0

Formb Segmentc

2.2

2.8

Width (cm)

Comments OG#21 sample 13, 2nd fragment OG#21 sample 13, 7 smaller fragments OG#21 sample 14, complete roundwood segment, plus 3 detached fragments of bark; two sides cut, removing outer growth increments; diameter/width across from two cut sides is 1.6 cm OG#21 sample 14, additional fragments, n=3 OG#21 sample 14, n=2 charcoal fragments OG#21 sample 15, one complete radial OG#21 sample 16, one 3/4 section of radial OG#21 sample 17, one radial, split/cut down length in half, plus one additional charcoal fragment, not clearly associated but same taxon

21s

10

F.80

small wood

yes

ROG

RW

1/3

2.2

5.5

21t

10

F.80

small wood

no

ROG

RW

1/2

2.7

4.0

21v

10

F.80

small wood

no

ROG

RW

1/2

2.0

1.9

21w

10

F.80

small wood

no

ROG

RW

1/3

21x

10

F.80

small wood

yes

ROG

RW

1/8

21y

10

F.80

small wood

no

WOG

RW

1/2

2.7

2.62

OG#21 sample 18, one roundwood broken in two, not clearly cut/split, slightly curved in form & w/branch node off one side; 3 additional ROG fragments are associated OG#21 sample 19, one radial segment & possibly associated additional specimen, not clearly split/cut OG#21 sample 21, split/cut (?) in half down length, & one additional fragment of ROG charcoal OG#21 sample 22, one fragment of roundwood charcoal OG#21 sample 23, radial plus 2 fragments detached bark & 5 additional ROG charcoal OG#21 sample 24, two specimens: radial, possibly cut/split, & WOG fragment, probably same specimen (continued)

Table 5.5—Continued

OG#

Field designation

Taxona

timber, zone 2 to top of zone 4 timber, zone 2 to top of zone 4 timber in 3 segments

yes

ROG

RW

1/2

no

Pinus

RW

1/2

no

ROG

RW

1

6.7

timber in 3 segments timber in 3 segments

yes

Pinus

RW

1/2

4.9

no

Robinia

RW

1

2

timber in 3 segments

no

Carya, true

19

2

timber in 3 segments

yes

ROG

RW

1/2

4.5

20

2

wood

yes

Carya, pecan

RW

1

2.4

Zone

25

18

2/4

26

18

2/4

27a

19

2

27b

19

2

27b

19

2

27b

19

27c

29

Formb Segmentc

Radius Diameter (cm) (cm)

Bark

Unit

5.1

Length (cm) 43.0

Width (cm)

Comments fragment with curved surface, nail (nail or drill?) hole, incomplete (3 photos)

27.0

Robinia type 2 small fragments of hickory charcoal

10.0

cut/split down centerline? (3 photos); lying immediately beneath is large bark, oriented perpendicular to wood, HW

30

20

2

wood timber

no

31

20

2

wood

no

ROG

RW

1/2

32a

20

2

no

ROG

RW

1/8

32b

20

2

yes

ROG

RW

32c

20

2

wood timber wood timber wood timber

no

Pinus

slat-like

32d

20

2

no

ROG

RW

32e

20

2

wood timber wood timber

no

Carya, pecan

RW

32f

20

2

no

ROG

RW

1/2

1.6

OG#32 sample 5

32g

20

2

no

ROG

RW

1/2

2.1

OG#32 sample 5

32h

20

2

wood timber wood timber wood timber

yes

ROG

RW

1/2

3.7

Castanea planklike

15.0

2.1 1.7

18.0

3.4d

11.0

partly collapsed burned timber, squarish-rectangular in section, flat surfaced on top (wide) side fragmented into 7 segments, plus smaller fragments OG#32 sample 1 OG#32 sample 2, partial segment OG#32 sample 3, 10 slat-like, very flat fragments of pine charcoal (these are not flat from growth ring separation; suggests a flat timber) OG#32 sample 3, 1 partial segment OG#32 sample 4, 2 partial segments

4.3

OG#32 sample 6 (continued)

Table 5.5—Continued

OG#

Field designation

Length (cm)

Width (cm)

Taxona

no

ROG

RW

1/8

no

ROG

RW

1/8

1.2

2.4d

OG#32 sample 8

no

ROG

RW

1/4

2.2

4.4d

OG#32 sample 9

no

ROG

RW

OG#32 samples 10 & 11, identical, partial segment

cane & wood mass

yes

Fraxinus

RW

2

cane & wood mass

no

WOG

partial roundwood in 5 segments, 2 large fragments of smooth bark w/lateralmerged warts/ridges may go w/this (4 photos) one blocky (cube) of charcoal

2

small wood

yes

WOG

Zone

32i

20

2

32j

20

2

32k

20

2

32l

20

2

33a

22

2

33b

22

34a

22

wood timber wood timber wood timber wood timber

Formb Segmentc

Radius Diameter (cm) (cm)

Bark

Unit

RW

1/2

Comments OG#32 sample 7

3.8

27.0

OG#34 sample 1

34b

22

2

small wood

no

WOG

RW

1/2

2.3

OG#34 sample 2

34c

22

2

small wood

no

WOG

RW

1/2

3.1

OG#34 sample 2

34d

22

2

small wood

no

WOG

RW

1/2

2.9

OG#34 sample 2

35a

22

2

small wood

no

Carya, pecan

RW

1/2

2.0

OG#35 sample 1

35b

22

2

small wood

no

ROG

RW

1/2

3.1

35d

22

2

small wood

yes

ROG

RW

1/4

4.3

OG#35 sample 2 OG#35 sample 4

35f

22

2

small wood

yes

Carya, pecan

RW

38

46

2

large timber

no

Pinus

RW

1

8.8

62.0

falling apart badly, diameter & length very provisional

39

46

2

timber

(no)

Pinus

RW

1

6.4

21.0

5 photos

40a

26

2

—

no

Carya, true

41a

28

2

wood cluster

no

ROG

RW

41b

28

2

wood cluster

no

ROG

RW

41c

28

2

no

WOG

RW

41d

28

2

wood cluster wood cluster

42a

27

2

wood cluster

no

Carya, true

42b

27

2

wood cluster

no

Castanea

42c

27

2

wood cluster

yes

Carya, true

RW

42d

27

2

wood cluster

yes

ROG

RW

no

7.5

3.0 1/2

1

2.2

6.2

14.0

3.3

14.0

4.4d

20.0

Pinaceae slat-like

RW

17.0

1

3.0

2.0

3.0

3.2

OG#35 sample 6, bark with some wood attached

one blocky (cube) of charcoal OG#41 sample 1, roundwood lying beside/adjacent to next (2 photos) OG#41 sample 1, roundwood lying beside/adjacent to previous (2 photos) OG#41 sample 2, roundwood in 6+ fragments OG#41 sample 3, pine or spruce, fragmented into 3 parts; flat and slat-like OG#42 sample 1 OG#42 sample 1, 5+ charcoal fragments, all Castanea

1/4

1.3

2.6d

OG#42 sample 2, bark fragments, n=4 3.1

2.9

OG#42 sample 3 (continued)

Table 5.5—Continued

Field designation

Bark

Taxona

wood cluster

no

Carya, true

RW

1

2

wood cluster

yes

Carya, true

RW

1/2

27

2

no

Castanea

42h

27

2

no

Pinus

42i

27

2

yes

ROG

RW

42j

27

2

yes

ROG

RW

42k

27

2

wood cluster wood cluster wood cluster wood cluster wood cluster

no

Carya, true

42l

27

2

no

Castanea

42m

27

2

no

ROG

42n

27

2

no

Carya, true

42o

27

2

no

Carya, true

OG#

Unit

Zone

42e

27

2

42f

27

42g

wood cluster wood cluster wood cluster wood cluster

Formb Segmentc

RW

Radius Diameter (cm) (cm)

Length (cm)

Width (cm)

Comments OG#42 sample 3

3.4

1/2

3.5

10.7

OG#42 sample 4, roundwood in 5 segments OG#42 sample 4, 2 charcoal fragments OG#42 sample 4, 1 charcoal fragment OG#42 sample 5 OG#42 sample 6, partial roundwood segment OG#42 sample 6, 3 charcoal fragments

1/4

1.8

3.6d

OG#42 sample 6, 3 charcoal fragments OG#42 sample 7 OG#42 sample 8, 2 charcoal fragments

RW

43

35

2

timber

no

Pinus

RW

44a

29

2

timber

no

Pinus

RW

1

44b

29

2

timber

no

WOG

45a

29

2

—

no

Pinus

RW

1

2

45b

29

2

no

ROG

6.2 3.6

30.0

7.2d

circumference may have been larger, very friable (specimen keeps fragmenting) 20 fragments of oak charcoal

47

30

2

timber

no

Pinus

RW

1

7.2

25.0

48a

30

2

small wood

no

Carya, true

RW

1/2

2

10.0

48b

30

2

small wood

no

Pinus

slat-like

48c

30

2

small wood

no

Castanea

49

30

2

small wood

yes

Carya, true

51

45

2

bark & wood

yes

Castanea

RW

6.5

1/2

large roundwood in 5 segments (2 photos) 50+ fragments of pine, probably all from 1 original timber; largest remnant of original is source of measurements 3 fragments of oak charcoal

1.7

cut or split straight down midsection 1.6

15.0 5.5

3.5

thickness 6.4 mm; nail (?) hole, diameter 4.55 mm, with differential coloration around rim (from nail head?) (5 photos) stray smaller charcoal fragments sample fragmented into 7 segments; cut or split straight down midsection large bark segment with merged lenticles/ridges perpendicular to long axis (3 photos) (continued)

Table 5.5—Continued

OG#

Field designation

Radius Diameter (cm) (cm)

Taxona

Formb Segmentc

wood from south half of feature

no

Pinus

slat-like

2

wood

no

WOG

RW

1

3.6

29.0

52/53

2

timber

no

ROG

RW

1

5.6

35.0

51

2

timber

no

Castanea

RW

1

4.4

20.0

Zone

54

—

F.91

55

52/53

56 57

110 samples total Average diameter of 60 samples:e

a

Length (cm)

Bark

Unit

Width (cm)

Comments linear wood segment, lightly charred on one end (thin sections recovered for subfamily assignment) (4 photos) roundwood in alignment, broken into 5 segments roundwood broken into 4 large segments; one end is blunt and the other is pointed, but shape may be taphonomic (created incidentally during burning)

3.3 8.8

max.

1.3

min.

ROG=red oak anatomical group; WOG=white oak anatomical group; two anatomical groups of hickory include “true” and “pecan”; HW=hardwood indeterminate. RW=roundwood (radial, branch or stem); for descriptions of “plank-like” and “slat-like” (thin, narrow), see the text. c Proportion of full circle/circumference (1=whole, 1/2=half, etc.). d Data estimated from radius measurements. e Standard deviation: 1.67. b

Wood Selection and Technology in Structures 1 and 5 · 189

group, and another 13 percent of that material came from the white oak group. Pine is an additional 13 percent of the Structure 5 nonpost wood remains. A similar pattern was observed for Structure 1, with 42 percent red oak, 27 percent white oak, and 20 percent pine constituting the nonpost wood assemblage. Chestnut and the two types of hickory are also present in the nonpost construction material from both buildings and range from about 6 percent to 4 percent of the respective assemblages. Any absolute correspondences of detail between the two buildings begin to break down when considering additional aspects of construction. Particularly with regard to indications of carpentry practices, the details of the two buildings begin to diverge. As might be expected if there were distinctions in the amount of time and care that went into the construction of the two buildings, the nonpost roundwood demonstrates some qualitative differences in terms of material proportions, age estimates, and specific details of how the material was harvested and put to use. Individual specimens from Structure 1 range from 2.4 centimeters to 14 centimeters in diameter (thus up to about the size of the upright posts), while those from Structure 5 overlap in size but show a smaller range overall, 1.3–8.8 centimeters in diameter (see tables 5.4 and 5.5). Moreover, there is a general impression of smaller and younger material in association with Structure 5, whether considering the nonpost roundwood separately or taking into account the upright posts. The average diameter for nonpost roundwood from Structure 5 is 3.3 centimeters (n=60; standard deviation, 1.671 [table 5.5]). In contrast, the average diameter for Structure 1 nonpost roundwood is 6.6 centimeters (n=25; standard deviation, 3.12 [table 5.4]), or double that of the other structure. This contrast in overall size of construction material likewise holds true when the upright posts are factored in. Thus, the full assemblage of roundwood for Structure 5, including posts, has an average diameter of 4.12 centimeters (n=66). The average diameter of all Structure 1 roundwood is 8.17 centimeters (n=43), which again is twice as large as that of Structure 5 (figure 5.3). This discrepancy in the diameter classes is also reflected, somewhat predictably, in the age estimates for the two assemblages. To reiterate, the posts from Structure 1 were determined to be 15–100 years in age. Ages for the nonpost class from this structure demonstrate a wider range, with growth ring counts from as low as 8 years to as many as 132 years (average 24.3 years, standard deviation 26.8 [table 5.6]). The 132-year-old specimen is a large, plank-like timber (OG#44) from Unit 50 (see table 5.4); excluding it, the average age for nonpost roundwood is 19.4 years (range, 8–70 years;

Figure 5.3. Histogram comparing roundwood diameters from Structure 1 (n=43) and Structure 5 (n=66).

Table 5.6. Growth ring observations and harvest periods for Structure 1 samples Unit

Zone

Taxona

Formb

1

7

4

WOG

RW

2

7

4

WOG

RW

3

7

4

WOG

RW

OG#

Ring count

17

Ring observations last 3–5 rings are very stressed with virtually no LW, thus season of harvest may be misleading (and see next) not certain of full radius (ring count) & absolute terminal ring; last 4–5 rings highly stressed with minimal or no LW formation; small portion of last increment shows some initial LW last 3–5 rings are very stressed with virtually no LW, thus harvest period uncertain; probably from same tree (or branch?) as OG#1 (and see OG#2, which may provide correct or more accurate reflection)

Seasonc (EW)

LW

(EW)

Ring count

Unit

Zone

Taxona

Formb

4

7

4

WOG

RW

15

13

6

Pinus

RW E½

16

13

6

Pinus

RW W½

70

19a

40

4

ROG

RW

15

OG#

19b

40

4

ROG

RW

19c

40

4

WOG

RW

10

19e

40

4

ROG

RW

17

19g

40

4

WOG

RW

28

20a

40

4

ROG

RW

14

Ring observations LW well developed in last ring & bark interface observed 18 rings over 18.30 mm (i.e., 1.016/mm, about 1/mm or 10/cm); two representative rings measured directly at 0.67 mm and 0.91 mm (thus about ½ mm and just under 1 mm) pith appears to be present (fragmented exposure of base); est. count with above, radius ~7 cm (ca. 70 years), growth rings frequently narrow (as smaller one above), thus may be older pith to outer ring; last ring with well-developed LW, appears to include bark interface locally

latewood developing in last increment (bark interface observed) some rings very narrow with almost no LW; last increment includes well-developed LW but uncertain whether it is the absolute last pith present; rings wide; last has full complement of LW but not clear whether bark interface present

Seasonc LW

LW

(LW) LW (LW)

(LW)

(continued)

Table 5.6—Continued

OG#

Unit

Zone

Taxona

Formb

Ring count

20b

40

4

ROG

RW

14

20c

40

4

ROG

RW

14

21a

40

4

Pinus

center post

56

21b

40

4

ROG

RW

23

21d

40

4

Carya, true

RW

24

12

4

Pinus

RW

27

12

4

WOG

RW

29

17

5

Castanea

post

31

13

4

Pinus

RW

20

32b

5

5

Pinus

post

86

35

10

4

WOG

RW

25

38

6

3–4

Pinus

RW

25

Ring observations pith present, rings ending in well-developed LW 13 rings counted, probably extends to 14 23 rings in 4.09 cm yields 5.6 years/cm × radius, or 56 years minimum (diameter is approximate, but good radius [see table 5.2], thus actual diameter is probably more like 20 cm) radius includes pith and bark incomplete segment, 10 rings total, minimal arc/ curvature full LW at or very close to bark interface

Seasonc LW

slightly off-center pith 19 rings counted, pith obscure but present and includes at least 1 more narrow ring (slow growth), 33+ rings, loose count near pith; may be another 5–10 rings there; 16 rings in 2.52 cm; ca. 86.6 years old exterior to pith (radius); rings about 1 mm, one measured at 1.19 mm innermost rings very wide; bark beetle channels present, but absolute final increment is unclear

(EW)

(EW)

Ring count

Unit

Zone

Taxona

Formb

43

50

4

ROG

plank-like

44

50

4

ROG

plank-like

132

47

14

4

Pinus

post

100

55

38

4

ROG

RW

58

20

4

WOG

RW

63

19

5

ROG

post

64

19

4

ROG

RW

16

67

19

4

WOG

RW

8

68

19

4

WOG

RW

9

73

41

4

ROG

RW

28

74

41

4

ROG

RW

11

OG#

Ring observations very narrow terminal increments (final years of growth) with virtually no LW, including last ring, thus harvest period uncertain 15 rings in 16.62 mm, thus 1.10 rings/mm or ~11/cm × diameter (width), or ca. 132 years 10 rings in 13.87 mm near center (clearest set), and rings appear fairly uniform (about the same growth rate) beyond that, thus 1.387/ mm or ~14/cm × radius, or ca. 99.86 years well-developed LW at bark interface and no indication of subsequent EW formation relatively juvenile wood (younger branch/stem) ring form/position appears central; post may have been larger circumference 16 counted but definitively terminal ring is not evident (may have been older/larger) slight ring curvature (minimal arc) many rings, extremely narrow with little or no LW, ~0.89/mm absolute terminal increment uncertain

Seasonc

(LW)

LW

LW LW

(continued)

Table 5.6—Continued

OG#

Unit

Zone

Taxona

Formb

Ring count

76a

41

4

ROG

RW

31

76b

41

4

ROG

RW

78

38

4

ROG

RW

79

38 & 50

4

ROG

RW

81

26

3

ROG

RW

82

26

5

Castanea

post

84

38

4

ROG

plank-like

85

72

4

Robinia

post

15

90

50

4

ROG

RW

15

91

74

2

Pinus

RW

8

92

26

5

Pinus

RW

8

93

65

5

Pinus

center post

73

16

Ring observations uniformly very narrow growth increments, ~0.91/mm similar details as noted for OG#76a, probably from 1 original full LW is clear, possibly earliest of EW but very uncertain mostly very low arc (curvature) to growth rings uncertain as to presence of terminal increment ring widths fairly uniform, ~2–2.5 mm tangential surface is broad face (e.g., plane sawn timber) ring count is minimum; uncertain of presence of final increment/year final increment has LW but uncertain whether any further development of EW; rings 11–13 are extremely narrow with little or no LW exceptionally wide rings (4.75/mm) wide rings (6.7/mm), low (minimal) arc compression wood: rings highly eccentric; narrow ring side ~10/1.3 cm (10 per 13 mm, or 1.3/mm), wide ring side ~5 in same span (i.e., 2.6/mm); tree was ca. 73 years old, extrapolating from narrow ring side (56.23 mm radius × 1.3, or 73.09 years)

Seasonc

(LW)

(LW)

(LW)

Taxona

Formb

OG#

Unit

Zone

2003a

18

4

Castanea plank-like

2003b

18

4

Castanea plank-like

a

Ring count

Ring observations

Seasonc

tangential (plane sawn) plank with very narrow rings, ~0.5 mm wide with dense packing of LW pores in very narrow LW portions of increments, EW pores in 1–2 rows; minimal or essentially no ring curvature; ca. 21–28 years across width of plank; 6 rings/7.28 mm yields 1.2/mm, 5 rings/7.1 mm yields 1.4/mm, 5.5 rings/5.91 mm yields ~1.07/mm; comparative specimens: 4 rings/14.3 mm, ~0.27/mm; ring widths ranging from 2.62 mm to 5 mm 4.5 rings over 8.34 mm yields ~1.8 rings/mm, similar to OG#2003a and likewise low/no arc to rings (these two timbers probably derive from the same original bole)

ROG=red oak anatomical group; WOG=white oak anatomical group; two anatomical groups of hickory include “true” and “pecan”; HW=hardwood indeterminate. b RW=roundwood (radial, branch or stem); “plank-like” indicates a slat (thin, narrow). c Season, i.e., period of terminal growth or harvest: EW=earlywood (spring), LW=latewood (summer–fall). Parentheses indicate tentative assignment.

196 · Lee Ann Newsom

Figure 5.4. Age classes for all roundwood, including posts, from Structure 1 (n=26) and Structure 5 (n=31).

n=22). In contrast, no material from Structure 5 exceeds 50 years of age (and recall that the entire structure has been excavated). One previously mentioned post (OG#65) was counted at 47 years, while the rest of the assemblage has a range of 5–28 years, with an average of 12.6 years (standard deviation 6.32 [table 5.7]). These statistics exclude six roundwood items (specifically, those indicated in tables 5.5 and 5.7), as possible fragments of other specimens. Because these are more or less duplicates, excluding them from statistical analysis avoids double counting specimens that may represent the same piece before fragmentation. The contrast in age estimates between Structures 1 and 5 is graphically portrayed in figure 5.4. Construction Process Another line of evidence that distinguishes the construction of the two buildings is postharvest finishing or dressing of the timbers. The “segment” column in tables 5.4 and 5.5 indicates the proportion retained of the full cross section—that is, the original circumference—for each specimen of roundwood, exclusive of the posts. This is scored as 1 for complete specimens and otherwise as the fraction remaining of the full circumference (3/4, 1/2, 1/4, and 1/8).2 The column is left blank for more fragmentary specimens or material that appears to represent further processing into the lumber category, those designated as slat- or plank-like forms. The roundwood segment data are summarized for both buildings in table 5.8. These

Table 5.7. Growth ring observations and harvest periods for Structure 5 samples Unit

Zone

Taxona

Formb

1

2

2

WOG

RW

strong development of LW

LW

8

11

2

WOG

RW

(LW)

12

10

2

ROG

RW

19

14

2

WOG

RW

21a

10

F.80

ROG

RW

24

21b

10

F.80

ROG

RW

13

full development of LW but uncertain whether final increment present uncertain whether final increment present EW clear, but uncertain whether any LW formed pith present, outermost rings narrow, sometimes with minor LW formation; more rapid growth toward interior (juvenile wood phenomenon) pith present

21f

10

F.80

ROG

RW

7

pith present

21f

10

F.80

ROG

RW

7

probably part of OG#21a

21h

10

F.80

ROG

RW

8

21i

10

F.80

ROG

RW

20

21l

10

F.80

ROG

RW

7

21l

10

F.80

WOG

RW

19

21m

10

F.80

ROG

RW

5

21n

10

F.80

ROG

RW

8

21n

10

F.80

ROG

RW

12

21o

10

F.80

ROG

RW

7

21p

10

F.80

ROG

RW

7

21q

10

F.80

ROG

RW

8

21r

10

F.80

ROG

RW

6

21s

10

F.80

ROG

RW

8

OG#

Ring Ring observations count

9

Seasonc

(EW)

very narrow rings (20–21 years) specimen may be part of OG#21h

oval cross section with pith in true center (rings not eccentric)—bark retained on one “compressed” side, absent on broader surfaces; truncation/trimming of growth rings accounts for shape in cross section may be part of OG#21h

pith present; may go with OG#21q

(continued)

Table 5.7—Continued

Unit

Zone

Taxona

Formb

21t

10

F.80

ROG

RW

7

21v

10

F.80

ROG

RW

7

21w

10

F.80

ROG

RW

8

21y

10

F.80

WOG

RW

8

26

18

2/4

Pinus

RW

27b

18

2/4

Pinus

RW

29

20

2

Carya, pecan

RW

21

31

20

2

ROG

RW

10

32a

20

2

ROG

RW

14

32b

20

2

ROG

RW

32g

20

2

ROG

RW

9

32h

20

2

ROG

RW

7

32j

20

2

ROG

RW

13

32k

20

2

ROG

RW

7

34a

22

2

WOG

RW

34d

22

2

WOG

RW

22

35b

22

2

ROG

RW

22

39

46

2

Pinus

RW

42f

27

2

Carya, true

RW

17

42m

27

2

ROG

RW

16

44a

29

2

Pinus

RW

9

OG#

Ring Ring observations count

Seasonc

count may be as much as 10 years, but friable and fragmenting possibly part of OG#21t

bark beetle channels observed spanning several rings, thus uncertain as to last increment

pith present LW LW

LW

eccentric form but rings are not eccentric; specimen appears flattened (inside is unburned & may account for partial collapse); bark beetle channels

3.6 cm from inner to outer surface, including at least 9 wide rings, ~3 mm wide each

EW

Wood Selection and Technology in Structures 1 and 5 · 199

OG#

Unit

Zone

Taxona

Formb

45a

29

2

Pinus

RW

47

30

2

Pinus

RW

55

52/53

2

WOG

RW

15

57

51

2

RW

28

59

2

4

Castanea Pinus

post

65

16

4

Pinus

post

66

4

4

Pinus

post

Ring Ring observations count 9

47

pith present plus 9+ years growth over ~2 cm; wide rings (may go with OG#44, closely situated) bark uncertain; beetle channel on one tangential surface

Seasonc

(EW)

off-center pith (reaction wood) rings wide, low curvature (arc); radius is partial coming apart; ring curvature (low arc) suggests large diameter

a

ROG=red oak anatomical group; WOG=white oak anatomical group; two anatomical groups of hickory include “true” and “pecan”; HW=hardwood indeterminate. b RW=roundwood (radial, branch or stem). c Season, i.e., period of terminal growth or harvest: EW=earlywood (spring), LW=latewood (summer– fall). Parentheses indicate tentative assignment.

data appear to suggest a fundamental difference between the two structures in the way raw material was processed for construction purposes. The majority (47 percent) of the nonpost roundwood from Structure 1 is fully or essentially complete and scored as 1; about 28 percent are half segments. In contrast, a greater proportion of the roundwood from Structure 5 is reduced to half or finer segments, 45 percent being half segments (table 5.8).3 Many of the half segments from both structures suggest deliberate cutting or splitting down the center line, as indicated in the “comments” column of tables 5.4 and 5.5. This operation would have produced two useable lengths of wood for each original bole or branch segment that was harvested, thus representing a more efficient use of wood construction materials and perhaps also a time-saving exercise, since it required the harvesting of less raw material. Yet this operation might also have produced a weaker structure, suggesting a perceived need for expediency in the planning of Structure 5. Another reason the issue of wood reduction deserves special attention is the possibility that differential use of metal tools, rather than traditional

200 · Lee Ann Newsom

Table 5.8. Proportion classes for roundwood Segments Whole

Threequarters

Onehalf

Onethird

Onequarter

Oneeighth

Onesixteenth

n

22

0

13

1

4

6

1

47

46.8

0.0

27.7

2.1

8.5

12.8

2.1

21

3

32

4

5

6

0

29.6

4.2

45.1

5.6

7.0

8.5

0.0

Structure 1 Number Percentage Structure 5 Number Percentage

implements of stone, distinguishes the construction process for the two buildings. To assess this possibility, I examined carpentry evidence, including the presence and form of tool marks (the “comments” columns of tables 5.4 and 5.5 contain notes on additional observations and potential evidence for carpentry activities). OG#21 from Structure 5 consisted of multiple wood specimens that were collected as individual bags or small groups of specimens bagged together depending on proximity to one another. Four of these samples have roundwood or other items that display clear tool marks or otherwise suggest the use of very sharp implements, and several others are half or quarter segments for which the separation is comparatively long and true—that is, crisp or decisive enough to suggest cutting with a sharp-edged metal blade rather than simply splitting along the grain. The most notable specimens derive from the samples designated OG#21c, #21h, #21k, and #21o (see table 5.5). Sample OG#21c includes nineteen MNI altogether, among which are eight larger fragments that also exhibit relatively sharp sloping bevel cuts (transversely angled across the grain) that would have served to separate the wood—that is, to cut it free from the rest of the branch or trunk member. This is clearly the result of neither natural breakage nor chopping with stone adzes. One of these specimens (OG#21c; figure 5.5) has, in addition to the beveled end, separate facets on the outer surfaces that may have been associated with dressing or bark removal, though because of charring distortion this is somewhat uncertain. Sample OG#21h includes a segment of roundwood with a long, even cut down its midline that exposed the pith for most of the length, approximately 11 centimeters, which is reminiscent of long machete cuts (figure 5.6). Samples OG#21i and OG#21k include similarly modified segments of

71

Wood Selection and Technology in Structures 1 and 5 · 201

Figure 5.5. OG#21c (Structure 5) showing beveled end and separate facets on outer surfaces.

Figure 5.6. OG#21h (Structure 5) showing long, even cut down midline reminiscent of machete cuts.

roundwood, and the latter also exhibits a facet (14.35 × 12.88 millimeters) on the outer (tangential) surface at one end (see table 5.5). Sample OG#21o is a complete segment that retains the bark on one outer surface, yet on two other parallel surfaces the growth rings are truncated, with the outermost growth having been removed by cuts down the length of the sample. Whether this trimming operation was intended simply to remove bark or instead to intentionally shape the wood, the modification resulted in an oval cross section. This specimen is one of the few examples of hickory in the structure. Another sample (OG#21f) includes a segment of roundwood with a cut end, though this is somewhat obscured because of burn distortion or subsequent taphonomic processes.

202 · Lee Ann Newsom

Figure 5.7. OG#25 (Structure 5) showing partial curvature and apparent edge of a hole that penetrated perpendicular to the longitudinal axis, possibly from the use of a nail or other fastener.

A few additional samples from Structure 5 require discussion. Two samples have circular holes that suggest nails or other fastenings. The sample designated OG#25—from Unit 18, zone 2/top of zone 4—is a half segment of roundwood in several fragments (figure 5.7). One of the fragments has a curved opening suggestive of a nail or drill; the sample is incomplete, however, and thus verifying whether the opening was cultural or natural is difficult. The anatomical structure clearly rules out the possibility that this was a branch trace in which the knot had fallen away or separated. OG#48b, part of a “small wood” collection from Unit 30, zone 2, is the second sample with a small opening or hole penetrating the wood horizontally across the grain (figure 5.8). This is a relatively flat, somewhat prismatic and slat-like specimen of pine with a hole (again, only partially complete) that measures 4.55 millimeters in diameter. The rim around the edge of the hole is discolored relative to the surrounding wood, which is highly unusual in charcoal and suggests that indeed something with an overlapping lip (such as the head of a nail) once occupied the hole and was present at the time the wood burned. In addition to this specimen, twelve other slat-like items of pine were recovered: one from Feature 91 (OG#54), ten from Unit 20 (OG#32c), and one (OG#41d) from Unit 28 (see table 5.5). The Feature 91 specimen is only partly carbonized—lightly charred on one end—while the other two are fully carbonized. One “plank-like” chestnut timber was recovered from

Wood Selection and Technology in Structures 1 and 5 · 203

Figure 5.8. OG#48b (Structure 5), a slat-like wood specimen with a small hole penetrating the wood horizontally across the grain; the lighter rim around the edge suggests the overlapping head of a nail.

Structure 5 (OG#30, Unit 20, zone 2); it measures 15 centimeters long by 11 centimeters wide (see table 5.4). Structure 1 yielded several samples that bear tool marks and other evidence of carpentry activity. A half segment of white oak roundwood (OG#3, Unit 7 [see table 5.4; figure 5.9]) was clearly worked to a point at one end, which is evident from the presence of four clear facets. These tool marks appear on the convex surface and range in length from about 10 centimeters to 25 centimeters and are about 8.5–16 millimeters wide (table 5.9). The wood is fully carbonized and very flat on the opposite side. A specimen of pecan roundwood from Unit 40 (OG#19i) also appears to have facet marks, in this case running parallel with the grain along the length of the exterior (figure 5.10). Three clear facets lie immediately adjacent to one another, and they range from 6.7 millimeters to 8.2 millimeters wide (table 5.9); together, these marks suggest the shaping and smoothing of a wooden implement like a tool handle. Within each are finer striations about 0.76 millimeters wide that probably resulted from a rough bit surface. With the possible exception of the facets on OG#19i, just detailed, all of the other Structure 1 samples I have just commented on could potentially have been made with aboriginal stone adzes, although metal implements remain a distinct possibility. A whole red oak roundwood specimen from Unit 40 of Structure 1 (OG#20a [see table 5.4]) has a sharply bevel-cut end, and a half segment

Table 5.9. Tool mark measurements Structure 1

1

OG# 3

19i

Unit 7

40

Zone 4

4

Type facet

Length (mm) 11.91

Width (mm) Comments 8.57

facet

24.43

14.14

facet

10.23

16.15

facet

25.38

8.87

facet

6.71

facet

6.71

facet

8.29

1

20a

40

4

bevel

1

65

19

4

facet

finer striations in facet, ~0.76 mm finer striations in facet, ~0.76 mm finer striations in facet, ~0.76 mm

facet facet facet 1

81

26

3

saw

5

21c

10

F.80

bevel

linear-narrow marks extending across face

facet 5

21k

10

F.80

facet

14.34

12.88

Figure 5.9. OG#3 (Structure 1), a half segment of white oak roundwood with a pointed end formed by a long bevel on one edge and four clear facets on the opposite edge.

Wood Selection and Technology in Structures 1 and 5 · 205

Figure 5.10. OG#19i (Structure 1), a fragment of pecan roundwood with facet marks running parallel with the grain, along the length of the exterior (the central facet is distinguished by brackets; note adjacent facets on either side and finer striations within individual facets).

of red oak roundwood lying immediately beside it (OG#20b) has a long, sharp cut down the center line where a lateral knot or branch appears to have been deliberately excised by a sharp transverse cut that suggests the use of a metal blade. Two red oak “plank-like” timbers from Unit 50 (OG#43 and OG#44) are about 12–13 centimeters wide and were partially trimmed by the removal of some of the outer growth increments; this resulted in somewhat angular, squarish forms that are narrowed toward one end, perhaps the result of sharpening or shaping (figure 5.11). The OG#44 timber produced a ring count of 132 years (see table 5.4). Very similar to these two finished, roughly quadrilateral timbers are the upright pine posts OG#32b and OG#47. OG#32b (figure 5.12) is roughly square in outline, having been at least crudely shaped, although clear tool marks are lacking and have perhaps been obscured by the carbonization process. OG#47 is a large timber with a 100-year ring count (see table 5.2). While it has the appearance of dimensional compression, this was found to have actually resulted from trimming on one side that removed the outer growth increments along that surface. OG#84 from Unit 38, Zone 4 (figure 5.13; see table 5.4 and figure 4.19b), designated in the field as a “doorway timber,”

Figure 5.11. OG#44 (Structure 1), worked timber or plank.

Figure 5.12. OG#32b (Structure 1), a squared post shown in transverse (cross section) perspective.

Wood Selection and Technology in Structures 1 and 5 · 207

Figure 5.13. OG#84 (Structure 1), a doorway threshold or doorsill.

is another large, crudely shaped and “plank-like” red oak timber. Three timbers from Unit 19, zone 4 of Structure 1 provide good evidence of wood workmanship. OG#64 and OG#66 are large red oak and pine timbers, respectively, that bear suggestions of shaping or bark removal in the presence of facets at the ends of OG#64 and vague tool marks along the lateral edges of both (see table 5.4 comments). Finally, three large red oak timbers—a post from Unit 38 (OG#53) and two timbers from Units 26 and 38, respectively (OG#81 and OG#78)—are especially revealing (figure 5.14). These have sharp, diagonal cuts across their ends that spanned fully and cleanly across their entire circumferences of 12.8 centimeters (OG#53), 14.14 centimeters (OG#78), and 34.55 centimeters (OG#81) (computed areas of 13.20, 15.90, and 95.03 square centimeters, respectively). The last, largest timber also bears very straightedged, narrow marks across the cut face that strongly suggest saw marks (figure 5.15). Moreover, there is no suggestion of the several adze cuts that would have been required to fell this timber using indigenous stone implements or even a metal ax. This particular timber is probably the strongest evidence for the use of European carpentry tools. In general, the presence of these essentially lopped-off or sawn timbers appears to distinguish the fundamental carpentry—or at least the procurement of raw materials—for Structure 1 from that for Structure 5, mirroring the greater overall size and age of material associated with the former structure.

Figure 5.14. Structure 1 timbers OG#78 (a, 4.5 cm in diameter) and OG#81 (b, 11 cm in diameter) shown in profile demonstrating very clean, complete separation across the entire circumference of woody stems.

Figure 5.15. OG#81 (Structure 1) shown in cross section with parallel saw marks spanning the full diameter.

Wood Selection and Technology in Structures 1 and 5 · 209

Figure 5.16. OG#10 (Structure 1), an upright chestnut post with a large (1 cm in diameter), angular hole (indicated by arrow) penetrating the grain and suggesting the use of an iron spike or bolt.

A very large chestnut plank was also recovered from Structure 1 in 2003 (samples listed at the end of table 5.4 as OG#2003a; see figure 4.4b). This lumber was minimally 76 centimeters long, 24 centimeters wide, and approximately 3–4 centimeters thick. Technically this falls within the “board” category, but it is very close in thickness to the second category for larger lumber (as described above, material classified as “dimension”). The relative position of the growth rings makes it clear that this is a tangential plank, or plane-sawn timber. At the end of this chapter I elaborate on the distinction between plane and quarter sawing and what this may indicate in terms of the wood technology represented by this particular timber. Meanwhile, a related observation is that the growth rings exhibit virtually no curvature, indicating that this plank derived from a tree of considerable girth and age. Unfortunately, any saw or other tool marks are now obscured. One additional detail regarding the modification of individual Structure 1 timbers concerns an upright chestnut post designated OG#10 from Unit 7 (figure 5.16; see table 5.2). A large hole penetrates the wood near the core; it is roughly circular and about 1 centimeter wide but has one very straight

210 · Lee Ann Newsom

side with a sharply angled corner, suggesting a cultural origin. This hole was perhaps made by something square, such as an iron spike, rather than a round, necessarily natural object. Time/Season of Harvest and Construction The majority of the samples from both structures for which the final growth increment is preserved (see tables 5.6 and 5.7) indicate the presence of latewood in the terminal growth increment. This is specifically the portion of the growth increment that forms during the final months of the growing season, that is, summer to early/mid-fall, depending on local conditions. Four specimens from Structure 1 may signal harvest during the spring (see table 5.6), as do three from Structure 5, and one of these—a roundwood pine timber—more than the others (OG#39, figure 5.17 [see tables 5.5 and

Figure 5.17. OG#39 (Structure 1) timber showing (a) bark beetle channels (at 7× magnification) traversing the surface and an exit hole partially preserved on one edge (at bottom of image); and (b) wood in cross section with the final growth ring at the top (arrow) and formation of earlywood (springwood) that has been locally eliminated by beetle channels.

Wood Selection and Technology in Structures 1 and 5 · 211

5.7]). None of these items is a post, however, so whether the temporal assignments represent the actual period of construction or subsequent additions, repairs, or enhancements is unclear. Many of the oak specimens, along with a number of the chestnut items, exhibit very narrow growth increments for which there is little or no latewood formation in the final and/or penultimate years of growth (see “ring observations” columns in tables 5.6 and 5.7). Such increments correlate with circumstances that are highly stressful to plant growth and development, such as a series of consecutive droughts or very cool conditions. There is a strong possibility that this correlates with the Jamestown drought episodes, specifically, with the low-rainfall period of the early to mid-1560s (Stahle et al. 1998, 2000). Other Construction Materials Secondary and tertiary construction elements used in the two structures include cane stems (Arundinaria sp.), sheets of hardwood bark, and possibly moss (table 5.10). Fragments of interwoven split cane may represent

Table 5.10. Tertiary construction elements from Structures 1 and 5 Structure

Field OG# Unit Zone designation Bark

1

21e

40

4

wood packed around center post

yes

HW

5

10

11

2

bark

yes

HW

5

21d

11

2

bark

Taxona

monocot

Length Width (cm) (cm) Comments OG#21 bag 1 samples 4 and 5, two large fragments of hardwood bark seven 3.5 mm carbonized fragments of bark OG#21 sample 3, monocot node segment, 6.72 mm in diameter, cf. Arundinaria (cane) (continued)

212 · Lee Ann Newsom

Table 5.10—Continued

Length Width (cm) (cm) Comments

Structure

Field OG# Unit Zone designation Bark

5

21u

11

2

bark

yes

HW

5

35c

11

2

small wood

yes

HW

5

35e

11

2

small wood

yes

HW

9.0

5

40b

11

2

—

Arundinaria

9.5

5

46

29

2

wood & bark

a

yes

Taxona

HW

3.0

OG#21 sample 20, one fragment of roundwood bark, probably oak (xylem absent) OG#35 sample 3, relatively smooth bark, cf. Castanea OG#35 sample 5, large bark in linear alignment, wood absent but structure suggests Carya cane stem, 2 fragments including node, diameter ~1.7 cm all bark (6 fragmentary), probably oak

HW=hardwood indeterminate.

matting that was used to cover the interior floors of both buildings. Such matting was perhaps also used as wall coverings. Individual segments of cane stem, especially the node region, may have been fashioned into sheets of wattle-work to cover the walls and roofs of the buildings. Cattail (Typha sp.) and possibly another type of monocot such as a sedge or rush have also been identified in the floor deposits (in collaboration with Sherwood’s microstratigraphic work) and perhaps also served as interwoven matting

Wood Selection and Technology in Structures 1 and 5 · 213

or floor covering. I further consider the presence of these additional herbaceous stems in the discussion that follows this section. The individual specimens of bark that I have examined have all been from hardwoods, with provisional assignments to oak, hickory, and chestnut (table 5.10). Finally, in the course of the microdebris analysis from Sherwood’s microstratigraphic analysis (chapter 4, this volume), there is a suggestion for the presence of a lower vascular plant such as a moss or bryophyte. If present by direct intent, the moss may have been used as packing and insulation. Insulation material in Structure 1 may also have included bald-faced hornets’ nests, fragments of one having been recovered attached to a red oak timber (OG#20b, figure 5.18).

Figure 5.18. OG#20b (Structure 1), fragments of a bald-faced hornets’ nest possibly used as building insulation material.

214 · Lee Ann Newsom

Fuelwood (Hearth Samples) Only a small subset of hearth-associated samples has thus far been analyzed. These have proven to be very similar to the general construction debris from the two structures. Red oak roundwood constitutes the majority of the charcoal MNI (88 percent) from the Structure 1 hearth, Feature 119 (table 5.11). These numbers are probably inflated because of fragmentation, however, and much of the red oak count may derive from one original segment of roundwood. Pine and the white oak group were also recovered from this feature. Five wood taxa were recovered from the Structure 5 hearth, Feature 91, including red and white oak, chestnut, and both anatomical forms of hickory (table 5.11). Together, the six wood taxa from these hearths are fully consistent with the identifications of the various building construction elements. In particular, Structure 5 samples OG#60b–60d are highly congruent with the structural debris. This suggests that either the same basic suite of woods used for construction was readily available and in sufficient supply to have also been used for fuel purposes, or that these items were originally secondary or tertiary building elements that collapsed on and became incidentally associated with the hearth feature. Nevertheless, if these woods were sufficiently abundant locally to serve dual or multiple purposes, they are high-quality fuel sources, being among the most dense woods available in the eastern deciduous forest environment (greater density equates with fuel efficiency [Little 1983; NAS 1980]). Sample OG#60a, which is an entire segment of pecan hickory roundwood that was about twenty years old at harvest and 27 centimeters in length by 5.32 centimeters in diameter, is notable. This specimen narrows at both ends owing to having been cut (as indicated by the faceted ends) to separate it from a longer section of bole/trunk (there is no indication that this was branch wood); thus, it may represent raw material that was cut into fuel-appropriate lengths. Since this specimen did not further break apart in the course of burning, as typically happens with repeated or enduring burns, I infer that it was not subject to repeated or enduring fire and heating such as occurs with the reuse of a prepared hearth and sometimes with continuously lit fires. Thus, if indeed fuelwood, this specimen must represent the final or near-penultimate use of the hearth. This item conceivably could have been unburned material that was placed in the hearth in preparation for lighting a new fire when the building conflagration happened, causing the wood to burn for the first time.

Table 5.11. Charcoal from fuelwood contexts

Pine Structure 1

Taxon (N) Chestnut Hickory, Hickory, Red oak pecan true group

OG# —

Feature 119– West A

Zone 1

Bark no

1

3

1

—

119– West B

1

yes

1

30

1

—

4

no

5

60a

119–East half 91

—

no

5

60b

91

—

no

5

60c

91

—

yes

5

60d

91

—

no

5

60e

91

—

no

White oak group

6

Comments FS 349, bag 545, fragmentary roundwood FS 350, bag 547, beetle channels in pine indicate terminal latewood; roundwood fragments may all indicate 1 original no analyzable wood

1

1 2 1 1

1

sample #1, large roundwood, complete segment ~27 cm long × ~5.32 cm in diameter; narrows at both ends due to cutting to separate from trunk (i.e., cut into fuel-appropriate portions); ca. 20 years old sample #2, small roundwood, ~half segment, ~10 cm long sample #3, 3 roundwood one-third to one-fifth segments, radii ~12–16 cm, lengths 6–10 mm sample #4, 1 roundwood quarter segment, radius ~21 mm, length ~3 cm sample #5, roundwood, 10 fragments identified, all indicate 1 original

216 · Lee Ann Newsom

To reiterate, Structures 1 and 5 are both considered to have been part of the Spanish compound at Joara, though Structure 5 was probably built later than Structure 1. Structure 5 is immediately distinguished from Structure 1 by its lack of a clearly defined house basin, as well as by the absence of a well-defined central roof support infrastructure such as was formed by the large pine center posts of Structure 1. Instead, at least two but perhaps three of Structure 5’s key supports were insubstantial enough to have required a second support post. Otherwise, on the basis of this analysis the two buildings were found to share several general characteristics such as the primary suite of taxa used, with red oak being the most commonly identified taxon overall, followed by pine and white oak in that order, then the rest of the wood taxa. Under finer scrutiny, however, it became apparent that the respective wood assemblages from the two buildings also bear important distinguishing characteristics. Pine was the wood of choice for the major structural supports of both buildings, particularly the center posts that bore the greatest burden of weight from the crowning roof assembly. Given the likely abundance of hardwoods in the area, the selective use of pine for this purpose was perhaps due to its basic growth form, with the tall, straight boles that typically lack horizontal branches along the lower portions being almost ready-made posts. Its use also probably reflects recognition of the inherent strength and durability of hard/yellow pine wood, especially the heartwood. According to Alexis Panshin and Carl de Zeeuw (1980: 446), this wood is highly suited for use as poles, mine timbers, piling, railroad ties, and structural timbers such as joists, trestles, beams, and stringers in bridge, dock, warehouse, factory, and other heavy construction contexts because of its strength, stiffness, and hardness. American chestnut was the second most used wood for posts in the Berry structures, followed by black locust, red oak, and to a lesser extent, white oak. Red oak dominated the nonpost roundwood assemblage from both structures, constituting wood used for other architectural components of the two buildings, presumably minor supports, framing, and rafters. In addition to roundwood elements, including posts, other specimens were individually distinguished in the lumber category of finished wood products. This included planks in the board category (among which is one—the large chestnut timber from Structure 1—that approaches the size class of “dimension”) and a few others found only in Structure 5 that were designated as “slat-like,” technically in the “strip” category.

Wood Selection and Technology in Structures 1 and 5 · 217

Close scrutiny of the wood technology and carpentry evidence revealed details that further distinguish Structures 1 and 5. In particular, the two buildings contrast markedly in terms of the relative ages and diameter classes of the woods used in their construction. The builders of Structure 5 generally used raw material of smaller proportions than that from Structure 1, with the posts and other roundwood from Structure 5 indicating younger and smaller-diameter material overall (the diameters being consistently about half of those of Structure 1). Given that much of the Structure 5 raw material was of a smaller girth than that used for Structure 1, we might expect this material to have been used whole, with the complete circumference left intact and the item deployed in the form of a post, pole, or rod. Instead, only about 30 percent of the Structure 5 material was used intact, with another 45 percent reduced to half segments and the remaining 25 percent as one-third or finer splits. In contrast, 50 percent of the Structure 1 roundwood (post and nonpost) was deployed complete, about 30 percent was split into halves, and the rest was reduced to finer portions. As discussed above, this material reduction may have left Structure 5 as a relatively weak or flimsy building. Whether this apparent lack of structural integrity was by design or the result of need is uncertain, but this matter is considered further below. First, it is useful and instructive to examine some of the details of the Berry site’s architecture in the broader context of Native American construction practices prior to European contact.

Interregional Context and Comparison Wood Types The use of pine in Native American durable construction, particularly for central support members, has been widely documented at other sites around the region, especially in Alabama, Georgia, and Florida, where this wood has been verified repeatedly as a key element in several Mississippian or contemporaneous structures (Lacquement 2004, 2007b; Lacquement, ed. 2007; Newsom and Quitmeyer 1991). The extensive use of pine at many of these sites probably relates to their location in the pine savannah forest zone of the southeastern Coastal Plain, in addition to pine’s suitability for construction purposes. A superb comparative sample exists in David Hally’s (2008) intensive analysis of Mississippian domestic and public architecture, especially those buildings reported through his excavations at the

218 · Lee Ann Newsom

King and Leake sites in northwestern Georgia.4 Although these sites are located in the Valley and Ridge physiographic province, north and west of the pine savannah belt, pine is—or historically was—a common constituent of the predominantly oak-pine or oak-hickory-pine forest characterizing this area (Hally 2008: 27). The same is true of the Berry site’s western Piedmont locale, with pine having been a regular though generally minor element of the former chestnut-dominated hardwood forest (Freinkel 2007). Lacquement’s (2004, 2007b) research indicates that pine was preferred for larger, standing or set pole/post construction. Conversely, pine was avoided for bent pole forms of construction, possibly because of its adverse wood technological properties, as demonstrated experimentally (Lacquement 2004, 2005; see also Reed 2007). As noted above, pine was used for all three of the central support posts excavated in Berry Structure 1, as well as for several of those posts that formed the outer wall structure. Likewise, pine was the wood most commonly used for posts in Berry Structure 5. Altogether, pine constituted about 40 percent of the posts recovered from the Berry site structures, and while at least four other hardwood taxa were also used for posts in these buildings, no other softwoods (conifers) were identified in the Berry assemblage. In contrast, structures at the Kincaid site in southern Illinois included posts of both red cedar (Juniperus sp.) and bald cypress (Taxodium sp.) (Brennan 2007); at the Mitchell site, also in southern Illinois, structures were principally supported by large cypress posts (Newsom, personal observations). Michael O’Brien (2001: 156) reports that bald cypress was preferentially used for large support posts at the Snodgrass site in Missouri. Otherwise, archaeological evidence for the architectural use of these two additional conifers in the Midwest or elsewhere is lacking to my knowledge, although I can point to one ethnographic record of Siouans in eastern North Carolina making use of “very long poles” of both pine and cedar in house construction (Swanton 1946: 410–411; see also Reed 2007). Thus, it would be unsurprising to find one or another type of cedar as more architectural remains from the Berry site are analyzed. Structures excavated at the Hovey Lake site in southwestern Indiana (Munson et al. 2009), like many others north of the southern, pine-dominated forest belt (Reed 2007), were built exclusively of hardwoods. Both red- and white-type oaks were used for upright posts in the two Berry site structures, and red oak in particular was predominantly used for the other nonpost construction elements constituting both structures (40 to 50 percent of the respective assemblages [see tables 5.4 and 5.5]). The red

Wood Selection and Technology in Structures 1 and 5 · 219

oak dominance at Berry is consistent with wood construction for a set of relatively late buildings at the aforementioned Hovey Lake site, as well as for the wood assemblages from Powers phase structures at the Turner and Snodgrass sites in southeastern Missouri (O’Brien 2001: 151–157). Otherwise, previous research on Mississippian structures indicates a preponderance of hickory and the white oak anatomical group; while ash and a few other hardwoods have also received mention, they appear to have been used to a much lesser degree (e.g., Lacquement 2007b; Reed 2007). Black locust seems to have been preferentially used for posts around the general region, including at the Hovey Lake site. Black locust, pine, and sassafras were all considered preferable for durable post construction in Chickasaw winter houses (Hudson 1976: 214). Similarly, Chickasaw summer houses had posts made of “rot-resistant pitch pine or black locust” (Hudson 1976: 216). “Pitch pine” is another name in the forest products industry for the southern hard or yellow pines, the same type identified among the Berry post samples. Alternatively, this vernacular may have been a direct reference to a particular species of pine, Pinus rigida, which under any circumstance belongs to the southern hard/yellow pine group. The now virtually extinct American chestnut was also documented for the two structures at the Berry Site, including both post and nonpost construction elements. This tree was for millennia the dominant species in the Appalachian Mountain forest ecosystem, before having been largely eliminated by the chestnut blight introduced early in the previous century (Freinkel 2007). Thus it is not surprising to find chestnut in the Berry wood assemblage, given the site’s location in the foothills of the Appalachian heartland. Chestnut, in fact, was the second most intensively used wood for posts in Berry Structure 1 (about 28 percent), and it was also fashioned into a large plank and several other substantial timbers, as earlier described. There exists a parallel here in particular with the Hovey Lake site, since chestnut wood was a major construction material there, as well (Munson et al. 2009). Hovey Lake may be the only other site in the region with evidence of chestnut wood in aboriginal construction. This is surprising, since chestnut is an excellent building material by almost any measure and shares many characteristics with oak, both anatomically and in terms of its wood technological properties (USDA Forest Service 1974). Chestnut was once a highly sought-after and economical wood, given its formerly dense forest stands, although advanced tools and technologies were required to harvest and convert the boles of particularly

220 · Lee Ann Newsom

large individuals into various classes of lumber. These advances resulted in great lumber yields per tree and over vast acreage, such that chestnut was a huge boon to the American timber industry while it lasted (Freinkel 2007). Before the blight, chestnut was the principal domestic source of tannin, which was obtained by soaking wood chips in hot water and evaporating the resulting liquor. Chestnut was also used for fiberboard, poles, fence posts, and railroad ties, as well as for furniture and other smaller wood articles (Panshin and de Zeeuw 1980: 560). Chestnut was considered especially suitable for railroad ties because of its natural durability, strength, and hardness, and it was a leading wood for coffin manufacture because of its outstanding durability and good working qualities. Given both its former widespread availability and the fact that it was a superior wood for durable construction, chestnut timber’s lack of documentation in Mississippian and other Native American architecture in the Eastern Woodlands—aside from our work at Berry and that of Cheryl Munson and colleagues (2009) at Hovey Lake—is striking. For example, the known prehistoric sites in the general range of the tree with structures that have been excavated and the wood types documented (mainly in eastern Tennessee, northern Alabama, and Georgia) have revealed no record of chestnut use. Other states with documented Mississippian structures (Missouri and Illinois, for example) are largely or wholly beyond the tree’s geographic range. Perhaps chestnut went unused as a wood resource because of technological constraints or some cultural proscriptions. Another possibility is that chestnut has been overlooked by individuals working to identify archaeological wood since it has strong similarities in anatomical structure with the white oak anatomical group and may thus be readily misidentified as wood of that genus. Building Components For the most part, the Berry site structures predictably and understandably have several features in common with well-documented examples built by southeastern Indians. Given the relative abundance, availability, and use of basic raw materials, as well as the potential for shared or overlapping cultural traditions, there is every reason to expect that buildings constructed at Joara fully or partly by Joarans would be consistent with the general pattern from neighboring areas. To this point, the square floor plan of Structure 1, its floor basin, its four prominent central support posts, and the definitive entryway—together with its inferred cane wattle and daub walls

Wood Selection and Technology in Structures 1 and 5 · 221

and bark or cane matting and roofing—are characteristics that are consistent with Native American building traditions in the region (e.g., Muller 1997: 186–190; and see Hudson 1976: 91, 213–216; O’Brien 2001). In general, Structure 1 conforms to what Lacquement refers to as a “large individually set post design” (2007a: 4). Structure 5 deviates from this pattern by the apparent lack of a house floor basin and its less substantial central supports. Hally (2008: 50–53) records that each structure at the King site was erected above a substructure floor basin and that the overall footprint or floor plan of domestic structures is square with rounded corners. He further specifies that the exterior walls of the King site buildings were of single-set posts (as alluded to above). Considering the relative size and scale of individual domestic structures, Hally reports that exterior walls ranged from 5 meters to 10 meters in length, with an average of 7.3 meters.5 Berry Structure 1 is roughly similar in overall size, with about 56 square meters in floor plan and spanning about 7.5 meters on a side. To reiterate, Structure 5 was a bit smaller at about 50 square meters, or 7 meters on a side. Muller (1997: 186–187) reports a general range of about 20 to 35 square meters in area for typical Mississippian residential structures by the mid-thirteenth century (noting also that earlier structures were on average smaller [e.g., Snow and Stephenson 1998: 102]), indicating that the medians for floor area ranged from a low of about 12 square meters to over 52 square meters (see also Mehrer 1995: 76–79). Moreover, Michael O’Brien (2001: 143) specifies means of 26.4 square meters and 24.1 square meters for the Turner and Snodgrass sites, respectively. With regard to the selection of raw material for the building framework and roof trussing, Hally (2008: 315) notes that the most common species used for construction purposes at the King site, particularly for posts, was white pine (P. strobus, of the soft pine anatomical group [Panshin and de Zeeuw 1980: 439–440], which is distinct from the type of pine used for the Berry structures). He reports that fifteen carbonized remnants of posts from five domestic structures were all pine, and he also refers to a sample of nineteen posts from Structure 1 at the contemporaneous Leake site (citing Patton 1990: 26), all of which also proved to be white pine. Moreover, Hally was able to specify several additional details that provide further insights into wood selection and transformation as building material. With some exceptions, his general impression is that most of the pine posts derived from “young trees” (Hally 2008: 53) that were then used either whole or split into roughly equal proportions. Many posts were observed to have

222 · Lee Ann Newsom

been approximately squared or angular in cross section, probably the result of a trimming operation to remove bark and/or sapwood (the outermost growth increments).6 Over the course of his excavations at the King site, Hally (2008: 50, 52–53) recorded some 9,400 postholes, among which were several that were clearly intended and positioned as central support posts for domestic structures; other similarly substantial posts were positioned at the corners of rectangular buildings that he also interpreted as houses (Hally 2008: 121). The typical pattern for domestic structures appears to have been as previously described: four interior roof support posts that were evenly positioned in the center of the building, forming a square around the central hearth. Hally (2008: 118–119) further indicates that these interior roof support posts, as well as the large corner posts of the rectangular buildings, were consistently about 21–23 centimeters in diameter or greater; otherwise, the majority of postholes in the habitation zone, including exterior wall and other posts, were recorded as 15–18 centimeters in diameter. Berry Structure 1 appears to conform very well to this building template, having the four central roof supports and multiple additional posts serving as exterior wall supports. The three excavated central supports for Berry Structure 1 were also all pine, although of the hard/yellow anatomical group, and they were generally of about the same size as the central support posts documented for the King site, with the carbonized post remnants measuring 14–17 centimeters in diameter (average, 15.6 centimeters; standard deviation, 1.52 [see table 5.2]). To reiterate, the complete assemblage of carbonized posts from Berry Structure 1 (n=18 [table 5.2]) indicates a range of 4–17 centimeters with an average diameter of 10.37 centimeters (standard deviation, 3.86). Again, this fits quite well with the King site data. Hally (2008: 70–71) likewise reports on an assemblage of charred posts from the King site that represents more than four dozen exterior wall posts from six different structures. These were observed to have been primarily wedge shaped in cross section, the result of having been “split from debarked tree trunks” (Hally 2008: 71). In other words, the pine boles were felled and the bark removed, after which individual logs were split along the radii to form three individual wedge-shaped posts, and possibly more depending on the size of the original trunk. These post splits were partitioned from trunks with a maximum estimated diameter of 24 centimeters. Carbonized remnants of outer wall posts for structures at the Leake site (Patton 1990, cited in Hally 2008: 71), which were also found to be white pine, were similarly prepared; about 75 percent were split as indicated

Wood Selection and Technology in Structures 1 and 5 · 223

above, and collectively the posts were estimated to have measured about 14 centimeters in maximum dimension, very consistent with the split pine posts from the King site. I want to make mention also of one larger, public structure documented by Hally (2008: 126–127, 130), because it provides more information on post diameters. This is the building designated Public Structure 17, with a total floor area of 212 square meters. Public Structure 17 had eight roof support posts, seven of which were represented by large postholes spaced at uniform distances from one another and likewise from the exterior walls and hearth. Three or four smaller postholes existed in the area where the westcentral roof support post should have been. This is somewhat reminiscent of the paired support posts for the Berry site’s Structure 5, evidently a similar compensation for the use of smaller-diameter posts by compounding their presence in a single location. The building may also have had interior partitions that were defined by posts about 18 centimeters in diameter, while the diameter of the exterior wall posts averaged about 20 centimeters (the equivalent data for postholes representing the roof supports are not specified). Despite the fact that the majority of posts at the King site derived from relatively young stems, the above data clearly convey that builders at King and Leake did also manage to fell and dress some fairly substantial timbers. None of the diameters for posts or other roundwood from Berry exceeded 17 centimeters (see tables 5.2–5.5), although several of the Berry site timbers did represent trees of considerable age. Moreover, it is important to recognize that felling a large white pine, which is relatively soft, even textured, and straight grained (Panshin and de Zeeuw 1980: 441, table 11.2), is not the same thing as felling a similarly sized hard/yellow pine, which has tougher, denser latewood per unit-area of stem (Panshin and de Zeeuw 1980: 446, citing Kukachka 1960). Growth rate, especially in pines, is also very important because faster-grown stems will have a greater volume of softer, juvenile wood. Nor can harvesting pine boles be easily equated with felling mature hardwoods such as those that predominate in the Berry site assemblage, particularly the oaks, chestnut, and black locust, the last being especially hard and dense (Panshin and de Zeeuw 1980: table 11.4). In other words, diameter class is not the entire story, and it is important to understand the interplay of fundamental wood property distinctions between taxa as well as the relevant functional anatomy (age distinctions). From my perspective, these distinctions are also essential for detecting any influence of the European contingent at the Berry site on what would oth-

224 · Lee Ann Newsom

erwise appear to represent fully Native American construction practices (at least for Structure 1). I will return to this point after providing a bit more regional perspective on basic construction details, for the sake of further comparison. In discussing the superstructure or roofing components of domestic structures reported at the King site, Hally (2008: 54) suggests that notched and interlocking timbers were probably associated with the roof framework, and he notes that the roofs of such structures were likely covered with sheets of bark or thatch. Similarly, notched roof timbers have been reported at the Kincaid site (Brennan 2007), with the notching specifically located on the downward end of adjacent roof timbers. At least one notched pine timber was located among the burned building debris on the floor of Berry Structure 1, perhaps similarly part of the architectural roof complex (see figure 4.4a). As discussed above, this timber is relatively short (40 centimeters) and has conspicuous notching at both ends; that the notches face in opposite directions suggests that this specimen was some kind of interlocking element between two sets of timbers (akin perhaps to the Chickasaw winter house roof assembly that included “notched poles interwoven with split saplings or white oak splints” [Hudson 1976: 214]). Hally (2008: 92–93) also describes roofing evidence from log impressions in daub associated with Structure 1 at the Leake site. There, about 70 percent of impressions suggested round cross sections for the timbers deployed in roof construction, with most of the remaining examples indicating that the original boles were split into segments to produce two or more poles. Whether used whole or split, all logs seemed to have been prepared for construction by removal of the bark. Hally reports the use of smaller-diameter stem, roundwood, or subdivided roundwood segments as wall-wattle framework at the King site, noting that this framework was probably covered with woven cane mats or mud plaster. The split roundwood recovered from both the King and the Berry site may also be evidence for superstructure/roofing material, something that has proven elusive in descriptions and analysis of other structures documented in the general region (e.g., Lacquement, ed. 2007). The hickory samples from both Berry structures are consistently among the smaller-diameter, younger roundwood elements, including some split in half (see tables 5.4 and 5.5). Although the Berry site samples do not have notches or other definitive evidence of modifications to join them with adjacent rafters or framing, they are likely to represent some of the original rafters or other superstructure elements.

Wood Selection and Technology in Structures 1 and 5 · 225

As for the final roof and wall coverings, evidence suggesting that bark slabs were used throughout the region as a roofing material indicates that Native Americans removed bark specifically for that purpose (Lacquement, ed. 2007). Berry site roundwood specimens—including posts with facets and other cut marks along their outer or tangential surfaces—are presumably evidence of bark removal practices. We may infer that the sheets of carbonized hardwood bark recovered from both Berry structures (provisionally oak, hickory, and chestnut) likely served as shingles or wall siding, though as we will see, there is also evidence from Structure 5 for the use of roof shakes in the European tradition. Likewise, it is probably safe to assume that the cane stem fragments and samples of woven split cane represent elements that were incorporated in the superstructure, walls, and flooring, if not in the interior furnishings. Cattail (Typha sp.) and possibly another type of monocot have also been detected among the carbonized debris found on or in proximity to floor deposits in both Berry structures. This is consistent with archaeological reports from around the region that note the presence of cane, sedges, or robust grasslike forbs as finishing components or matting for Mississippian structures (e.g., Hudson 1976: 215; Reed 2007). For example, cattail was used as thatch or matting in a structure at the Kincaid site (Newsom, lab data 2009), and bulrush (Scirpus sp.) was identified as thatching at the Rench site, also in Illinois (McConaughy 2007). At least one ethnographic case indicates that cattail has long been an important material for thatching and floor covering in Anishinaabek summer and winter houses (Herron 2002),7 and dried grass was incorporated with the clay plaster used to coat the underside and outer surfaces of Chickasaw winter houses (e.g., Hudson 1976: 214). In addition, pine bark or grass thatch was systematically laid across the roof tops of the Chickasaw houses and held in place with rings of split saplings. Furthermore, and specifically concerning eighteenth-century Cherokee dwellings and the use of bark as a construction material, Bartram (1995 [1791]: 84) relates that chestnut bark sheets covered entire structures, the walls as well as the roofs. And to reiterate, there is a suggestion that moss and perhaps even a bald-faced hornets’ nest served as packing and insulation for Structure 1 at the Berry site. Finally, Hally (2008: 53) also documents important details about the entryways of domestic structures at the King site, specifically, narrow passageways that were only about 70 centimeters wide and extended about 1–1.5 meters beyond the exterior walls. He inferred that the walls lining these passageways were likely made of wooden planks placed on end in

226 · Lee Ann Newsom

shallow trenches and roofed over with sheets of bark. This is very reminiscent of the well-defined entryway for the Berry site’s Structure 1, which appears to have been of similar form (plank lined or sided—recall mention above of OG#84 from Unit 38, the large, red oak “doorway timber”) and of about these same dimensions (if a bit longer [see plate 7]; Hudson [1976: 214] mentions passageways of 6–7 feet in length, or more than 2 meters). Construction Process Continuing to think about the harvest and reduction of raw material, along with relative wood quality, I return now to a consideration of form (stem/ bole completeness), tool marks, and the maturation states of material recovered from the Berry structures. The practice of splitting larger boles to create more posts relative to each bole harvested, and then setting the split lengths of bole as individual poles (recalling those used for the outer wall posts at the King and Leake sites), is distinct from what was documented for the two Berry structures. All of the posts recorded for Berry Structures 1 and 5—both those that served as central supports and those that were part of the outer walls—were set as entire rounds, that is, boles complete in their original circumference, with the exception that the bark was removed. With respect to split versus whole trunks, Hally (2008: 71) makes an important point when he alludes to the presence of juvenile versus mature wood, calling attention to the fact that smaller boles such as the majority noted above for the King site were predominantly of juvenile wood. From a construction standpoint, these would have been weak in comparison with wedge segments from relatively older, mature trees. I later return to this question of juvenile wood anatomy, but lacking more information, determining whether the Berry posts with similar diameters were harvested from relatively older trees (versus the majority of the material documented for the King and Leake sites) is difficult. As indicated above, however, many of the Berry samples had well surpassed the juvenile stage of growth. If generally older, larger boles were felled for the Berry site structures—at least for Structure 1—it would be interesting to know whether that in itself is a reflection of the use of metal tools (recall my discussion of sharp cuts and the use of a saw in these samples). The relative maturity of wood boles felled for use, as reflected in age and diameter classes, may serve as a proxy for the amount of effort required to secure suitable wood for construction. Likewise, both the maturation state of individual timbers and the collective proportion of younger versus older material, including heartwood versus sapwood, may provide an indication

Wood Selection and Technology in Structures 1 and 5 · 227

of the overall durability or structural integrity of the architecture. All of this, in turn, provides further insights into wood selection, if not time investment, for building projects. In reiteration, the Berry Structure 1 posts ranged from 15 to 100 years of age at harvest, including an age spread for four pine posts of 56 to 100 years. The post diameters for this structure are correspondingly large, with the central pine supports ranging from 14 centimeters to 17 centimeters in diameter. The nonpost roundwood and other timbers from this same structure included fourteen individuals in the range of 15+ years, with two timbers being 70 and 132 years of age at harvest (see table 5.6). When the twenty-seven items from the whole assemblage having age estimates (discounting two 14-year-old roundwood specimens that probably equate with a third [see table 5.6]) are examined, the average age is 32.4 years, with a standard deviation of 32.2 years (minimum age is 8 years; maximum age is 132 years). The point I seek to convey is that wood formed during the first 14 or 15 years in the life of a tree, and perhaps a bit longer in some species, generally describes the period of juvenile growth and its associated wood anatomical variation. Juvenile wood is not fully consistent with or representative of the adult/mature condition. It is generally easier to cut and shape but tends to be weaker and much less durable than mature wood of the same taxon. Therefore, while much of the roundwood that was used for the various nonpost architectural elements of Structure 1 was certainly juvenile wood, the builders of the structure also chose to harvest some fairly substantial and mature trees, particularly for its upright posts. In contrast, the Structure 5 wood assemblage was characterized by younger material that was smaller in size than the Structure 1 assemblage. Recall that one pine center post from Structure 5 was 17 centimeters in diameter and at least 47 years of age at harvest (see table 5.3), but the overall assemblage demonstrates an average age of 13.7 years (n=31, discarding duplicates; standard deviation, 8.75 [minimum age is 5 years; maximum age is 47 years]). Compared with Structure 1, at least, Structure 5 included proportionately more juvenile wood, with the average age being about half of what characterized the other assemblage (13.7 years versus 32.4 years). Whatever the reasons, the prevalence of juvenile wood and wood of smaller diameter in the architecture of Structure 5 would have reduced the overall durability of the building. That is, the proportion of relatively immature growth used in its construction indicates that Structure 5 would have been much less sturdy than Structure 1 and perhaps rather flimsy, suggesting its use as a temporary or short-term building, one that perhaps served a

228 · Lee Ann Newsom

different purpose from the more permanent and enduring Structure 1 (see chapter 4 for a discussion of Structure 5 as a possible detached kitchen). Thinking back to the larger, older posts from the King and Leake sites, as well as to the very large posts from the Mitchell site and to those from the Turner and Snodgrass sites that were reportedly in the range of 35–40 centimeters in diameter (O’Brien 2001: 146–147), we know that precolonial Native American technology could accomplish the harvesting of very large boles. Experimental archaeology has demonstrated that such boles may be laboriously cut down with stone tools (Hammerstedt 2005), and archaeological research has revealed some of the equipment that may have been used to do so. Hally (2008: 481), for example, recovered woodworking tools from burials at the King site, including beaver incisors, unifacial end scrapers, and stone celts. Large trees may also have been girdled and left standing to die in place and then be toppled over for greater ease of harvest several months later, but this practice would have resulted in diminished wood quality and represented a lag time before the wood was available for use. As I have noted, several of the pine posts and other timbers that constituted Structure 1 at the Berry site represent reasonably large and mature trees. American chestnut in particular was a true forest giant, with mature individuals attaining 100 meters or more in height and having considerable girth (Freinkel 2007). Even though most of the wood incorporated into the Berry site architecture was from trees that lacked such monumental proportions, the chestnut uprights from Structure 1 did range from about 5.5 centimeters to 12 centimeters in diameter, and at least one individual was a minimum of sixteen years of age at harvest. At this point I want to return to the large chestnut plank from Structure 1 (OG#2003a [see figure 4.4b]). This plank preserves a minimum of twenty-eight years of growth from a tree that would have been several times that old at harvest, based on the positioning of the growth increments and the minimal growth ring curvature. The method—or, more precisely, the orientation—that was used to create this timber was earlier indicated as plane sawing, which is plank separation along the tangential dimension (Hoadley 2000). This contrasts with quarter sawing and other ways of dividing timber. Both plane and quarter sawing can be traced to colonial European and early American practices, back through even earlier European traditions. It is important to convey that this timber represents a tangentially oriented separation from the bole, perpendicular to the radius, that is less readily achieved in comparison with quarter-sawn cuts, which follow natural planes of weakness in a tree, such

Wood Selection and Technology in Structures 1 and 5 · 229

as the ray system. The wedge-shaped timbers that Hally describes from the King site, for example, are the equivalent of quarter-sawn timber, achieved by splitting the wood into thirds, quarters, or finer divisions along these inherent planes of weakness. If oriented correctly—and depending on the technology used, the size of the trunk, and the details of growth increments—plane sawing or splitting may sometimes take advantage of weaker zones within individual growth increments, but even so this method would still require a difficult maneuver given such a length of plank as that indicated by the Berry site specimen. In addition, this plank is mature, dense wood with relatively narrow growth increments and very minor curvature, indicating that it was harvested from a very large tree and probably from growth nearer the base of the bole (if it was harvested from higher up the stem, then this was indeed a tree of gigantic proportions). I am skeptical that such a large bole could have been effectively managed for creating large planks such as the Berry site specimen without metal carpentry tools, specifically, something like a crosscut saw. Indeed, saws, axes, wedges, and chisels were among the equipment accompanying Pardo’s expedition (see table 8.4), and this plank in particular is a good indication for the presence and use of these tools, probably in addition to European woodworking knowledge and practice. In its most basic design details, Berry Structure 1 thus conforms quite well to traditional Native American construction practices, lending weight to the inference that native Joarans were indeed involved in the building project. But the use of relatively mature, whole timbers as posts for the building, as well as indications of what are most likely European practices such as plane sawing, are details that distinguish it from Native American architecture and may reflect instead the presence of Europeans and the use of metal tools. Such woodworking equipment would have enabled the builders of this structure—whether Europeans, native Joarans, or both working together—to secure larger boles from some relatively dense and strong woods for more-substantial posts that were then set in the ground as full rounds. Recall, too, the timber from Structure 1 (OG#81) that exhibits definitive saw marks over its full cross section; several others already mentioned suggest the use of a saw and other metal tools to perform raw material extraction and reduction. For example, several specimens from Structure 1 have very sharp bevels, and one had a swath of excised growth rings that included a knot (that is, it was worked across the grain). At least five of the timbers, including plank-like forms and posts, had been trimmed into somewhat angular forms, including the doorway timber and OG#44, the

230 · Lee Ann Newsom

latter of which was minimally 132 years old. The shaping of timbers into roughly quadrilateral cross sections by the removal of bark and secondary limbs (that is, perpendicular growth) does not in itself distinguish this material from fully Native American assemblages, but the prevalence of long, sharply defined cuts and the presence of the plane-sawn, chestnut board— among other examples—strongly deviate from traditional patterns. Structure 1 also had the timber with a square hole penetrating transversely and intersecting five growth increments, suggesting that someone had hammered a large metal spike into the top of the timber. Although Structure 5 differed in several respects from Structure 1 (as discussed above), many examples of roundwood from Structure 5, like those of Structure 1, had been cleanly cut down the center line through the pith, and numerous other timbers exhibited precise cuts—especially bevels—and not simply splits along the grain. The construction debris from this building also included one plank-like timber, the two specimens with apparent nail holes, and thirteen slat-like forms (“strips”) all made from pine. These last items, one of which is among those with an apparent nail hole, are highly suggestive of wooden shakes or shingles. The use of shakes, cut and split into fairly standard sizes to cover the roof or sides of a building, may represent an exclusively European tradition and differs markedly from the variously sized sheets of bark or woven cane that were used for such coverings in Native American architecture. We cannot be certain, but if these slats were indeed shakes or something similar, then these specimens and the suggested use of iron nails serve as strong evidence that Structure 5 was built exclusively by the members of the Pardo expedition, or at least that they were even more involved in its construction than in that of Structure 1. The morphology of the tool marks just described strongly suggests the use of metal tools—rather than those of stone or shell—for much of the wood recovered from the Berry structures. The cut marks on prehistoric wood debitage and artifacts, both of which are abundant on the Florida wet sites where I have had opportunities to observe and study them, tend to be shallow, short, narrow, and curved (e.g., Newsom et al. 2013). In contrast, those displayed on the Berry site material are relatively long, wide, and severe (that is, sharp and angular), much like the tooling marks I have observed on debitage from Spanish colonial deposits at St. Augustine, Florida (Newsom 1989b), and at an early fortification in Old San Juan, Puerto Rico (Newsom 1996). Moreover, the Berry site assemblage contains several examples of completely lopped-off roundwood sections, with clean separa-

Wood Selection and Technology in Structures 1 and 5 · 231

tion perpendicular to the grain, as well as other examples in which the wood was cleanly cut in two right down or near the center line. While these specimens do not necessarily indicate the use of metal tools, in my experience studying large assemblages of carbonized, aboriginal building materials such as wood structural debris from the Powers phase sites in Missouri, they are inconsistent with Native American practices. Thus, the tool marks and evidence of material reduction on these Berry site specimens are better explained by the use of European tools than by the use of aboriginal woodworking equipment.

Conclusions A combination of traditional Native American construction practices and evidence for the use of European carpentry tools and techniques suggests that a joint effort ensued between Joarans and Spaniards in the building of Structure 1; that is, some degree of mutual cooperation most parsimoniously explains this structure. Structure 5 is somewhat odd, although it may conform to what Lacquement (2007a,b) refers to as a “composite” building type. Nevertheless, Structure 5 deviates in several important respects from traditional Native American architecture, particularly in the suggested use of wooden shakes or siding and iron nails. This structure is most plausibly explained as a building assembled by the Spanish contingent, and the question remains as to whether its smaller size and less durable nature resulted from a perceived need for expedience or were related to increasing tensions between Joarans and Spaniards that might have limited access to raw material. Of course, neither of these possibilities is mutually exclusive.

Notes 1. Ash (Fraxinus sp.) wood may also be assigned to either of two anatomical groups, one representing wetland taxa and the other “upland” forms (Panshin and de Zeeuw 1980); however, the material from the Berry site could not be distinguished to this level. 2. All upright posts were set as complete segments of bole (i.e., trunk), which is evident from their intact cross sections, or in some cases by the clear suggestion of completeness from the associated fragmentary debris and/or sediment impressions showing that the entire circumference was originally present; all are therefore scored as “1.” 3. I am intentionally ignoring finer segments because of the potential for fragmentation into quarters and finer segments with carbonization. 4. There are few published analyses of architectural components from contemporaneous sites in western North Carolina. Hally’s published discussions of architectural remains

232 · Lee Ann Newsom

in northwestern Georgia offer far and away the best comparative data for the Berry site wood. 5. All of Hally’s figures, reported in standard English units, are converted here to metric. 6. Stripping and cutting away the outermost growth rings and bark is a practice that has a long history in European and Native American wood technology (to leave these components intact invites insect problems because the bark and sapwood contain the photosynthates, i.e., carbohydrate-rich metabolites undergoing storage or transport along the length of the trunk, branch, and root network). 7. The Anishinaabek are American Indian peoples of the northern Great Lakes region, specifically, Ojibway, Odawa, and Potawatomi.

IV What They Ate Politics, Food, and Provisioning

The woman who, in time, would come to know herself as Luisa Méndez, but who for now was a captive among her enemies and the strangers who had befriended them, rose from her work and opened and closed her fingers, opened and closed her fingers. It was midday, and since dawn she had labored crushing hickory nuts and kneading their meat and oil and small bits of shell into a thick paste that she rolled into balls the size of her fist. Later that day, she and the other women held captive—taken down from the mountains beyond this accursed place when the homes of their mothers and their mothers’ mothers were burned to the ground—would make a rich soup from the paste; then her enemies and their new brothers would feast into the night. But tomorrow, most of these new men would depart, and she with them. For weeks she had accompanied them, from the smoldering ruins of her own town, her beloved Maniatique, to Chiaha, where she was astounded when her captors were met by so many others of their own kind. Yet even in such numbers, their fear of Coosa and his allies, lying in wait along the path ahead, had driven them back here to this place of her enemies, this Joara. When she recalled her captivity in years to come, she would pull her rosary close and cross herself. Now, though, the anger and pain still sharp, she silently called down on them all, old foes and new alike, the fires that had consumed Maniatique.

*

*

*

Food—whether procuring it, preparing it, or consuming it—was simultaneously among the most integrative and yet potentially dangerous domains

234 · Part IV. What They Ate: Politics, Food, and Provisioning

of the colonial encounter, for the men of Fort San Juan and their hosts alike (e.g., Earle 2012). Pardo was provided with only 800 pounds of biscuit (about 6 pounds per soldier), 24 liters of wine, and 10 cheeses (Hudson 1990: 341). As a result, the men whom Pardo left behind were all but dependent on Indians for their daily subsistence needs. In part 4, “What They Ate,” Gayle Fritz and Heather Lapham present their analyses of plant and animal remains, respectively, from contexts inside and outside of the Spanish compound. In chapter 6, “People, Plants, and Early Frontier Food,” Fritz discusses the archaeobotanical evidence of food preparation and consumption, observing that most of those plant foods consumed within the compound area were prepared by native women within the compound itself. These foods include maize, beans, squash, and various wild nuts that were used in stews and gruels. Significantly, her analysis indicates that plant remains recovered from structures and features in the compound area are little different from those recovered from nonSpanish parts of the site. That is, the meals that the native women of Joara prepared for Pardo’s soldiers varied little from the meals they prepared for their own families. Lapham’s analysis of the faunal remains tells quite a different story. In chapter 7, “Fauna, Subsistence, and Survival,” she shows that animal remains recovered from the Spanish compound differ in important ways from such remains recovered elsewhere on the site. Most important, the quantity of bear meat observed in compound features is far higher than in features located south of the compound; moreover, there is less evidence of bear meat processing in the compound, and the choicest cuts of meat—particularly limb bones—dominate the assemblage of bear remains. For native groups across the American South, bear meat was generally considered a high-quality food suitable for entertaining guests. If, as we can safely assume, most of this bear meat was procured by adult men, then the men of Joara were gifting this delicacy to soldiers in the compound. The pattern of provisioning identified by Fritz and Lapham, with Indian women preparing plant foods and meals within the Spanish compound and Indian men providing select cuts of bear meat processed outside the compound, gains additional significance when interpreted through the temporal lens of early and late features. Plant food processing and consumption varies little from early to late contexts within the Spanish compound, suggesting that native women played a major role in the soldiers’ daily subsistence from the founding of Fort San Juan to its ultimate, untimely end. Indeed, some women may have been living with soldiers inside the compound. The amount of bear meat, however, falls dramatically in late contexts, suggesting that native men

Part IV. What They Ate: Politics, Food, and Provisioning · 235

had stopped provisioning the soldiers with this delicacy. This insight coincides with the slapdash construction of Structure 5—a later building that evidence suggests was built by soldiers working on their own. Some native men and women may thus have had quite different experiences with and reactions to the Spanish occupation, a divergence that hastened the end of Fort San Juan.

This page intentionally left blank

6 People, Plants, and Early Frontier Food Gayle J. Fritz

The archaeology of early interactions among indigenous American Indians, Africans, Spaniards and other Europeans in eastern North America is extensive and impressive, and food plays a major role in ongoing efforts to understand social dynamics and changes that occurred as hemispheres collided. In the region now known as the southeastern United States, pioneering studies by Kathleen Deagan (1973, 1983), Stanley South (1982), Elizabeth Reitz and C. Margaret Scarry (Reitz and Scarry 1985; Scarry and Reitz 1990), Lee Ann Newsom (1989a), Paul Gardner (1980, 1982), and others have provided data and insights from excavations at sixteenth-century settlements including St. Augustine, Fountain of Youth, and Santa Elena. Deagan and her colleagues formulated the “St. Augustine pattern” with its elements of syncretism and creolization resulting from long-term or less-permanent relationships between native women and colonizing men. Indigenous wives, concubines, and servants prepared traditional foods in the private, domestic sphere, while European men of high status hosted public meals that featured Old World ingredients and imported serving wares (Deagan 1983). The St. Augustine pattern informed, inspired, and influenced teams of archaeologists working at early colonial sites across the Western Hemisphere, although Barbara Voss (2008) and others (e.g., Jamieson and Sayre 2010) recognize the need to shift away from a focus on dualistic categories such as public versus private spheres and to emphasize economic domains other than intermarriage that may be more appropriate for regions beyond La Florida. Deagan’s St. Augustine pattern, despite inevitable refinements and improvements by later scholars, established gender roles as extremely important in the archaeology of colonialism, and this model deserves credit for the continuing focus on women as cultural brokers and social agents with control over many types of food.

238 · Gayle J. Fritz

St. Augustine and Santa Elena were founded in 1565 and 1566, respectively, as part of the plan of Pedro Menéndez de Avilés to secure La Florida for Spain and to exploit its resources. Captain Juan Pardo arrived at Joara in December 1566, so very little time had elapsed since the first permanent Spanish settlements on the North American Atlantic coast were built. The Spanish presence at Fort San Juan lasted for less than eighteen months and included only one full growing season. Sixteenth-century European presence at the Berry site was far more fleeting and less substantial than at either of the longer-lived coastal towns. All of the soldiers at Joara (the maximum contingent at any one time was 130) certainly interacted with native women as well as men, but there was insufficient time for stable, independent households to be established that exemplify mestizaje, the genetic and cultural legacy of native women in creating a new domestic world order. Still, we know that the lone survivor of the attack on Fort San Juan, Juan Martín de Badajoz, married a woman (Teresa Martín) whom he met either at Joara or at one of the other towns between Santa Elena and the Appalachian Mountains (Hudson 1990). Despite Fort San Juan’s brief temporal span and essential military nature, Joaran women likely played a central role in basic domestic activities, especially those concerning food. The emphasis that Deagan and others have placed on gender therefore continues to be salient for Berry site research. Joara’s mico, who was male, promised Pardo that he would fill a house with maize for the soldiers, but maize was probably grown by women, and women produced, gathered, and processed the other plants that sustained both hosts and visitors, making these women active social agents (Hudson 1976). Voss (2008: 868) advises that gender should be understood as “a politically and socially charged arena of conflict and negotiation in Spanish colonial settlements,” and this is true even for the relatively short-lived Spanish colony at the Berry site. Most of the meals consumed within the compound took place during Pardo’s absence, with thirty or fewer men present; many meals involved only the ten Spaniards left behind by Hernando Moyano when he went out on prospecting or military forays. If we assume that women prepared and served these meals, shifting their tasks temporarily from the thriving adjacent town (where they also fed their families), then daily opportunities existed for interactions such as language teaching and learning, game playing, and the formation of various levels of friendships. Analysis of plant and animal remains from St. Augustine and Santa Elena has provided valuable comparative databases for understanding how

People, Plants, and Early Frontier Food · 239

sixteenth-century Europeans adjusted their diets to new environments and colonial milieus, and a wealth of fascinating information is summarized by Reitz and Scarry (1985; see also Scarry and Reitz 1990). Still, it is important to keep in mind that however infrequently the supply shipments arrived at the coastal settlements, and however imbalanced was the distribution of the most desirable foods among high- and low-status military men and other colonists, no equivalent stores reached Pardo’s forts in the interior. Moreover, there was insufficient time for peach trees to have grown to maturity at either St. Augustine or Santa Elena or for peach pits carried inland from coastal sites to have grown to maturity in the interior. Therefore, Old World foods played a much smaller role than they did on the coast, and any manifestations of status differences at Fort San Juan may reflect native rules and expectations as much as, or even more than, imposed Spanish patterns.

Methods of Field and Laboratory Analysis Charred macrobotanical remains were recovered by flotation using a modified Shell Mound Archaeological Project type of system (Watson 1976) with water pumped from Upper Creek, which flows past the eastern edge of the Berry site. Standard 10 liter flotation samples were taken from each zone of units excavated in Structures 1 and 5 and from each level of each feature inside and outside of the structures. Light fractions were floated through a 0.425 millimeter geological sieve, while heavy fractions were caught in window-screen fabric (1.6 millimeter mesh) lining the inside of the flotation barrel. Eighty-two samples are reported here: twenty from Structure 1; eighteen from Structure 5; two from a pit feature (Feature 92) underneath Structure 5; six from pit features in the central courtyard area of the Spanish compound referred to in this chapter as the “central pits” (Features 23, 25, 69, and 83; these pits, together with Feature 92, are included among Lapham’s “early features” [see chapter 7, this volume]); twenty-six from pit features immediately to the west of Structure 5, referred to here as the “western pits” (Features 66, 103, 106, 109, 111; these pits, excepting Feature 109, are included among Lapham’s “late features” [see chapter 7, this volume]); and ten from Feature 112, which is a large pit south of Structure 5. Two of the samples from Structure 5 and three from central pits were smaller than 10 liters in sediment volume floated, resulting in a total of 804 liters of sediment floated for the assemblage reported here. In the laboratory, each light fraction and each heavy fraction was weighed to the nearest 0.01 gram on an electronic scale, and the contents

240 · Gayle J. Fritz

were passed through a series of USDA geological sieves with mesh openings ranging from 4.37 to 0.35 millimeters. Remains larger than 2 millimeters were completely sorted according to category, counted, and weighed, with the exception of wood charcoal and bark, which were weighed but not counted. Below the 2 millimeter mesh size, contents of the various geological sieves were scanned, and all charred seeds or recognizable seed or cultigen (maize or squash rind, for example) fragments were pulled, but because of the impossibility of identifying such fragmentary nonseed material, the contents were not completely sorted. Acorn shell was sorted down to the 1.4 millimeter sieve size to accommodate its fragility and its lesser likelihood (in comparison with hickory shell or walnut) of being well represented in the larger than 2 millimeter splits.

Excavation and Sampling Contexts Structure 1 (twenty samples analyzed) and Structure 5 (eighteen samples) are thought to have been built for occupation by Pardo and his men, with Structure 1 having a deeper basin and Structure 5 generally appearing less sturdy and less skillfully constructed (see chapters 4 and 5, this volume). Structure 1 may have been one of the earliest buildings at Fort San Juan, involving the full cooperation and labor of native Joarans on behalf of the Spanish soldiers, whereas Structure 5 might have been a later and more unilaterally Spanish episode of construction reflecting a period of lessfavorable relations. Both buildings, however, were cleared of most artifacts prior to burning. The plant remains consist of wood, bark, resin, cane culm, other stem types, and remains of food and other material deposited in the floor that escaped being swept during routine housekeeping or final clearance before destruction. Feature 92 (two samples analyzed) lies under Structure 5 and apparently predates it, evidently having been filled with refuse while the soldiers lived in one of the earlier buildings, such as Structure 1, located to the east. Based on his analysis of pottery from this and many of the other pit features at the site, Trevor Martin (2010; see also Moore et al., chapter 9, this volume) identified Feature 92 as a context that possibly reflects postfeasting deposition. Potsherd density was high there, the proportion of nonlocal pottery was high, and there were more serving bowls than other vessel forms. Six samples came from pit features in an open courtyard-like area between the burned structures: one from Feature 23, three from Feature 25, one from Feature 69, and one from Feature 83. Special attention was paid

People, Plants, and Early Frontier Food · 241

to Feature 25 because Martin recognized additional evidence of feasting in this location. These features also yielded higher numbers of bear bones than elsewhere (Lapham, chapter 7, this volume). Although Feature 25 does not stand out in terms of unusual proportions of plant remains, the average density of food remains, especially corn and acorn, from this and nearby pits is far higher than any other pooled context examined during the present study. Intrasite variability is discussed in greater depth below. Whether the pit features west of Structure 5 had been dug originally for trash deposition is unclear, but they seem to have acquired that function before the native revolt. They are similar to the central pits in lacking the high densities of architectural debris such as wood, bark, and cane stem that were common in Structures 1 and 5. Although the western pits were not as rich as those centrally located between the structures, they contained higher average densities of plant food remains than did either of the buildings. Three samples were analyzed from Feature 66, four from Feature 103, nine from Feature 106 (divided among Sections 1–6 of that feature), two from Feature 109, and eight from Feature 111. Most of the pits were shallow and consisted of one or a few zones and levels, but excavations into Feature 111 went down nine levels. Plant remains were very sparse in the deeper levels of Feature 111, lowering the average ubiquity and density values of the western pits. Feature 112 is a large pit just south of Structure 5 that is distinctive in terms of the density of nutshell and other organic remains and the presence of large Pisgah potsherds. It measures approximately 2 meters in diameter and 45–50 centimeters deep (below the base of the plow zone). As described by Martin (2010) and discussed elsewhere in this volume (Moore et al., chapter 9), this feature had a single deep depositional zone (Zone 1), at the bottom of which were sherds from two extremely large jars. Below the sherds was a smaller, spatially restricted zone (alternatively called Zone 2 or Zone 1/2) with fill that was described as greasy. On the basis of this evidence for rapid deposition and the large jars, Martin argues that the contents may represent feasting. Seven of the ten analyzed 10 liter flotation samples came from Zone 1, one from Zone 1/2, and one from Zone 2. The tenth (FS 1002, Bag 192) was labeled Zone 3.

Plant Foods in the Spanish Compound No Old World cultigens have been identified in the Berry site Spanish compound assemblage, which is not surprising given the meager supplies listed

242 · Gayle J. Fritz

in the historic accounts and the fact that no pack animals were available. Only biscuit (hard tack), cheese, and wine were issued from the storehouse at Santa Elena (Hudson 1990: 126–127), and none of these would have left any macrobotanical evidence. The first Pardo expedition departed Santa Elena on December 1, 1566, with 793 pounds of biscuit, 23.8 liters of wine, and 10 cheeses. The second expedition, which left Santa Elena the following September, carried only 85 pounds of biscuit but 48.4 liters of wine. By this time, Pardo had good reason to expect that he could rely on stores of maize and other foods provided by the native chiefs. Excavations at Santa Elena have yielded figs, European hazelnuts, watermelon and other melon seeds, peach pits, peas, domesticated grapes (Vitis vinifera), olives, and wheat (Gardner 1980, 1982; Reitz and Scarry 1985: table 2), but most of these remains were fortuitously preserved in waterlogged barrel wells, none of which has yet been found at the Berry site. Santa Elena, as noted above, was occupied far longer than Fort San Juan, and the Spaniards there received regular shipments from abroad. Nutshell Acorn shell is outnumbered and heavily outweighed by hickory nutshell in the structures and most pit features north of the mound, with the central pit group being the only context where acorn outnumbers hickory (table 6.1). Nevertheless, acorn shell is well represented in the Spanish compound assemblage, occurring in 100 percent of the samples from the central pits, Feature 92, and Feature 112; in 80.8 percent of samples from the pits west of Structure 5; and in 83.3 percent of samples from Structure 5 (figure 6.1). Structure 1, however, is exceptional in its low ubiquity (15.0 percent) and extremely low density (0.02 fragments per liter of sediment floated) of acorn shell, and the western pits yielded far lower numbers of acorn shell than did the central pits. Feature 112 was very rich in acorn, with an overall density of 5.20 fragments per liter, and the central pits were even richer, with average density of 11.38 fragments per liter (figure 6.2). Kristen Gremillion also found considerable quantities of acorn shell and high overall ubiquity (72.7 percent) in the assemblage from the mound area at the Berry site, prompting her to speculate (after adjusting for differential preservation favoring thicker hickory nutshell and higher representation of acorn nutmeat per nutshell fragment) that acorns originally may have outnumbered hickory nuts as food resources at the Berry site (Gremillion 2002: 305). Although we have no way of knowing which nut type actually contributed more to the diet, it is evident that acorns as well as hickory nuts

People, Plants, and Early Frontier Food · 243

Figure 6.1. Ubiquity values (percentage of samples in which an item occurs) of maize and nutshell.

were important foods and that both the Spaniards and the native Joarans consumed considerable amounts of both. Maps in The Manual of the Vascular Flora of the Carolinas by Albert Radford and colleagues (1968) show seven species of oak trees native to Burke County and several others growing in nearby counties. All acorns in this region would have required leaching to remove the bitter tannins. Soldiers at Fort San Juan seem unlikely to have sought out or specifically requested acorns as food, given the typical grain-based Iberian diet and the possibility that acorns carried a stigma as low-status or famine food. Ubiquities and average densities, however, demonstrate that some forms of acorn-based products made their way into the provisions. Whether these foods were breads, soups, or porridges, they were evidently palatable enough to the Spaniards to be accepted. The low visibility of acorn shell in units from Structure 1, however, warrants further discussion, especially in light of the fact that most of these samples yielded hickory nutshell and maize. If this was one of the first

Table 6.1. Macrobotanical remains from the Berry site Spanish compound Structure 1

Structure 5

20

18

200

169

Samples (n): Samples (liters): Common name

Scientific name

Count

Wt. (g)

Count

Wt. (g)

347

2.21

157

0.85

Domesticates

Squash

Zea mays ssp. mays Cucurbita sp.

Bottle gourd

Lagenaria siceraria

Beans (SNE)

Phaseolus vulgaris

3

Tobacco

Nicotiana sp.

2

Hickory

Carya spp.

99

1.04

212

2.05

Acorn

Quercus spp.

4

0.03

53

0.30

Walnut

Juglans nigra

5

0.24

Juglandaceae

walnut family

3

0.03

7

0.07

Hazelnut

Corylus sp.

3

0.02

Maize

1

Nuts

Castanea americana Fruits (number of seed fragments)

Poss. chestnut

Grape

Vitis sp.

14

9

Maypops

36

41

1

1

Sumac

Passiflora incarnata Diospyros virginiana Rhus sp.

Bramble

Rubus sp.

1

Persimmon

1

Other seeds Total seeds

277

423

Grass seeds

67

333

Other Wood

204.91

77.64

Bark

73.69

24.39

Cane stem

8

0.05

306

2.45

Feature 92

Central pits

Western pits

Feature 112

2

6

26

10

20

55

260

100

Count

Wt. (g)

Count

Wt. (g)

Count

Wt. (g)

Count

Wt. (g)

69

0.33

1,122

6.94

1,337

5.80

116

0.56

3

0.03

4

0.03

1

0.01

1

4

7

40

0.44

401

7.31

707

9.82

1,646

27.75

24

0.08

626

1.61

229

0.81

520

1.54

1

0.03

3

0.07

1

0.02

118

6.65

2

0.05

13

0.17

93

1.25

1

0.03

1

0.01

3

0.08

12

0.06

17

0.02

2

0.02

1

0.02

1

36

16

4

5

32

82

22

1

55

7

14

5

1

11

1 31

76

126

87

15

4

9

3

10

11.01

39.97

69.67

92.91

0.51

1.04

1.10

0.83

0.04

0.54

113

0.60

3

0.03

246 · Gayle J. Fritz

Figure 6.2. Density values (number of fragments divided by liters of soil floated) of maize and nutshell.

buildings to be built for occupation by Pardo’s men, it might indicate that soldiers viewed acorns with suspicion or distaste until after they had grown accustomed or resigned to native dishes. However, the higher visibility of acorn in the central pits, also interpreted as temporally early manifestations, and the relatively low frequencies and density of acorn in the western pits, thought to be later depositions, complicate attempts at making sense

People, Plants, and Early Frontier Food · 247

of the variability observed. Another scenario might involve physical separation of acorn-processing activity areas from Structure 1. Acorn shell can be cleanly removed from the edible cotyledons before further processing or cooking occur; therefore, its presence in so many samples from the pits west of Structure 5 and from Structure 5 itself—not to mention unusually high densities in the central courtyard pit features and in the Feature 112 “feasting pit”—leads me to suggest that it was regularly processed in this part of the site. Fragments of hickory nutshell were recovered from all samples examined except one sample from Structure 1 and one from Structure 5 (see figure 6.1). The presence of at least one hickory shell fragment even in most Structure 1 samples, despite low densities of edible plant types overall in Structure 1 (see figure 6.2), indicates that dishes including processed hickory nuts were probably routinely consumed there. Unlike acorn shell and walnut shell, which do not necessarily enter the foods made from their nutmeats, mass processing of hickory nuts and conversion to oil, milk, or kunuchee-type balls (Fritz et al. 2001; Scarry 2003) makes it likely that even if hickories were processed elsewhere, quite a few small shell fragments could be discarded inside the residential zone. The quantities and relatively large sizes of hickory nutshell pieces from Feature 112 are especially likely to represent processing there. The 100 liters of sediment floated from this feature yielded 1,646 fragments of hickory nutshell weighing a total of 27.75 grams, with a corresponding density of 16.46 fragments per liter or, if calculated by weight, 0.28 grams per liter. This is nearly three times the total weight of 9.85 grams of hickory shell from a much larger volume of sediment (265 liters) floated from the mound area as reported by Gremillion (2002). Nutshell densities in Feature 112 are far greater than those in the other pits west of Structure 5 or in either of the structures, although densities in the central pits were also high. Radford and colleagues (1968) show three species of hickory (Carya cordiformis, C. pallida, and C. tomentosa) growing in Burke County, North Carolina, and three others (C. glabra, C. ovata, and C. ovalis) in nearby counties. I do not divide hickory nutshell into thick-shelled and thinshelled groups, because neither pecan (C. illinoinensis) nor water hickory (C. aquatica), both with thin shells, is native to the western Piedmont or southern Appalachian region. Some of the hickory nutshell in these samples, however, especially in Feature 112, is quite thin (