The Archaeology of the Caddo [1 ed.] 9780803240469, 9780803220966

Citation preview

Copyright © 2012. Nebraska Paperback. All rights reserved. The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved. The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

The Archaeology of the Caddo

Copyright © 2012. Nebraska Paperback. All rights reserved.

Edited by Timothy K. Perttula and Chester P. Walker

University of Nebraska Press Lincoln and London

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

© 2012 by the Board of Regents of the University of Nebraska All rights reserved Manufactured in the United States of America Chapter 8, “The Evolution of a Caddo Community in Northeastern Texas,” by Timothy K. Perttula and Robert Rogers, was originally published in American Antiquity 72.1 (2007): 71–94. It has been slightly modified to conform with the editorial guidelines followed for the volume.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Library of Congress Cataloging-in-Publication Data The archaeology of the Caddo / edited by Timothy K. Perttula and Chester P. Walker. p. cm. Includes bibliographical references and index. isbn 978-0-8032-2096-6 (pbk.: alk. paper) 1. Caddo Indians—History. 2. Caddo Indians— Antiquities. 3. Caddo Indians—Social life and customs. 4. Excavations (Archaeology— Great Plains). 5. Great Plains—Antiquities. I. Perttula, Timothy K. II. Walker, Chester P. e99.c12a73 2012 976'.01—dc23 2011052083

Set in Quadraat and Quadraat Sans. Designed by A. Shahan.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Contents List of Figures List of Tables Foreword 1. The Archaeology of the Caddo in Southwest Arkansas, Northwest Louisiana, Eastern Oklahoma, and East Texas: An Introduction to the Volume timo thy k . perttula 2. Form and Structure in Prehistoric Caddo Pottery Design ann m. early

Copyright © 2012. Nebraska Paperback. All rights reserved.

3. At the House of the Priest: Faunal Remains from the Crenshaw Site (3MI6), Southwest Arkansas h. edwin jackson, susan l . sco tt, and frank f. schambach

vii xiii xv

1 26

47

4. Bioarchaeological Evidence of Subsistence Strategies among the East Texas Caddo diane wilson

86

5. Spiro Reconsidered: Sacred Economy at the Western Frontier of the Eastern Woodlands james a . brown

117

6. Viewshed Characteristics of Caddo Mounds in the Arkansas Basin gregory vogel

139

7. Exploring Prehistoric Caddo Communities through Archaeogeophysics chester p. walker and duncan p. mckinnon

177

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

8. The Evolution of a Caddo Community in Northeast Texas timo thy k . perttula and robert rogers 9. Settlement Patterns and Variation in Caddo Pottery Decoration: A Case Study of the Willow Chute Bayou Locality jeffrey s. girard

239

10. Caddo in the Saline River Valley of Arkansas: The Borderlands Project and the Hughes Site mary beth trubitt

288

11. Spatial Patterns of Caddo Mound Sites in the West Gulf Coastal Plain of Arkansas jami j. lockhart

313

12. Decisions in Landscape Setting Selection of the Prehistoric Caddo of Southeastern Oklahoma: A gis Analysis robert l . brooks 13. The Character of Fifteenth- to Seventeenth-Century Caddo Communities in the Big Cypress Creek Basin of Northeast Texas timo thy k . perttula

Copyright © 2012. Nebraska Paperback. All rights reserved.

209

14. The Belcher Phase: Sixteenth- and Seventeenth-Century Caddo Occupation of the Red River Valley in Northwest Louisiana and Southwest Arkansas david b. kelley 15. The Terán Map and Caddo Cosmology george sabo iii References Cited Contributors Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

335

363

411 431

449 499 501

Copyright © 2012. Nebraska Paperback. All rights reserved.

Figures

1-1. Distribution of the main Mississippian period groups in eastern North America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1-2. Sites and areas mentioned in the text in the Southern and Northern Caddo areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1-3. The geographic extent within the Southern and Northern Caddo areas discussed by contributors in this book . . . . . . . . 15 1-4. Looking south-southwest at the Battle mound site in southwest Arkansas in 2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 2-1. Map of Arkansas showing the area where the sample of Friendship Engraved, var. Freeman vessels used in this study came from, and the 1938–1950 collecting area of the Henderson State University Museum and Hodges pottery collections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 2-2. Friendship Engraved, var. Freeman carinated bowl, showing scribing lines laying out the design fields . . . . . . . . . . . . . . . . . 33 2-3. The first three steps in decoration . . . . . . . . . . . . . . . . . . . . . . . 34 2-4. The design choices found in the study sample . . . . . . . . . . . . . 35 2-5. Finishing stages and options . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 2-6. Final decorative choices available for the oval-within-anoval design option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 2-7. Examples of “nongrammatical” vessels . . . . . . . . . . . . . . . . . . 41 2-8. Friendship Engraved, var. Meador bowl . . . . . . . . . . . . . . . . . . . 42 2-9. Examples of stacked or fused vessels . . . . . . . . . . . . . . . . . . . . 45 3-1. Location of the Crenshaw site . . . . . . . . . . . . . . . . . . . . . . . . . . 51 3-2. The Crenshaw site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 3-3. Simplified plan view of the 1969 and 1983 Arkansas Archeological Survey Excavations at Crenshaw . . . . . . . . . . . . 53 3-4. Size distribution of fish remains from Crenshaw, distributed according to family . . . . . . . . . . . . . . . . . . . . . . . . . 70 3-5. Deer anatomical distribution in structure and midden fan samples, grouped by food utility . . . . . . . . . . . . . . . . . . . . . . . . 75

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

3-6. Comparison of percent nisp of faunal assemblages from Crenshaw and four Caddo sites in the Red River region . . . . . 77 3-7. Comparison of mni of faunal assemblages from Crenshaw and four Caddo sites in the Red River region . . . . . . . . . . . . . . 78 3-8. Comparison of relative bone weight contributions by family to the Crenshaw and McLelland faunal assemblages . . . . . . . . 79 4-1. The relationship between apatite delta 13C and delta 15N collagen values through time . . . . . . . . . . . . . . . . . . . . . . . . . . 104 4-2. The relationship between apatite delta 13C and delta 15N collagen values through time . . . . . . . . . . . . . . . . . . . . . . . . . . 105 5-1. The Craig mound from the northwest during the winter of 1913–1914 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 5-2. The Great Mortuary floor showing the approximate location of the major cult figures and other furniture recovered by the relic hunters among the litter burial constructions and other features uncovered by the wpa . . . . . . . . . . . . . . . . . . . . . . . . 129 6-1. Relief map showing locations of sites employed in this study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 6-2. Viewshed size from the ground surface . . . . . . . . . . . . . . . . . 153 6-3. Viewshed size from mound summits . . . . . . . . . . . . . . . . . . . 154 6-4. Gain in mound viewsheds . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 6-5. Mound viewsheds plotted by echelon . . . . . . . . . . . . . . . . . . . 157 6-6. Mound viewsheds ranked against the statistical background of 99 generated samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 6-7. Relief map showing viewshed from the summit of Brown mound at Spiro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 6-8. Relief map showing viewshed from the Craig mound at Spiro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 6-9. Relief map showing viewshed from the Cavanaugh mound . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 6-10. Topographic cross-section across the Poteau/Arkansas bottoms, with mound base and summit elevations . . . . . . . . 168 6-11. Relief map showing viewsheds from Mound I–II at Norman, Mound Unit 7 at Harlan, and Fort Davis . . . . . . . . . . . . . . . . . 170 6-12. View from the bottoms approximately 2 km north of Ewing Chapel Cemetery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 7-1. The magnetometer survey areas at the George C. Davis site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

7-2. Interpretive map showing the architectural features identified in the magnetometer surveys and past archeological excavations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 7-3. Examples of button houses at the George C. Davis site . . . . . 185 7-4. The excavated and geophysical architectural features from the George C. Davis site, including the structures excavated under Mound A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 7-5. Locations of possible plazas and community spaces at the George C. Davis site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 7-6. The location of Area A and Area B at the Hill Farm site, channel lakes of the Red River, and an ancient abandoned stream channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190 7-7. Plan of Area A collection blocks and interpretation of the geophysical data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 7-8. Plan of Area B collection blocks and interpretation of the geophysical data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 7-9. Detail of the western portion of the 1691 Terán map showing the old abandoned channel of the Red River and the household compounds that may represent the likely area of the Hill Farm site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 7-10. Total coverage of the magnetic gradiometry survey at the Battle site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 7-11. Clusters of possible structures, a community cemetery, a potential compound fence, a borrow pit, and a possible linear causeway identified in the Battle site archaeogeophysical data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 7-12. House 6 at the Belcher site compared with a small circular structure with a possible extended entranceway identified in the Battle site data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 7-13. A complex grouping of high magnetic values forms a possible farmstead with a compound fence . . . . . . . . . . . . . . 200 7-14. Numerous circular structures located on two rises east of the large mound . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 7-15. Structure 1 at the McLelland site compared with two large circular structures identified in the Battle mound site data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204 7-16. Structure at the Werner site compared with the several concentric circular patterns identified in the Battle mound site data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

7-17. A large cluster of pits possibly representing a large comcommunity cemetery about 300 m east of the large mound at the Battle site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 8-1. Oak Hill Village in the Caddo area of northeast Texas and other important contemporaneous Caddo sites and phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 8-2. Plan of structures and features at the Oak Hill Village . . . . . . 213 8-3. Structure groups A–H and structures 2, 5, and 10 at the Oak Hill Village . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214 8-4. Plans of the different kinds of structures at Oak Hill Village . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 8-5. Three overlapping circular structures in the northeastern part of the site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 8-6. Circular structures in the western part of the site . . . . . . . . . . 222 8-7. Overlapping structures in the northwestern part of the Oak Hill Village site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 8-8. Unraveling the sequence of structures in the northwestern part of the Oak Hill site, from A (oldest) to E (youngest) . . . . 224 8-9. Inferred distribution of the earliest village at the Oak Hill Village site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 8-10. Inferred distribution of structures in the middle-era village . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 8-11. Inferred distribution of the latest village at the Oak Hill Village site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 8-12. Middle Caddo period ceramic complexes in northeast Texas, including a complex on the Sabine River and tributaries where the Oak Hill Village site is located . . . . . . . . . . . . . . . . 235 8-13. Changes in atmospheric delta 14C and climatic minima (peaks in atmospheric delta 14C) during the three village components at the Oak Hill Village site . . . . . . . . . . . . . . . . . 237 9-1. Location of the Willow Chute Bayou locality . . . . . . . . . . . . . . 240 9-2. Terán map of 1691–1692 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 9-3. Land surfaces in the Willow Chute Bayou locality . . . . . . . . . . 245 9-4. Distribution of recorded archaeological sites along Willow Chute Bayou . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248 9-5. Contour map of the Vanceville mound site . . . . . . . . . . . . . . . 260 9-6. Profile of test pit 96B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264 9-7. Profile of test pit 96C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268 9-8. Percentage stratigraphy for contexts at the Vanceville site . . . 270

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

9-9. Scatter plot of first two dimensions for the correspondence analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278 9-10. Frequency seriation of collections used in the correspondence analysis with order following dim1 scores . . . . . . . . . 280 9-11. Schematic view of the Willow Chute locality with sites coded according to dim1 scores in the correspondence analysis . . . 282 10-1. Portion of Arkansas and surrounding states with physiographic regions and locations of sites discussed in text . . . . 290 10-2. Photograph of main mound at the Hughes site, taken in 1972 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292 10-3. Partial vessel, Mound Place Incised and Brushed type, from 1972 testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293 10-4. Topographic plan map of the Hughes site . . . . . . . . . . . . . . . 294 10-5. Excavation unit N199E244, feature 16 . . . . . . . . . . . . . . . . . . . 296 10-6. Probability curves of 2-sigma calibrated age ranges for radiocarbon assays from the Hughes site compared with other area sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298 10-7. Common ceramic patterns found in the Hughes site sherd assemblage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302 10-8. Class C rim patterns, Class B body patterns . . . . . . . . . . . . . . 304 10-9. Class E sherds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 10-10. Arrow points from the Hughes site . . . . . . . . . . . . . . . . . . . . . 309 11-1. Statistical clusters of mound sites compared with similar clusters grouped by watershed boundaries . . . . . . . . . . . . . . . 321 11-2. Environmental similarity model for Caddo mound sites in the West Gulf Coastal Plain of Arkansas . . . . . . . . . . . . . . . . . 328 11-3. Habitation, activity, and accessibility spaces of prehistoric Caddo culture in the West Gulf Coastal Plain of Arkansas based on Caddo mound site density and environmental similarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331 12-1. Study area of Choctaw and McCurtain Counties and streams and rivers of southeastern Oklahoma . . . . . . . . . . . . . . . . . . . 338 12-2. Distribution of Caddo settlements in Choctaw and McCurtain Counties, southeastern Oklahoma . . . . . . . . . . . . . . . . . . 348 12-3. Elevation of Caddo settlements by built environment . . . . . . 356 13-1. The distribution of the Titus phase . . . . . . . . . . . . . . . . . . . . . 364 13-2. Titus phase Caddo political communities in the Big Cypress Creek basin of northeast Texas . . . . . . . . . . . . . . . . . . . . . . . . 368

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

13-3. The important Titus phase village and political community center at the Shelby site on Greasy Creek . . . . . . . . . . . . . . . . 370 13-4. Mound C at the Harroun site . . . . . . . . . . . . . . . . . . . . . . . . . . 384 13-5. Number of recorded burials from Titus phase cemetery sites on different drainages in northeast Texas . . . . . . . . . . . . 394 13-6. Variation in the size of Titus phase cemeteries . . . . . . . . . . . . 395 13-7. The structure of Titus phase community cemeteries . . . . . . . 398 13-8. Distribution of Titus phase cemeteries with burials of individuals of presumed high social rank . . . . . . . . . . . . . . . . 406 14-1. The location of Belcher phase sites mentioned in the text . . . 413 14-2. Belcher phase structure patterns . . . . . . . . . . . . . . . . . . . . . . . 416 14-3. Belcher phase pottery from the Belcher site . . . . . . . . . . . . . . 418 14-4. Marine shell ornaments from the Belcher site . . . . . . . . . . . . 426 15-1. Facsimile of the Terán map, 1691 . . . . . . . . . . . . . . . . . . . . . . 434

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Tables

1-1. Caddo chronological framework . . . . . . . . . . . . . . . . . . . . . . . . 13 3-1. Counts and weights of identified taxa from Caddo I contexts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 3-2. Charring of Early Caddo deer postcranial elements by body part . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 3-3. nisp, nne, and mni for whitetail deer from all analyzed Early Caddo deposits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 3-4. Percent nisp for taxa from five Caddo sites . . . . . . . . . . . . . . . 76 3-5. Unusual taxa recovered from elite contexts at Crenshaw, Lubbub Creek, and Toqua sites . . . . . . . . . . . . . . . . . . . . . . . . . 82 4-1. Average dental attrition scores for sites in northeast Texas . . . 90 4-2. Caries data for sites discussed in the Blackland Prairie, Post Oak Savannah, and Pineywoods of northeast Texas . . . . . 94 4-3. Stable isotope data from the study region . . . . . . . . . . . . . . . . . 99 4-4. Mean isotopic values through time alongside caries rates and dental wear scores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 4-5. Stable isotope means in the Blackland Prairie, Post Oak Savannah, and Pineywoods . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 4-6. Stable isotope means in the east Texas river basins . . . . . . . . 110 4-7. Stable isotope means for Caddo mound and non-mound sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .111 6-1. Recorded mounds in the Arkansas Basin, ordered by state site number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 6-2. All mound viewsheds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 6-3. Parameters used to generate the predictive model of mound locations within the Arkansas and Neosho River regions . . . 158 6-4. Parameters used to generate the predictive model of mound locations within the Ozark Plateau and Ouachita Mountain regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 6-5. Results of Monte Carlo analysis . . . . . . . . . . . . . . . . . . . . . . . . 160 8-1. Details of the circular structures at the Oak Hill Village site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-2. Radiocarbon dates from the Oak Hill Village site . . . . . . . . . . 219 9-1. Summary of radiocarbon assays from the Vanceville mound site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266 9-2. Sherd percentages for contexts at the Vanceville mound site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 9-3. Counts and percentages of sherd categories for collections in the Willow Chute locality . . . . . . . . . . . . . . . . . . . . . . . . . . . 274 9-4. Summary of correspondence analysis . . . . . . . . . . . . . . . . . . . 277 10-1. Results of radiocarbon dating of Hughes site samples . . . . . . 297 10-2. Ceramic sherd surface decoration by temper . . . . . . . . . . . . . 300 11-1. Geographic Information System data layers . . . . . . . . . . . . . . 317 11-2. Caddo mound coincidence with Arkansas physiography using χ2 statistic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 11-3. Caddo mounds and land use/land cover in the West Gulf Coastal Plain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325 11-4. Summary of environmental variables associated with most Caddo mounds in the West Gulf Coastal Plain . . . . . . . . . . . . 325 11-5. Statistics for the seven-class environmental similarity model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 12-1. Distribution of Caddo sites by soil association . . . . . . . . . . . . 353 12-2. Soil associations and the constructed landscape. . . . . . . . . . . 354 12-3. Distribution of the Caddo constructed landscape by landform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355 12-4. Distribution of Caddo sites by slope . . . . . . . . . . . . . . . . . . . . 357 12-5. Distribution of Caddo constructed landscape by slope direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358 12-6. Distribution of Caddo constructed landscape by nearest water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359 13-1. Titus phase radiocarbon dates . . . . . . . . . . . . . . . . . . . . . . . . . 375 13-2. Titus phase radiocarbon dates from earthen mounds in the Pineywoods and Post Oak Savannah of northeast Texas . . . . . 378 13-3. Titus phase cemeteries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389 13-4. Frequencies of funerary objects in selected Titus phase cemeteries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402 13-5. Notable burials from Titus phase cemeteries and kind of mortuary treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404 14-1. Frequency of ceramic types in four Belcher phase assemblages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Foreword The archaeology of the Caddo Indian peoples that lived in the forested habitats of southwest Arkansas, northwest Louisiana, eastern Oklahoma, and eastern Texas, in the far western reaches of the southeastern United States, has been the topic of archaeological inquiry since the early twentieth century. The study of Caddo archaeology over the years has been important in inspiring questions and interpretations of the native history of the Caddo peoples beginning some 2,500 years ago, as well as their relationships with Woodland and Mississippian cultures in the lower Mississippi valley and the eastern United States, as well as possible relationships with northern Mexican and Mesoamerican cultures. These interpretations have flowed from the early study of such impressive Caddo mound centers at the George C. Davis site (in east Texas), the Gahagan, Mounds Plantation, and Belcher sites (in northwest Louisiana), the Crenshaw and Battle sites (southwest Arkansas), and the Harlan and Spiro sites (in eastern Oklahoma), to the more wide-ranging and eclectic archaeological and geophysical studies of many sites, features, and material culture assemblages that are the foundation of Caddo archaeology today. Nevertheless, while the character of Caddo archaeology, and the histories of generations of Caddo peoples embodied in preserved archaeological sites, is better known today than ever before, it is fair to say that the appreciation of the diversity of cultural practices and traditions that came to characterize the pre-Columbian Caddo world is only now being realized. As new advances come to light on a range of new and old research themes and issues and on previously recorded and new sites, it is an exciting time to be involved in the study of Caddo archaeology This book presents new advances in the native history of the Caddo Indian peoples, focusing on key sites and several research approaches and themes, among them better ways to understand mortuary practices, ceramic analysis, reconstruction of settlement and regional histories of different Caddo communities, Geographic Information Systems and xv

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

geophysical landscape studies at several spatial scales, and the cosmological significance of mound and structure placements. Our purpose in compiling this book is to bring the unique and compelling story of the Caddo Indians to a broad audience, including those interested in Native American life who may not know of the Caddo peoples and their proud heritage. The native history of the Caddo has all too often been portrayed as little more than that of a peripheral Mississippian culture, and we wish to remedy this by placing the focus of the book squarely on the archaeology of the Caddo’s ancestral world, one that can be defined beginning around A.D. 800–900 over an area perhaps covering as much as 80,000 square miles of four states. In ancestral times, the Caddo Indians were mound builders, expert traders and artisans, and accomplished farmers, as well as the most socially complex Native Americans living between the Mississippi River Mississippian societies and the ancestral Puebloan peoples of the American Southwest. When Europeans came among the Caddo in the late seventeenth century, they relied on the goodwill of the Caddo to explore what became Arkansas, Louisiana, Oklahoma, and Texas, as well as the diplomatic and economic skills of the people. The Caddo’s rewards were disease, depredations, and territorial dispossession at the hands of French, Spanish, English, and American speculators, mercenaries, priests, traders, and land developers. By 1835, the Caddo’s fate in the land that had once been theirs became clear, and the policies of the Republic of Texas and the United States between them led to their forced exodus from Texas to Indian Territory in 1859. The Caddo live to this day in their new Oklahoma home. We take an archaeological perspective in relating the native history of the Caddo Indian peoples because this is the best way to convey, as well as to experience, the long sweep and dense chronicle of a politically and religiously astute Native American group. This archaeological chronicle of the Caddo Indian people hopes to bring to light their heritage, their creativity, and their political and religious abilities. We look forward to the continued archaeological study of the sites left behind and abandoned by the Caddo Indian peoples in their traditional homelands and to refining our understanding of how these peoples lived and thrived in this part of the southeastern United States.

xvi

Foreword

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved. The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

1. The Archaeology of the Caddo in Southwest Arkansas, Northwest Louisiana, Eastern Oklahoma, and East Texas An Introduction to the Volume

Copyright © 2012. Nebraska Paperback. All rights reserved.

timo thy k . perttula

This volume examines the archaeology of the Caddo Indians who lived in southwest Arkansas, northwest Louisiana, eastern Oklahoma, and east Texas from at least as early as 2,500 years ago, if not before then, ca. 3,300 years ago based on glottochronological dates and maize term evaluations (Brown 2006:table 47-4), until they were removed to Oklahoma in 1859.1 In the broadest terms, and seen as evolving from their Woodland period forefathers (see Early 2004; Schambach 2001, 2002; Story 1990; Wyckoff 1980), the Caddo archaeological and cultural tradition represents “an archaeological concept . . . recognizable primarily on the basis of a set of long-standing and distinctive cultural, social, and political elements that have temporal, spatial, and geographic connotations” (Perttula 1992:7). What these elements are, and how they are represented in the archaeological record of the Caddo area after ca. A.D. 800, has no simple answer, primarily because the archaeology of the Caddo by region is quite variable in material culture expressions (especially in its stylistically diverse fine ware engraved ceramic vessels), social and political practices, use of landscapes, subsistence strategies and use of cultivated plants, interaction with neighbors (both Caddo and non-Caddo, especially with Mississippians to the east), and the tempo of cultural changes. The chapters in this book take up the challenge of examining Caddo archaeology through thematic, regional, and microhistorical perspectives. It has been common practice by archaeologists since at least the 1940s (e.g., Krieger 1947:199) to refer to the archaeology of this broad 1

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

area as “Caddoan” or as the “Caddoan area,” even as it was recognized that these terms are more than problematic, primarily due to the reluctance to link a linguistic label (i.e., “Caddoan”) with the archaeological record of indigenous peoples who lived in a specific geographic area. “Caddoan” is a linguistic term that has been in use since John Wesley Powell’s 1891 pioneering language studies, and it refers to a language family that includes two branches: the Northern Caddoan, including the Pawnee, Arikara, Kitsai, and Wichita languages, and the Southern Caddoan, including the Caddo language (Goddard 1996:319). The Adai language, of the Adaes group that lived in northwest Louisiana in historic times (Gregory 1983; Gregory et al. 2004), has an uncertain but unlikely relationship to the Caddo language (Wallace Chafe, September 2010 personal communication). It is the archaeology of the Caddo peoples that speak the Southern Caddoan language (and its many dialects) that is the primary focus of this book. We will employ the term “Caddo” to refer to the peoples who lived in the area in question, rather than the hackneyed term “Caddoan,” and also use the term “Caddo” to refer to the many archaeological sites and abundant material they left behind. The use of the term “Caddo” conveys the belief that the peoples who lived in east Texas, northwest Louisiana, southwest Arkansas, and eastern Oklahoma, centered on the Red River and its tributaries, shared a common cultural heritage and native history. The annual Caddo Conference is the principal venue is which Caddo people and archaeologists who study their native history and culture “remember that the contributions of Caddo people continue to teach us how to better understand what we mutually share” (Gregory 2009:2): a deep and abiding interest in the native history and cultural traditions of the Caddo peoples. The term “native history” (cf. Trigger 1980, 1985) as used in this book refers to the totality of the archaeological history of the Caddo peoples from at least 2,500 years ago to the present in the geographic area that has come to be recognized as the traditional territory of the Caddo people. In other words, we view archaeology as long-term history (Mitchell and Scheiber 2010:18–19). The term “prehistoric” is also employed by several authors in this book. By it, they are referring to the period in the archaeological record before A.D. 1680 and the first sustained contact with Europeans. By using that term, they do not mean to imply that the Caddo peoples did not have their own history, they do not intend 2

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

1-1. Distribution of the main Mississippian period groups in eastern North America (after Fagan 1995:437).

to “frame native culture change in terms of European processes and European experiences,” or that “they embody the belief that the forces unleashed by the arrival of Europeans were both novel and irresistible” (Mitchell and Scheiber 2010:13). The term “pre-Columbian” is found in George Sabo’s chapter 15, employed as a counterpoint to the term “prehistoric.” This word, however, still has strong connotations that history in the New World began with Columbus in 1492, “the year the world began” (Fernandez-Armesto 2009:315), and thus is not widely used herein. To many archaeologists who work in the eastern and southeastern United States and the Great Plains today, the archaeology of the Caddo is poorly understood. Maps of the Caddo area in relationship to the Mississippian Southeast are often inaccurate or incomplete with respect to the locations of important Caddo sites as well as the territorial “boundaries” or limits of the Caddo area (i.e., Cobb 2003:fig. 1; Cobb and Giles 2009:fig. 3-1; Milner et al. 2001:figs. 2-1–2-3; Payne and Scarry Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

3

Copyright © 2012. Nebraska Paperback. All rights reserved.

1998:fig. 2-1; Scarry 1999:figs. 5-3–5-5; White and Weinstein 2008:fig. 10). The Caddo Indians are rarely mentioned in more accessible archaeological publications. This may be due to the absence of an overarching and modern synthesis of their native history—or at least a synthesis that has been written in the last 35–60 years (e.g., Newell and Krieger 1949; Webb 1959; Wyckoff 1974). Also, much of the present-day consideration of Caddo archaeology and native history is confined to the “gray” (i.e., limited distribution) archaeological literature and the technical writings of archaeologists, ethnographers, linguists, and physical anthropologists.2 Instead, much of what is thought to be understood about the Caddo peoples from an archaeological perspective is primarily a product of what seems to be a misplaced focus and emphasis solely on the spectacular mortuary findings and Southeastern Ceremonial Complex (e.g., King 2007; Reilly and Garber 2007) artifacts from the Spiro site on the Arkansas River in eastern Oklahoma (e.g., Brown 1996, 2007; see also Brown, chap. 5, this volume) as representative of the Caddo archaeological record, rather than a proper focus on communicating to others the considerable diversity that underlies and characterizes the Caddo archaeological tradition (Story 1990:320). Still others see Caddo archaeology as a western manifestation of the Mississippian world (see fig. 1-1), perhaps even peripheral to the core Mississippian groups of the central Mississippi valley, but contemporaneous with Plaquemine (see Rees and Livingood 2007), Middle Mississippian (see Butler and Welch 2006; Pauketat 2004, 2007), and South Appalachian Mississippian aboriginal groups. Blitz (2010:7) describes the Caddo as being situated “at the far edges of the Mississippian world.” Caddo archaeologists, in the main, however, have argued that the evolutionary development of the Caddo cultural tradition in the ninth century, from an indigenous Woodland world, took place independently of the emergence of Mississippian period developments in the southeastern United States, although they consider Caddo society to be southeastern in character because of related political histories, similar platform and burial mound constructions, the social hierarchy of populations (including priests [Xinesi] and chiefs [Caddi] as the elite), and an eventual reliance on cultivated plants. Rogers (1991:224) called it a “Caddoanization” of local groups within this broad region along the far western edge of the Southeast. Helen Tanner (1993:6) has perceptively noted 4

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

that “the Caddo became the most western of the great chiefdoms that developed in what is now the southeastern United States. . . . Of all the great southeastern chiefdoms, the Caddo was the only one west of the Mississippi River. No other comparable social and political organization existed between the Caddo country and the land of the Anasazi in New Mexico.” Consequently patterns of cultural change that have been identified among cultural groups within the Caddo archaeological area are not thought to be directly comparable with, nor temporally synchronous with, those of Mississippian groups (e.g., Blitz 2010; Livingood 2008; Pauketat 2005). There are clearly discernible sociopolitical and trade relationships between the Caddo and a number of aboriginal groups living in the Southeast in prehistoric and early historic times, and Caddo societies share much with their Mississippian neighbors—including the eventual adoption of maize and the intensification of maize agricultural economies, as well as in systems of social authority and ceremony (Butler and Welch 2006; Livingood 2008:3–11). But the Caddo area is manifestly different in several intriguing ways. We find a 1,000-year record of broad cultural change and social complexity that is manifest in cultural, ethnic, and presumed genetic continuity among peoples who maintained their own distinctive sociopolitical and economic dynamic and had no need for fortified settlements (Dye 2008:12). They continued to build mounds into the seventeenth century and had well-populated settlement areas, whereas Mississippian polities fought and competed for power and tribute (cf. Cobb and Giles 2009; Dye 2006, 2008; Dye and King 2007; Pauketat 2004; Brown 2006; King 2006) and abandoned many parts of the Southeast shortly after 1450. This book’s purpose is to present the most current findings, from a number of approaches and regions, concerning what is known about the diverse Caddo archaeological record. The chapters range from the Arkansas River basin, to the Pineywoods of east Texas, and to the traditional and centrally placed Caddo heartland along the Red River and its major tributaries. The contributors to this volume have devoted much of their careers to the study of the Caddo Indian peoples. Historical, Ethnographic, and Archaeological Context Before the middle of the nineteenth century the term “Caddo” denoted only one of at least 25 distinct but closely affiliated groups centered around the Red River in Texas, Arkansas, Louisiana, and Oklahoma (see Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

5

Copyright © 2012. Nebraska Paperback. All rights reserved.

1-2. Sites and areas mentioned in the text in the Southern and Northern Caddo areas.

Swanton 1942). The term derives from the French abbreviation of “Kadohadacho,” a word meaning “real chief ” or “real Caddo” in the Kadohadacho dialect. European chroniclers referred to the Caddo groups as the Hasinai, Kadohadacho, and Natchitoches confederacies, although the “confederacies” are better interpreted in modern parlance as kin-based affiliated groups or bands of Caddo communities. The many Hasinai groups lived in the Neches and Angelina River valleys in east Texas, the 6

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Kadohadacho groups on the Red River in the Great Bend area, and the Natchitoches groups on the Red River in the vicinity of the French post of Natchitoches (Fort St. Jean Baptiste aux Natchitos), established in 1714 (see Swanton 1942:fig. 1). The first European description of the Caddo peoples came in 1542 from diarists traveling with the de Soto entrada, then led by Luis de Moscoso Alvarado (Hernando de Soto had died in the spring of 1542). The Spanish described several of the Caddo groups as having dense populations living in scattered settlements and having abundant corn reserves (Hudson 1997; Young and Hoffman 1993). Recent investigations of many pre-1680 Caddo sites indicate that Caddo communities were widely dispersed throughout the major and minor stream valleys of the Caddoan area by around 800. Their roots can be traced to several ancestral Woodland Period culture groups of varying sociopolitical complexities, including the Fourche Maline (Schambach 2002), Mill Creek (Perttula and Nelson 2004a), and Mossy Grove (Story 1990) cultures, which began to settle down in dispersed communities throughout the region, to manufacture ceramics for cooking and storage of foodstuffs and to develop a horticultural way of life based on the raising of tropical cultigens (corn, squash, and beans) and certain native plants (Early 2004:560–566). By the early ninth century, however, the Caddo had begun to develop a series of fundamentally complex and sedentary societies, fueled by distinctive but still poorly known ideological and cosmological perspectives, creating an extensive archaeological record of their life that spanned at least nine centuries, or until they were forcibly removed from Louisiana and Texas, beginning in the 1830s (see Neighbours 1973, 1975). That record is marked by the remains of farmsteads, hamlets, villages, family and community cemeteries, many small and large mound centers with public structures on and off mound platforms, plazas, and the burials of the social and political elite in and off mounds, as well as a rich material culture, especially their well-crafted ceramic wares. The development of Caddo culture in ancestral times may have been the result of several factors, including (a) the rise, elaboration, and maintenance of complex social and political symbols of authority, ritual, and ceremony (centering on the construction, dismantling, remodeling, and use of earthen temple and burial mounds) and ideology; (b) the development of elite status positions within certain Caddo communities; (c) increased sedentary life and the establishment of both domestic and Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

7

Copyright © 2012. Nebraska Paperback. All rights reserved.

sacred “places . . . that lent order to the chaotic worlds of people and nature” (Pauketat 2007:198); and (d) the expanding reliance on tropical cultigens in the economy, with an intensification in maize production after the thirteenth century (see Wilson, chap. 4, this volume), accompanied by increases in population. Regardless of the processes involved, it is clear that after the tenth century, the Caddo groups in the main were complex and socially ranked societies with well-planned civic-ceremonial centers. These Caddo peoples conducted elaborate mortuary rituals and ceremonial practices at sacred places and engaged in extensive interregional trade and social interaction (Early 2004; Kidder 1998, 2004, 2007:203). Caddo societies shared much with their Mississippian and Plaquemine neighbors, particularly the eventual adoption of maize and the development of maize agricultural economies, as well as the aforementioned systems of social authority and ranking. In the times before European observations and written reports, the Caddos lived in dispersed communities of grass and cane-covered houses of various forms and construction methods. The communities were composed of isolated farmsteads, hamlets, a few larger villages with compounds of farmsteads, and the civic-ceremonial centers. These centers used earthen mounds as platforms for temple structures for civic and religious functions, for burials of the social and political elite, and for ceremonial fire mounds, all expressions of social complexity (Knight 2010:1). The largest communities and the most important civic-ceremonial centers were primarily located along the major streams: the Red, Arkansas, Little, Ouachita, and Sabine Rivers (see fig. 1-2). The Caddo peoples developed a successful horticultural economy based on the cultivation of maize, beans, and squash, as well as such native cultigens as maygrass, amaranth, chenopods, and sunflowers, but they also consumed a wide variety of wild plants and animals found in the wooded habitats of the region. By about 1300 most Caddo groups were consuming large amounts of maize (Perttula 2008; see Wilson, chap. 4, this volume). Several varieties of corn were cultivated, an early or “little corn,” harvested in July, and the “flour corn,” harvested in September at the harvest of the Great Corn. Deer was the most important source of meat to the Caddos, who also exploited bison (especially among the Arkansas basin Caddo, see Wyckoff 1980) and bear for their furs and meat. After 8

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

the introduction of the horse in the late seventeenth century, the Caddos began to participate in winter communal bison hunts on the prairies to the west of their settlements, and Caddo groups also became heavily invested in the deer hide trade with Europeans. They developed long-distance trade networks in prehistoric times. Important products were bison hides, salt, and bois d’arc bows, along with copper, stone, turquoise, and marine shell used for gorgets, cups, and dippers, as well as finished objects such as pottery vessels and large ceremonial bifaces made from high-quality cherts. Many of the more important trade items were obtained from great distances (e.g., turquoise from New Mexico, copper from the Great Lakes, marine shell from the Gulf Coast, and finished marine shell items and fired clay pipes from the Cahokia area). These items were often placed in the graves of the social and political elite. The Caddo peoples used clay, stone, bone, wood, shell, and other media to manufacture tools, clothing, ceramic vessels, basketry, ornaments, and other material items. The Caddos are particularly well known, and have been since the early archaeological investigations of C. B. Moore (1909, 1912), for the beautiful artistic and functional ceramic wares they made. Stone was fashioned into arrow points, and the Caddos also made ground stone celts and axes for use in girdling and removing trees and turning over the soil. They made bone into awls, beamers, digging implements, and hoes, as well as ornaments, beads, and whistles. Hoes and digging tools were also made of freshwater mussel shells, while marine shells obtained through trade were used in the production of shell pendants, gorgets, beads, and cups. The Caddos traced descent through the maternal line rather than the paternal. Matrilineality was reflected in kinship terms, as an individual’s father and father’s brothers were called by the same term as the mother and the mother’s sisters. The Caddos recognized and ranked a number of clans. Marriage typically occurred between members of different clans. Religious and political authority in historic Caddo society rested in a hierarchy of key positions within and between the various affiliated communities and groups. The xinesi inherited a position of spiritual leadership, the caddi the position of principal headman of a community (also a hereditary leadership position), and the canahas the position of subordinate headmen or village elders. The Caddo people turned to the xinesi for mediation and communication with the supreme Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

9

Copyright © 2012. Nebraska Paperback. All rights reserved.

god, the Caddi Ayo, for religious leadership and decision-making influence between allied villages and in leading certain special rites, including first-fruits, harvest, and naming ceremonies. The xinesi imbued everyday life with the supernatural. The caddi was primarily responsible for making political decisions for the community, sponsoring ceremonies, leading councils for war expeditions, and conducting the calumet (or peace pipe) ceremony with visitors to the communities. The most influential and politically astute Caddo leaders, or caddices, in historic times were Tinhiouen (ca. 1760–1789) and Dehahuit (ca. 1800–1833) of the Kadohadachos, and Iesh or José María (ca. 1842–62) of the Anadarko or Nadaco tribe. At the time that sustained Spanish and French contact began in the late seventeenth century (see Barr 2007:17–25), Caddo peoples lived mainly on the Red River and in east Texas, with the exception of the Ouachita and the Cahinnio groups on the Ouachita River (see Swanton 1942:fig. 1). European populations—living in missions, ranches, and trading posts—increased throughout the eighteenth and nineteenth centuries in the Red River valley and in the vicinity of Natchitoches and Nacogdoches, important fur trading centers, while epidemics between 1691 and 1816 greatly reduced Caddo populations (see Perttula 1992:70–76; Derrick and Wilson 2001; Barr 2007:59–60, 304n58). At the same time, the Caddo peoples participated in the fur trade, traded guns, horses, and other items to Europeans and other Indians, and developed new trade and economic networks. The resulting economic symbiosis between the Caddo groups and Europeans (cf. Gregory 1973) was an important means of interaction for the Caddo because great quantities of European goods became available to the Caddo. While the Hasinai Caddo groups continued to live through the 1830s in their traditional east Texas homeland in the Neches and Angelina River valleys, the Kadohadacho groups moved off the Red River in the 1790s to distance themselves from Osage depredations and slave-raiding. Their new settlements were established between the Sabine River and Caddo Lake, generally along the boundary between the territory of Louisiana and the Spanish province of Texas (Tiller 2008:fig. 1). Most of the Kadohadacho remained in the Caddo Lake area until about 1838, although with the cession of Caddo lands in Louisiana to the United States in 1835 and increased settlement by Anglo-Americans in the province of Texas, other Kadohadacho were forced to move to the Brazos River in what is now north central Texas to 10

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

avoid Anglo-America depredations. By the early 1840s, all Caddo groups had moved to the Brazos River area to remove themselves from white repressive measures and colonization efforts (see Neighbours 1975; La Vere 2004; Anderson 2005; Smith 2005). They remained there until they were placed on the Brazos Indian Reservation in 1855 by the U.S. government, and then, because of harassment from encroaching Texans, in 1859 the Caddos (about 1,050 people) were removed to the Washita River in Indian Territory (now west Oklahoma) with the help of Robert S. Neighbors, superintendent of Indian affairs in Texas. During the Civil War most of the Caddo groups abandoned the Indian Territory and resettled in south and east Kansas, but they moved back to the Wichita Reservation in 1867. By 1874 the boundaries of the Caddo reservation were defined, and the separate Caddo tribes agreed to unite. Under the terms of the General Allotment Act of 1887, the Caddo reservation was partitioned in 1902 into 160-acre allotments for each enrolled Caddo, and the remaining lands were opened for white settlement (see Smith 1996). The Caddo peoples continue to live in west Oklahoma, primarily in Caddo County near the Caddo Nation of Oklahoma Tribal Complex, outside Binger.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Chronological Framework The Caddo chronological framework has passed through a series of iterations since the pioneering typological and chronological schemes developed in the 1940s by Alex D. Krieger (1944, 1946; see also Suhm et al. 1954:151–227) and Clarence H. Webb (1945, 1959). That scheme was solidly grounded in the culture history practices of that era of North American archaeology (e.g., Lyman et al. 1997), and led to the definition of two Aspects (the chronologically earlier Gibson and the younger Fulton) and various foci, and the recognition of a range of components (i.e., the archaeological manifestation of a specific focus at particular sites), including mound centers, villages, and hamlets, as well as cemetery sites, based on the Midwestern Taxonomic System (mts) (McKern 1939). With increasing research in a variety of locations in the Caddo archaeological area, dissatisfaction with the Krieger and Webb classification scheme led to the creation of new Caddo temporal-spatial systematics. This was based on the introduction of the phase scheme (Willey and Phillips 1958; Phillips 1970) and a new type-variety system for ceramics Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

11

Copyright © 2012. Nebraska Paperback. All rights reserved.

(Schambach and Miller 1984) from the lower Mississippi valley into the Caddo area. These phases replaced the foci of the mts, sometimes (but not always) with substantive changes in definitions of archaeological traits characteristic of the phase (i.e., a formal content that almost always comprised a range of ceramic fine ware and utility ware types), their geographic space, and their estimated temporal duration. Components recognized at specific sites were the building blocks of phases, as they were with foci, and through the years the study of components within phases—and the comparison of the archaeological character of specific phases—have been employed by archaeologists when defining and unraveling the development of local and regional cultural-historical sequences of past Caddo groups. This concern with culture history, or the temporal breadth of Caddo native history as seen through the study of the preserved archaeological record, is still a major pursuit of Caddo archaeologists as new sites, localities, regions, and sub-areas continue to be investigated. This has proceeded along with the development of new technological and methodological approaches (cf. Lockhart 2007; Walker 2009) that are basically in step with archaeology as it is being practiced in North America (cf. Davis and Davis 2009:4). Theoretical developments with archaeological applications specific to the Caddo area have not had much currency. The broad Caddo chronological framework followed in this book is provided in table 1-1. It is deceptively simple, being divided into five periods of roughly equivalent temporal length, beginning with the Formative Caddo period at ca. A.D. 800–1000 (Story 1990) and ending with the Historic Caddo period. There have been few, if any, professional archaeological investigations on Caddo sites that date after ca. 1830 in the traditional Caddo homelands (but see Parsons et al. 2002; Marceaux and Perttula 2010), and none (that I am aware of ) on post-1859 Caddo sites in Indian territory. The strength of the Formative to Historic Caddo period scheme is that contemporaneous regional Caddo archaeological entities can be related one to another in a consistent manner. Within each of these periods are a series of content-based archaeological phases recognized at the scale of the region or locality, including the major regions: Arkansas basin, the western Ozark Highlands, the Great Bend, Little River, Northwest Louisiana, the Middle Red, the Ouachita River, the southern Ouachita 12

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Table 1-1. Caddo chronological framework Period

Dates (A.D.)

Formative Caddo Early Caddo Middle Caddo Late Caddo Historic Caddo

800–1000 1000–1200 1200–1400 1400–1680 1680–1860+

Mountains, Big Cypress Creek, the upper Sabine River, and the Neches River basin (see fig. 1-2; see also Perttula 1992:table 3). Unlike earlier period schemes (i.e., Caddo I–V, see Davis 1970), there is no unstated assumption in the chronological framework presented in table 1-1 that the Formative to Historic Caddo periods represent linear or evolutionary views of regional developments, or that archaeological developments within the Caddo area conform in any way from one region to another within the overall areal framework (Perttula 1992:58). It may well be the case, however, that these archaeological periods and their associated phases have outlived their usefulness in framing Caddo chronological and taxonomic problems and research questions. Dunnell (2008a:58–64), who has thoroughly belabored the archaeological value of phases in the lower Mississippi River valley (and elsewhere), granting them little currency in modern research efforts, has noted that “as radiocarbon dates become cheaper and the number of ways to obtain chronometric dates has increased, the phase should be out of work. It is now possible to assign virtually any assemblage to an independent time scale. The actual relations between data points can be studied instead of boxes of our own cryptic creation” (Dunnell 2008a:64). The day may come in the Caddo area when sufficiently robust sets of calibrated radiocarbon dates, thermoluminescence dates (see Feathers 2009), or optically stimulated luminescence dates are available from a broad range of contexts on Caddo sites of different ages to put the “phase” and “period” out of work. We are not there yet, as few Caddo sites have more than 10–20 absolute dates. Consequently, phases and periods will take on most of the heavy lifting in this book when there are discussions and comparisons made of the specific archaeological content of site assemblages and components in different regions and localities in the Caddo archaeological area. Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

13

Contents of the Volume

Copyright © 2012. Nebraska Paperback. All rights reserved.

The chapters in this volume run the gamut of archaeological issues and research problems of concern to Caddo archaeologists. Certainly a fundamental issue is the stylistic character of Caddo decorated ceramic vessels and the importance of understanding ceramic stylistic variation in exploring the temporal, geographical, communitywide, and sociocultural character of prehistoric groups in Arkansas, Louisiana, and Texas. Ann Early (chap. 2) considers the form and structure of prehistoric Caddo pottery in the mid-Ouachita region in southwest Arkansas. The complex decorative patterns and vessel shapes made by Caddo potters represent one of the great independent developments of North American Indian ceramic technology. Patterns and designs on ceramic vessels are regularly employed by archaeologists as sources of cultural and historical information, and this has certainly been the case in Caddo archaeological investigations since its early days. However, the sophisticated nonrepresentational ceramic decorative tradition of the Caddo may be brought into play to investigate other aspects of their culture. Early’s consideration of design organization and pattern analysis of mid-Ouachita Caddo ceramics explores a grammar of design and tests whether this underlying structure expresses more fundamental cultural concepts that united regional Caddo societies. Using a sample of 300 Friendship Engraved vessels, Early outlines the process of design building and investigates whether this process embodies broader symbolic principles expressed in other material and behavioral aspects of Caddo culture. In chapter 9, Jeffrey S. Girard examines a large archaeological ceramic database at a spatial scale concordant with a dispersed Caddo floodplain community in northwest Louisiana and discusses the implications of the ceramic findings for the study of inter-community variation and past Caddo social structure across the landscape. Archaeological investigations along Willow Chute Bayou have identified numerous archaeological sites (through repeated surface collections and test excavations) with similar ceramic assemblages that date between 1050 and 1450. The locality proved well suited to addressing questions concerning the meaning of ceramic variation at a community level because it is at a spatial scale consistent with Caddo communities described in historic records, and the locality is bounded by natural areas that have significant limita14

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

1-3. The geographic extent within the Southern and Northern Caddo areas discussed by contributors in this book.

tions for human habitation. Using primarily correspondence analysis, Girard attributes most of the variability between ceramic collections along Willow Chute Bayou to spatial shifts in Caddo settlement through time, beginning with the initial development of the community from a small aggregated hamlet to the establishment of a dispersed village. The Oak Hill Village (41RK214) in the Sabine River basin of northeast Texas is a ca. A.D. 1150–1450 Caddo settlement that was completely excavated in the mid-1990s prior to lignite mining activities. Analysis of Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

15

Copyright © 2012. Nebraska Paperback. All rights reserved.

the architectural remains, including more than 40 structures and several granaries, along with a series of key calibrated radiocarbon dates from habitation features, and finely detected changes in ceramic vessel stylistic decorative elements by Timothy K. Perttula and Robert Rogers (chap. 8), indicate that the village evolved as three temporally and spatially different communities composed of a number of separate households. Emerging in the latter two communities, after 1250, were important social institutions (a plaza, an earthen mound, and specialized structures with extended entranceways) that bound this singular Caddo community together for 150–200 years (ca. 1250–1450). Other contributions in the volume are concerned with the organization and interrelationship of Caddo communities at various spatial scales. This includes the evolution at the micro-scale of the aforementioned Oak Hill village in northeast Texas; the development and dispersion of a larger Middle Caddo period community in the Red River floodplain of northwest Louisiana; the intra-site organization of several mound centers in east Texas and southwest Arkansas as discerned through archaeogeophysical study; and the meaning of the regional distribution of Caddo mound centers and associated habitation sites in the Arkansas River basin, the Red River basin in southeastern Oklahoma, and the Red and Ouachita River basins in southwest Arkansas. One chapter considers the cosmological significance of the layout, organization, and directionality of structures, a temple mound, and open spaces at an important Caddo mound center and village on the Red River visited and mapped by a 1691–1692 Spanish entrada. In chapter 7, Chester P. Walker and Duncan McKinnon discuss recent archaeogeophysical investigations at several pivotal Caddo mound and village sites along the Red River and in east Texas, namely, the George C. Davis (41CE19), Hill Farm (41BW169, part of the Nasoni Caddo village known as the Hatchel site [41BW3]), and Battle mound (3LA1) sites. This work obtained primary datasets of geophysical features that allow a more detailed consideration of community organization at a landscape scale as well as a better appreciation of the character and intra-site distribution of Caddo architectural features. It is becoming increasingly economically feasible to use archaeogeophysics as a complementary investigative tool in the study of Caddo community spatial organization across the landscape, and thus the use of archaeogeophysical data as a primary dataset in landscape archaeology studies can obtain compre16

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

hensive images of overall site spatial layout and intra-organization of structures, mounds, and other cultural and geological features that are difficult to identify using traditional archaeological methods. Walker and McKinnon point out that where excavations have the ability to confirm interpretations of individual archaeogeophysical anomalies, a broad-scale map of the overall site spatial layout from archaeogeophysical investigations complements these datasets with an analysis of spatial relationships and patterning of cultural features across entire sites. The visual impact of mounds has long been underestimated. However, because many mounds were among the most prominent features of the landscape in ancient times, were covered periodically with brightly colored soils, were topped by substantial structures and occasionally with large fires, and were constructed in ways that made evident the effort in human planning and labor it took to construct them, large mounds were grand and impressive sights. In chapter 6, Gregory Vogel reviews visual aspects of Caddo mounds in the Arkansas River basin in eastern Oklahoma and west Arkansas, and presents a Geographic Information Systems (gis) model of mound viewsheds. Vogel determined that mounds along the Arkansas River and its major tributaries, particularly the mounds at larger and more elaborate civic-ceremonial centers, were located in areas where they could be seen from a distance. Mounds in the Ozark Plateau region, however, are more likely to have been built in areas with small viewsheds, suggesting that these monuments had different uses or meaning to Ozark Plateau aboriginal populations than was the case among the more substantial Arkansas River valley populations. Additionally, several mounds in the Arkansas River basin were situated where ground level at the base of the mound was hidden from the view of the adjacent bottomlands, but the mound itself, and any structure it may have supported, would have been visually prominent from the same bottomlands. The Caddo groups of southeastern Oklahoma made deliberate choices in construction of their built environment after 1000. For some 500–600 years, the Caddo built residential communities and mounds along the Red River and its principal southward-flowing tributaries. Robert Brooks (chap. 12) considers the distribution of these sites through a gis approach. Site selection criteria included soils, elevation, floodplain width, biotic habitats, and the presence of other nearby Caddo residences or mound centers. Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

17

Copyright © 2012. Nebraska Paperback. All rights reserved.

Recent geophysical surveys and excavations at the Tom Jones site (3HE40) in southwest Arkansas have provided information about intrasite organizational patterns and material culture (see Schambach 2003; Lockhart 2007). Jami J. Lockhart (chap. 11) extends the search for recognizable inter-site spatial patterning to the broader region, encompassing pre-1680 Caddo landscapes within Arkansas’s West Gulf Coastal Plain physiographic subdivision. Employing the statewide database of archaeological sites, gis technology, ethnographic and archaeological information, and exploratory data analysis techniques, Lockhart details a research approach for recognizing and integrating natural and cultural components of past landscapes. Quantifying relationships between environmental themes and the locations chosen by the Caddo for mound sites are used to develop an environmental similarity model, which, combined with a statistically derived area of accessibility, encompasses 92 percent of all Caddo sites within less than half of the West Gulf Coastal Plain. This model also delineates potential cultural buffer areas to the east that contain land environmentally similar to the land chosen by the Caddo, but in which there are no sites, and also identifies an atypical set (11 percent) of Caddo mounds sites that are entirely surrounded by, but not included in, the modeled area of environmental similarity. These few Caddo mound sites, among them the Tom Jones site, were located in the highest elevations of the West Gulf Coastal Plain in proximity to major watershed boundaries and small, isolated refugia of Blackland Prairie ecosystems. In the volume’s concluding chapter, George Sabo examines a map produced in connection with an early Spanish entrada to the Caddo, the 1691–1692 exploration of east Texas by Domingo Terán de los Ríos. The Terán map depicts a Nasoni Caddo village along the Red River, including farmstead compounds, the headman’s compound, and a temple mound complex at the western end of the village. Sabo convincingly argues that the position of the temple mound complex within the village marks a boundary separating the community not only from an external terrestrial world but also from its associated spiritual realm. The map is interpreted by Sabo as a cosmogram captured by the Spanish mapmaker that coincidentally incorporated a Nasoni Caddo view of their place in a spiritually transcended world. Chapters by Diane Wilson and Edwin Jackson, Susan L. Scott, and Frank Schambach consider different aspects of Caddo subsistence from 18

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

east Texas and the Crenshaw mound site, respectively. Their contributions point out that subsistence strategies and the utilization of domesticated plants and various wild plant and animal foods were tremendously variable among Caddo populations through time and across space, as well as equally diverse between elite and non-elite groups. The Caddo believe that cultivated plants such as corn and pumpkin are gifts from the earth that “they were to hold and use for their benefit. The two gifts most closely associated with the earth were corn and pumpkin” (Newkumet and Meredith 1988:30). The adoption and use of maize sometime before 900 is one of the key features that defines Caddo culture archaeologically, along with the cultivation of beans, pumpkin, and various oily and starchy seeds (chenopods, sunflower, sumpweed, and maygrass), because it is not seen in their hunter-gatherer predecessors, but yet there is growing archaeological and bioarchaeological evidence of significant variability in the utilization of maize by the east Texas Caddo. The analysis of recovered plant remains suggest that the Caddo began to use tropical cultigens after the ninth century, but intensive food production based on maize apparently became most important across almost all of the traditional Caddo homelands only after the thirteenth century. Diane Wilson (chap. 4) reviews the dental and stable isotope evidence from the bioarchaeological study of human remains for maize subsistence among east Texas Caddo peoples. Maize subsistence generally increased through time, but there was considerable variability by individual, site, and region. Her synthesis has led to an understanding of the bioarchaeological character of the Caddo people in parts of east Texas, especially in the upper Neches River basin, that is currently unparalleled anywhere in the larger Caddo archaeological area (cf. Wilson 2011). The bioarchaeological information she gathered regarding diet, health, and pathologies obtained during the course of the Lang Pasture site (41AN38) investigations in 2006 has an import well beyond the local archaeological context, because it provides a sweeping view of more than 800 years of Caddo life. The Lang Pasture bioarchaeological investigations were done in consultation with the Caddo Nation of Oklahoma. That consultation has allowed the Caddo peoples to keep abreast of the bioarchaeological approach and findings from the project as it moved forward and provided them an opportunity to communicate their Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

19

Copyright © 2012. Nebraska Paperback. All rights reserved.

perspectives on the meaning of the findings from these investigations. This is a first for the field of Caddo archaeology, and it is a step in the right direction of making room for the Caddo peoples to be full partners in future archaeological and bioarchaeological investigations in their traditional homelands. The Crenshaw site (3MI6) is a late Fourche Maline–Early Caddo period ceremonial center located on the Great Bend of the Red River in southwest Arkansas. Excavations some years ago by Frank Schambach uncovered a structure, “House of the Priest,” that was used between 1000 and 1200 as the residence of an individual with both elite social status and religious responsibilities. Adjacent to the structure was a truly impressive deposit of 2,042 antlers, representing a minimum of 1,021 white-tailed deer. As discussed by Edwin Jackson, Susan L. Scott, and Frank Schambach in chapter 3, the very well preserved faunal assemblage from the Crenshaw site provides a sense of the ways in which patterns of faunal procurement and consumption in this Caddo ceremonial center transcended economic considerations. The character of the faunal assemblage from those excavations have led Jackson et al. to strongly argue that remains found there were the product of individual social and political status and preferential access to quality wild animal foodstuffs for the elite, and decisions on wild animal procurement that would have helped create, maintain, and demonstrate the supernatural relationships that would have legitimized and supported the elite status quo. Ultimately, the faunal assemblage from the “House of the Priest” at the Crenshaw site represents a considerable departure from what would otherwise be expected as typical wild animal food dietary patterns for Early Caddo sedentary communities along the Red River and other parts of the Caddo area (cf. Hunter et al. 2002; Perttula and Bruseth 1983). Several chapters examine the character of specific sites and welldefined archaeological phases in different parts of the Caddo area. These studies nicely illustrate the diversity in material culture content, social relationships, subsistence, and settlement character that existed in broadly contemporaneous Late Caddo period groups. Mary Beth Trubitt (chap. 10) focuses on the Caddo archaeological record in the Saline River valley in south central Arkansas. The Saline River valley was something of a “borderlands” between the trans-Mississippi south and lower Mississippi valley archaeological areas, and between Caddo and Mississippian cultures in the Late Caddo/protohistoric pe20

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

riod. Recent archaeological research on this poorly known area has included work at the Hughes site (3SA11), an important mound center, designed to reach a better understanding of Caddo social and ceremonial systems, settlement patterning and economic organization, and relationships with neighboring groups. Excavations at the site provide a glimpse into activities conducted near the main mound, based on the excavation of portions of stratified burned structures that date to at least the fourteenth and fifteenth centuries. Trubitt summarizes the results of the Hughes site excavations and compares the archaeological record of the Saline River valley with the neighboring Middle Ouachita River valley, Ouachita Mountains, and Arkansas River valley. The Caddo peoples living in the Big Cypress Creek basin in northeast Texas from the fifteenth to the seventeenth centuries, known archaeologically as the Titus phase, are the topic of Timothy K. Perttula’s chapter 13. These Caddo, never ethnographically documented, were a strong and powerful group of peoples, not a society in eclipse. They were farmers living in dispersed communities, and they were active traders. These Caddo groups were among the most populous and socially complex of the many Caddo societies living at that time, and they were the westernmost aboriginal group that was sociopolitically akin to middle and late Mississippian polities in the southeastern United States. The Caddo communities living in the heartland of the Big Cypress Creek basin experienced rapid and sustained population growth during times of fluctuating climatic conditions. These dynamic farming communities dealt with climatic and subsistence stresses by effecting new means of holding their societies together, forming several stronger communities with larger mound centers, community cemeteries, and villages at key nexuses in the Big Cypress Creek basin. In chapter 14, David Kelley summarizes the archaeological character of the Belcher phase (ca. 1500–1700) in northwest Louisiana. The Belcher phase is one of the better known Late Caddo period archaeological manifestations, and it represents the peak of Caddo settlement in this part of the Red River valley. The formal archaeological content of the phase was originally based on Clarence H. Webb’s (1959) excavations at the Belcher mound site (16CD13) in northern Caddo Parish, Louisiana. His work provided the first detailed information on the nature of the ceramic and lithic assemblages, architectural patterns, and burial programs of the upper level of Belcher phase Caddo society. More Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

21

recent excavations at Caddo farmsteads belonging to the Belcher phase have gathered important new data on Caddo settlement and social hierarchies, leading to a more comprehensive view of the Belcher phase, especially in terms of settlement patterns, the subsistence character of these Caddo people, and material culture assemblages. Wrapping up the volume, chapter 5 by James A. Brown is concerned with the utterly unparalleled (i.e., unparalleled in the Caddo area, and unparalleled in the eastern and southeastern United States) assemblage of exotic mortuary goods recovered from the Spiro site in the Arkansas River basin in eastern Oklahoma at the western frontier of the eastern Woodlands. The changing interpretation of these mortuary goods has challenged customary ways of explaining the pre-Columbian past here and in the larger context of the native history of the eastern United States.3 Debate has centered on the archaeological interpretation of the findings in the Great Mortuary in the Craig mound at Spiro, which Brown demonstrates in a tour de force to be a three-tiered sacred monument organized around an axis mundi upright pole. Sacred objects were laid out in the Great Mortuary to represent the universe, the imago mundi. The accumulation of shell, copper, and other mortuary objects is best attributed to the attractions of the “sacred economy,” not the operation of a centralized chiefdom or the mobilization of a long-distance network of trade and exchange. Conclusions

Copyright © 2012. Nebraska Paperback. All rights reserved.

The chapters in this volume consider the long sweep of Caddo native history, from its origins and ancestral times, still clouded in mystery and limited archaeological investigations, to the early years of contact between the Caddo and European adventurers and explorers. Just as the Caddo built impressive monuments (fig. 1-4) to honor their leaders and their god above, Ayo-Caddi-Aymay, we hope that the contributions in this volume honor the strengths, creativity, and traditions of the Caddo peoples. In recent years, in pace with developments in the lower Mississippi valley (see Dunnell 2008b), the Southeast (Livingood 2008, 2010), and adjoining archaeological regions (Baugh and Perkins 2008), while the Caddo archaeological interest with temporal-spatial systematics has remained strong (along with approaches that can discern rapid ce22

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

1-4. Looking south-southwest at the Battle mound site (3LA1) in southwest Arkansas in 2007.

ramic stylistic changes with fine chronological control, see Schambach and Miller 1984), Caddo archaeologists have also turned to addressing broader and equally significant research problems about the Caddo peoples and past communities that concern such themes as settlement patterning, sociopolitical organization, construction and destruction of Caddo houses, ceremony and ritual, iconography, subsistence change, health and adaptive efficiency, and exchange networks (see Brown 1996, 2007, 2010; Early 2000, 2004; Emerson and Girard 2004; Kay and Sabo 2006; Lankford 2007a, 2007b, 2008; McKinnon 2009; Perttula 1996, 2008; Trubitt 2009). Even the discussion of Caddo warfare with Southern Plains people (but see Baugh 2007) has entered the Caddo archaeological field (Burnett 2010), to account for the dismembered remains of more than 350 individuals at the Crenshaw site (see Jackson et al., chap. 3, this volume). These developments have taken place in tandem with the burgeoning field of cultural resources management and its accompanying laws, regulations, and practices—under which most Caddo archaeological work is done these days—and in conjunction with the development and use of new scientific methods, the most significant among them being geophysical surveys of Caddo sites (Lockhart 2007; Perttula 2010; Perttula et al. 2008; Walker 2009; see Walker and McKinnon, chap. 7, this Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

23

volume). Caddo archaeologists have also come to more fully appreciate the social and cultural complexity and diversity that characterize the Caddo peoples’ heritage (cf. Newkumet and Meredith 1988; Carter 1995; Halfmoon 2004; Gonzalez et al. 2005; Sabo 2005; Gregory 2009). There is every reason to think that we will continue to learn a great deal more about the native history of the ancestors of the Caddo peoples who lived in the Caddo area, building upon the archaeological findings of those who came before us. To honor the modern Caddo peoples, as well as their ancestors, the royalties from this book are dedicated to the Caddo Nation of Oklahoma. Acknowledgments

Copyright © 2012. Nebraska Paperback. All rights reserved.

I would first wish to thank Elisabeth Chretien and Elaine Maruhn at the University of Nebraska Press for their assistance in bringing this volume to fruition. Chester P. Walker and I also appreciate the comments and efforts of the two anonymous book manuscript reviewers, as well as the work of each of the contributing authors to this book, and we thank the contributors for their patience as this volume was assembled. I also wish to thank the Society for American Archaeology (saa) and John Neikirk of the saa for permission to reproduce chapter 8 from a 2007 volume of American Antiquity and for permission to reproduce three photographs from Webb (1959) in chapter 14. Pictures of Record, Inc. also gave permission to reproduce three photographs in chapter 2. Sandra L. Hannum prepared several figures that are used here in chapter 1. Finally, I would like to thank the Caddo Nation of Oklahoma for everything they have done for me over the years, especially Bobby Gonzalez (Native American Graves Protection and Repatriation Act coordinator), Robert Cast (Tribal Historic Preservation Officer), LaRue Parker (past tribal chairwoman), and Elizabeth Edwards (current tribal chairwoman). The editors and authors hope this volume, in some small way, is a measure of payback and appreciation in return for their support of the archaeological community’s studies of the Caddo Indian people. Notes 1. The northern Caddo area (or Northern Caddoan area, as it is often referred to) even includes a bit of the Ozark Highlands in southwest Missouri (Ray and Lopinot 2008). 2. A recent Caddo bibliography of archaeological, bioarchaeological, ethno-

24

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

historical, ethnographic, and historical writings on the Caddo Indians has been published by the Arkansas Archeological Survey (Perttula et al. 2006), and it is now updated on a website developed by the Caddo Conference Organization. Archaeological research on the Caddo may also be found in the Caddo Archeology Journal, published by Stephen F. Austin State University (Nacogdoches, Texas), and generally in the publications of the Arkansas, Louisiana, Missouri, Oklahoma, and Texas archaeological societies. 3. Some archaeologists still give credence to the notion that the archaeology of the Arkansas River basin in eastern Oklahoma is linked with the Wichita people. For instance, Don G. Wyckoff (2008:377) has recently written that “some of us believe this cultural tradition [Arkansas Basin] has something to do with one, if not more, of the Wichita bands.” The presentation of the pros and cons of this issue of ethnic and cultural affiliation of the Arkansas Basin peoples that lived there between ca. 900 and the late seventeenth century (Rogers 2006; Baugh 2009) would take at least another volume to summarize, if even then. The reader should consult the chapters here by James A. Brown and Gregory Vogel for their perspectives, the writings of archaeologists who have worked in the Arkansas Basin, and the views of the Caddo Indian peoples.

Archaeology of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

25

2. Form and Structure in Prehistoric Caddo Pottery Design ann m. early

Complex decorative patterns and vessel shapes made by prehistoric Caddo potters represent one of the great independent and in situ developments of North American Indian ceramic technology. Archaeologists typically use patterns and designs as sources of cultural-historical information (cf. Phillips 1970; Schambach and Miller 1984). The highly sophisticated and nonrepresentational ceramic decorative tradition of the Caddo is information-rich and ought to be used to investigate other aspects of their culture. Design organization and construction offer an avenue to explore a grammar of design in a southeastern U.S. Indian culture and to test whether this structure expressed a more fundamental set of cultural concepts that were shared among regional societies, linking them into a Caddo whole. This chapter uses a sample of pottery vessels from a single decorative variety to test the information potential in one closely related pottery group. Copyright © 2012. Nebraska Paperback. All rights reserved.

Some General Observations on Caddo Ceramics One archaeological benchmark used to identify the emergence of the Caddo tradition is the appearance of new and distinctive pottery making practices. These include both new vessel shapes and new approaches to decoration. At about the same time, changes in cultural attitudes about the treatment of the dead included placing numerous pottery vessels in graves. As a result, more pottery was manufactured for both the living and the dead, and large numbers of complete vessels were buried, rather than going through a life cycle of manufacture, use, breakage, and fragmentary abandonment or disposal. Relic collecting, archaeological investigations, and natural erosion events have brought many thousands of complete vessels to light. Consequently skillfully fashioned and elab26

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

orately decorated pottery has become emblematic of pre-nineteenthcentury Caddo culture. Most archaeological interest has focused on ways to use pottery to delineate basic geographic and temporal relationships among Caddo sites and regions (cf. Suhm et al. 1954). Several chapters in this volume, including those by Perttula and Rogers (chap. 8), Girard (chap. 9), Trubitt (chap. 10), and Kelley (chap. 14), show how fine-grained observations of even basic decorative techniques observable in sherd collections can be used in studies of local settlement histories and social relationships. Moving beyond issues of typology, ceramics still have value to advance both local and regional Caddo studies. Research into technological and stylistic attributes of whole ceramics or sherd collections has lagged behind. Technological studies, while few, are more common than stylistic studies (cf. Arnold 1975; Fishbeck et al. 1989; Perttula 2002; Perttula et al. 2003; Perttula and Ferguson 2010) and have largely involved clay sourcing and the geographic distribution of key ceramics. Prewitt’s 1969 pioneering attempt at stylistic analysis of one group of ceramics is virtually the only study of its kind, although variation in decoration, along with other attributes, has been used in some larger Caddo studies (Thurmond 1990; Perttula 2000). This is unfortunate, because Caddo ceramics are information-rich, especially with respect to decoration, and have been available in quantities that are not easily duplicated in other regions of North America. Before proceeding with the specifics of this study, it is worthwhile to take note of some general characteristics of Caddo pottery. These observations are based on the relatively rare opportunity to see and compare thousands of complete Caddo pottery vessels in museums, archives, and publications. Two Pottery-Making Trajectories Throughout the Caddo tradition, potters made two distinctly different wares. The first, commonly called utilitarian ware, includes a variety of jar and bowl forms. Decoration, applied when the clay was plastic, consists of incising, brushing, punctation, and stamping. Sometimes only one technique is used on a vessel, but frequently two are used. During some centuries of the Caddo tradition, and in some river valleys, utilitarian rim and body areas were completely decorated, although this was not the case in all parts of the Caddo world. Form and Structure in Caddo Pottery Design

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

27

At the same time, potters were producing fine ware that included bottles, bowls, beakers, seed jars, and a variety of compound, or hybrid, vessel forms. A minority percentage of fine wares were highly polished but otherwise undecorated or were incised when the vessel was nearly fully dried. Most were engraved, which seems to have been executed on hardened vessels before actual firing. The two traditions maintained separate repertoires of vessel shapes, and there is no crossover between them. A sample of the range of variation of both utilitarian and fine ware ceramics, and the general location within the Caddo area where they were found, can be seen in Suhm and Jelks (1962), but no more comprehensive published catalog of Caddo pottery exists. Vessel Form Previews Design

Copyright © 2012. Nebraska Paperback. All rights reserved.

Vessels typically are built with crisply defined parts: transitions from lip to rim to body to base are cleanly delineated, both with breaks in vessel profile and changes in decoration. Caddo potters rarely modeled effigy vessels, although they did put anthropomorphic or zoomorphic rim tabs on some bowls. There are also rare examples of highly stylized, modeled human head bottles, as well as seed jars fashioned evidently as bears or dogs, but overall, Caddo pottery has clean profiles and nonrepresentational surface decorations. With regard to decoration, specific designs were executed with specific techniques on specific body parts of specific vessel shapes, with virtually no exception to what must have been universally understood rules. For instance, one group of designs will be found only on jar rims, will be executed only with incising and punctation, and will be found only on jars with a specifically nuanced profile and configuration that is perceptibly different from other contemporary jars found in adjoining river valleys. This means that when a Caddo potter began to build a vessel, the choice of decorative technique and design was already made, and the container conformed to a very narrow set of options in preparing a landscape of appropriate design fields for the decorator. It is worth noting that we do not know if the same person was both the fabricator and the decorator for pottery vessels, if these roles were separated during part or all of the Caddo tradition, or if gender or social situation had an effect on who made or decorated pottery. In addition to limiting design choices for the vessel as a whole, the vessel shape usually created two or more design fields on which only a limited range of combinations could be applied. Individual creativity or 28

Early

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

whim had little, if any, place in decorating Caddo pottery. Each design intended for a vessel rim or body had a limited set of companion design options. In some times and places in the Caddo area, design options involved using two techniques, one on rims and a second on companion bodies. This is particularly true of utilitarian vessels. Published Caddo pottery types that show this include Kiam Incised, Bullard Brushed, Nash Neck Banded, Maydelle Incised, La Rue Neck Banded, Harleton Appliquéd, and numerous untyped examples, all illustrated in Suhm and Jelks (1962). Most of these types are Late Caddo in age, but the practice of combining two decorative techniques on utilitarian wares is also found among Early Caddo utilitarian wares as illustrated by Newell and Krieger (1949:figs. 64 and 65) with ceramics from the East site (3CL21). Krieger photographed these vessels while they were still in private hands, before their transfer to the Henderson State University Museum collection. There were also strong correspondences between specific design motifs and particular decorative techniques. Some motifs were executed only by engraving, for instance, while others may have been made with either engraving or dry line incising.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Vessel Type Previews Decorative Technique Caddo potters followed a small set of choices in selecting particular decorative techniques on fine or utilitarian wares. Engraving is found exclusively on fine wares. More rarely, dry paste incising (lines drawn on firm but not completely dried paste, thereby compressing the clay while scribing a line rather than scratching it) and punctation were used with specific design motifs. Punctation, incising, brushing, and stamping—with a wide range of tools that produce a variety of distinctive visual and plastic effects—were used on utilitarian wares. Different techniques were often used in sequence on different decorative areas as vessels dried. Some early Caddo fine ware bowls, for instance, like some Crockett Curvilinear Incised specimens (Suhm and Jelks 1962:31–34 and plate 17b), may have dry paste incising in a horizontal band below the lip and a lower panel of engraved complex designs. Tandem Changes in Vessel Types and Designs through Time and across Space There is evidence that potters in each river valley were familiar with what their neighbors were producing and made local versions of a broadly acknowledged, appropriate suite of containers found across the Caddo Form and Structure in Caddo Pottery Design

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

29

Copyright © 2012. Nebraska Paperback. All rights reserved.

region. For instance, Middle Caddo period utilitarian wares Military Road Incised and Haley Complicated Incised (Suhm and Jelks 1962:59– 60, 107–108 and plates 30 and 54) are found in adjoining river valleys, the middle Ouachita valley and the Little Missouri–Great Bend of the Red areas, respectively, in southwestern Arkansas. The Little Missouri River is a major western tributary of the Ouachita River, and it is situated between the main channel of the Ouachita and the Great Bend of Red River. Feeder tributaries for all three drainages arise close to each other in the uplands, and it is only a short walk from one drainage basin to the other. Both pottery types consist of flat-based, globular-bodied jars with tall flaring rims and decoration on both rims and bodies. The decorations consist of both rectilinear and curvilinear design elements executed with both wet paste incising and punctation. The punctations differ in execution between the two types, however. This may be differences in the type of punctation tool or tools selected or differences in the timing of punctation in the moist paste. The overall effect is to produce visually different kinds of punctations on the two types. If samples of whole vessels of the two types are placed side by side, the characteristics unique to each type are easy to see. Haley Complicated Incised decoration consists of concentric circular design elements covering vessel bodies. Sometimes the circular elements are linked with whorls, guilloches, and other connecting elements. Designs are commonly subdivided into panels with punctated appliqué ridges or fillets. Vessel rims are frequently castellated and bear pairs of strap handles. In contrast, Military Road Incised jars have body decorations consisting of continuous horizontal or meandering bands around the circumference of the vessel. These designs are also frequently found on the vessel rims. Strap handles, castellation, and appliqué ridges dividing the body of the vessel into design fields are not present. These two pottery types are easily distinguishable when whole vessels are compared. On the sherd level, the differences may not be evident unless an analyst knows what the whole vessels look like. The Case Study Friendship Engraved The name “Friendship” was first proposed by Dr. Thomas L. Hodges and his wife, Charlotte, artifact collectors and Arkansas history supporters, for a group of pottery vessels that were found near their home in 30

Early

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

2-1. Map of Arkansas showing the area where the sample of Friendship Engraved, var. Freeman vessels used in this study came from, and the 1938–1950 collecting area of the Henderson State University Museum and Hodges pottery collections.

southwest Arkansas (Hodges and Hodges 1945:106). The Hodges collected nearly 1,500 pottery vessels, and more than 48,000 other artifacts, almost exclusively from a stretch of the Ouachita River valley about 50 km long (fig. 2-1). This location, downstream from the place where the river leaves the Ouachita Mountains and enters the uplands of the West Gulf Coastal Plain, features a broad alluvial valley and a dense distribution of pre–European contact Caddo settlements, and has been called the Middle Ouachita Archeological Region in southwest Arkansas (Schambach et al. 1982). Friendship is a modern hamlet situated in the uplands overlooking the alluvial valley near the midpoint in this region. The Hodges proposed “Friendship” for a group of engraved carinated bowls. Suhm et al. (1954) formalized Friendship Engraved as a pottery Form and Structure in Caddo Pottery Design

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

31

Copyright © 2012. Nebraska Paperback. All rights reserved.

type, adding a group of engraved bottles and other vessel types to the carinated bowls (Suhm et al. 1954:274 and plate 21). The type specimens came from the Hodges collection and from sites in the upper Ouachita River valley in the Ouachita Mountains (Suhm and Jelks 1962:45 and plate 23). Distinguishing criteria for the type included rectilinear engraved designs and the use of cross-hatching as a filler or background treatment, but a wide variety of designs were included in the type. When archaeologists became more active in the region around 1970, they embarked on efforts to replace the ceramic framework of Suhm et al. (1954) with an approach based on Philip Phillips’s Type-Variety system, a process still under way. This change provided a means to partition Suhm et al.’s broad types into smaller and more formally uniform groupings. In the case of Friendship Engraved, Frank Schambach (1970:412–420) suggested divisions, and Cynthia Weber drafted a series of new types and varieties based on examples among the roughly 700 minimally provenienced whole vessels in the Henderson State University Museum in Arkadelphia (Weber 1972; Weber and Loring 1971). These manuscripts remain unpublished, but some of the proposed types and varieties were formalized in subsequent studies (cf. Early 1988; Early, ed. 1993) when collections and contextual information were available. The Henderson State University Museum pottery collection was originally assembled by a group of local collectors in the 1930s and 1940s from prehistoric Caddo sites situated largely within about 15 km of Arkadelphia (see fig. 2-1), and it was donated to the University when the museum was formed in 1954. The Henderson State University Museum and Hodges collections come from sites that overlap geographically and culturally, but were separate collecting areas frequented by two groups of amateur diggers and collectors active in the Ouachita River valley. Suhm et al. (1954) drew on firsthand examination of both collections, as well as the University of Arkansas collections, when laying out their Caddo pottery typology. As a result, several dozen type specimens illustrated in what has come to be known as their “Texas Handbook” of regional pottery and stone tool types are in these two large collections from a small geographic area in southwest Arkansas. The Freeman variety of Friendship Engraved is the vessel that Dr. and Mrs. Hodges used originally to illustrate their Friendship Engraved type (Hodges and Hodges 1945:fig. 1). Weber provided a preliminary description of this variety (Weber and Loring 1971), but a formal description 32

Early

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

2-2. Friendship Engraved, var. Freeman carinated bowl, showing scribing lines laying out the design fields.

first appeared in the Hardman site ceramic analysis (Early 1993:89), when at long last an excavated context was available. This is the most common example of any Caddo fine ware to be found in this region.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Friendship Engraved, Variety Freeman Vessel Configuration I assembled a sample of 251 whole vessels belonging to this variety for the study. Most were available for firsthand observation in the Henderson State University Museum and Hodges collections, the latter now owned by the Joint Educational Consortium of Arkadelphia, Arkansas. I collected information also from black-and-white prints and color slides of vessels from numerous other collections. All but 2 of the vessels in this sample came from a 30 km stretch of the Ouachita River valley (see fig. 2-1); 185 vessels are from 38 known archaeological sites, and another 64 are assignable to county but not to site locations. These carinated bowls have shallow, convex bodies below the angle of carination (fig. 2-2). Instead of flat, disk bases that were foundational to vessel construction, the body area in contact with the support surface is gently convex and just slightly thicker than the rest of the body. The angle of carination is sharp and clearly marked by a change of profile. The rim is straight to very slightly convex. It is situated at an angle about 90 degrees from the horizontal plane of the support surface, and in some examples it is leaning slightly toward the center of the vessel. There is a separately fashioned lip, so small as to be nearly vestigial and flaring outward away from the center of the vessel. There are no rim tabs, lugs, or effigies on the vessel rim, nor are there punctations or other embellishments on the lip. Form and Structure in Caddo Pottery Design

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

33

2-3. The first three steps in decoration: first, choosing a two- or three-panel format; second, dividing the panels into smaller, oval design fields; and third, choosing either an oval-within-an-oval or a single horizontal line element to fill the design field.

Copyright © 2012. Nebraska Paperback. All rights reserved.

The decorative field on these vessels is the entire rim area between the point of carination and the lip. Decoration consists of superimposed horizontal panels, two of which contain rectilinear elements running around the circumference of the vessel (fig. 2-2). Some vessels have pigment rubbed into the engraved lines, most commonly red or white. Bodies and vessel interiors are burnished or polished, but not otherwise decorated. Interior and exterior colors almost universally range from a rich mahogany brown to a very deep brown grading into black. Decorators engraved this design with a sure and heavy hand. Individual lines are deeply scored and core colors thus exposed offer vivid color and texture contrast against the polished surface where pigment has not been used. The fine interval cross-hatching gives a textural effect to the vessel surface, a dimensionality that is also found in contemporary utilitarian pottery. The heavy engraving reveals the sequence of design construction in the variable depth and overlap of lines displayed on most vessels (see fig. 2-2). This information and the range of finished products present in the study sample were key to reconstructing the sequence of steps through which vessels were used and discarded. Decoration Step One: First Decision The first step in decorating the bowls was to divide the rim into two or three horizontal panels (fig. 2-3). In the case of vessels with two divisions, each was filled with a repeating set of decorative elements, and 34

Early

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

2-4. The design choices found in the study sample. Each pair of columns shows the number and array of finishing options for each major design element and panel arrangement and the total number of each option in the study sample. The left member in each pair shows options with three panels, and the right member shows the two-panel options. Columns one and two show the options used with the “conjoined oval” element; columns three and four have the horizontal line element; columns five and six show vessels with the scroll element; and columns seven and eight show options associated with the “floating oval” element. Each row shows the number of vessels with the same finishing designs applied with major design elements. Empty cells are combinations not found in the study sample.

Copyright © 2012. Nebraska Paperback. All rights reserved.

the upper and lower sets were laid out to mirror but offset each other. When there were three divisions, the middle panel was filled with crosshatching to contrast with the other two. There are only 42 (17 percent) two-panel examples in the sample, indicating that the three-panel choice was by far the preferred option (fig. 2-4). Subsequent decorating choices found on the two-panel vessels are identical to those on the more common three-panel vessels, however, showing that the overall trajectory of decoration was the same. There may be a temporal factor in the appearance of these two options. A richly furnished grave at the Hardman site (3CL418) dating to the transition between the Mid-Ouachita and Social Hill phases has only two-panel examples present, along with a new short-rimmed type of carinated bowl, now called Cook Engraved, that lacks the cross-hatched background and most other variety Freeman decorative elements (Early 1993:87 and figure 54a-b). Cook Engraved vessels replace Friendship Engraved carinated bowls in Social Hill Phase assemblages after A.D. 1500. A few hundred meters away from the Hardman site discovery, Schambach salvaged a grave at the Bayou Sel site (3CL27) in 1966 that had a ceramic assemblage nearly identical to the Hardman grave (Early 1993:116 and fig. 71). There are also a small number of vessels that have central panels considerably thinner than the adjoining upper and lower panels (Early 1993:111 and fig. 67). This indicates that the middle panel was “squeezed out” while the vessel rim became shorter. Decoration Step Two: Framing the Design Area Copyright © 2012. Nebraska Paperback. All rights reserved.

This decorative step took place on both two- and three-panel bowls. The upper and lower panels were divided into decorative units with pairs of brackets. Spaces are left between opposing brackets, and brackets were placed so as to overlap adjoining bracketed units above or below. Each panel was divided into several decorative units: the most common number is four, but three- and five-unit examples exist. Decoration Step Two: One Finishing Option One option available at this point was to fill in the middle panel (if present) and spaces between bracketed units with cross-hatching, to complete one version of the vessel and put it into use. This trajectory was rarely pursued, because there are only two items in the sample that are 36

Early

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

finished in this manner, both with the three-panel option (see fig. 2-4, top row). All of the remaining vessels in the sample went through at least one or two subsequent steps to finish the decoration process. Decoration Step Three: Bracketed Designs, Primary Structure For the majority of vessels still in the design process, the next step was to choose one of two approaches to creating design elements within the bracketed spaces. Either a simple horizontal line or a narrow oval was scribed within each bracketed unit. Both design elements are shorter and narrower than the unit, so neither the line nor the oval touch either the edges of the panel or the end brackets. This step took place on both two- and three-panel vessels. Only 20 percent (n = 50) of the vessels in the sample have the “floating oval” element, most of them (n = 47) with the three-panel design framework. These vessels went through one more decorative stage before the background was filled in and they were put in use. This stage is mentioned below since it parallels the last design stage for the two most common design elements, the scroll and “conjoined oval,” found in the sample. The remaining vessels, with the horizontal line choice, went through two more stages before the background stage was reached.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Decoration Step Four: Bracketed Designs, Secondary Structures After the horizontal line bisecting the bracketed area was in place, decorators made one of four choices to refine this design element and complete the vessel (fig. 2-5). The first choice was to leave the horizontal line unembellished, fill in background cross-hatching, and put the vessel into use. This choice is found on both two-panel and three-panel vessels, 13 on the former and 8 on the latter, making up 8 percent of the total sample (see fig. 2-4, columns three and four, top row). The second choice was to embellish the horizontal line with a row of short ticks, predominantly oriented at an acute angle to the horizontal, and situated either above or below the line. Cross-hatching the background completed the vessel for use. There are 43 examples of this product in the sample, 17 percent of the total, and all but 3 are three-panel vessels (see figs. 2-4 and 2-5). The third and fourth choices available to decorators linked the horizontal line element to the surrounding oval with a set of brackets that were scribed at each end of the line. One option was to scribe linking Form and Structure in Caddo Pottery Design

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

37

2-5. Finishing stages and options. Moving from left to right, arrows show the sequence of choices and array of options in finishing the design and releasing the vessel to use. The column at the right shows the final appropriate design element choices known.

Copyright © 2012. Nebraska Paperback. All rights reserved.

brackets that parallel the oval end brackets, thereby creating a smaller oval within the larger one that is joined to the top or bottom of the larger oval, thereby making a pair of “conjoined” ovals. The alternative and second option was to scribe brackets that oppose the arc of the larger oval ends, thereby creating a simple horizontal scroll element within the oval (see fig. 2-5). In the case of both the “conjoined oval” and the scroll choices, these design elements are oriented toward either the top or bottom of the rim rather than to the center. Ovals and scrolls in the upper panel are oriented toward the lip. When the design elements are in the lower panel, they are oriented toward the point of carination. The visual effect is to “oppose” the design elements, both in terms of placement around the circumference of the vessel and in terms of verticality on the rim. Once these choices were applied, the majority of vessels went through one more stage in decoration before they were ready for the final application of background engraving. Some vessels, however, had back38

Early

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

2-6. Final decorative choices available for the oval-within-an-oval design option.

grounds applied without additional embellishment of the central design elements. There are three examples of the “conjoined oval” option, two of which are on three-panel vessels. The simple scroll design appears on 11 vessels, 9 of which have two decorative panels. Combined, these make up only 6 percent of the total sample. For the remaining vessels, the ovals and scrolls were “embellished” or finished with one of six decorative choices. These choices were also applicable to the oval-within-an-oval option that was created at Decoration Step 3 in opposition to the single line filler (fig. 2-6). Figure 2-4 displays the distribution and frequency of each finishing option as it appears on two- and three-panel vessels with each major decorative element. At the top of the figure is a finished example of a three-panel design with primary decorative oval fields that lack further additions. Beneath this are eight columns showing finished decorations. Looking from left to right, there are four pairs of columns. Each pair shows in turn three-panel and two-panel vessels, and each pair of columns shows one of the four major design elements used within the primary decorative oval field. For example, columns one and two show three-panel and two-panel examples with the “conjoined oval” design element, respectively. Columns three and four show three-panel and two-panel examples with horizontal line elements in the primary decorative oval element. The third pair of columns show examples with horizontal scroll elements and the last pair of columns have examples of oval-within-an-oval element vessels. Each column in this figure also displays the variety of final additions Form and Structure in Caddo Pottery Design

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

39

Copyright © 2012. Nebraska Paperback. All rights reserved.

applied to a particular decorative element and the number of examples in the pottery sample. Each row shows a single finishing detail and the range of vessel types on which it is known to occur. This figure indicates that for three of the major design elements, the conjoined oval, the scroll, and the oval-within-an-oval, there are the same three finishing options that could be chosen and that are present in the sample. These decorative elements are placed within the smaller ovals or in the space between the ends of the scrolls. The choices were (1) a double row of short ticked lines, (2) a row of single vertical lines that segment the space into equal divisions, or (3) a row of doubled vertical lines that segment the space into equal divisions (compare figs. 2-5 and 2-6). In addition, two of the three major design elements share two other finishing options in the sample. These are a horizontal row of closely spaced multiple vertical lines that segment the space into equal divisions and a pair of multiple vertical lines placed close to either end of this interior decorative space. There are three examples of a finishing option that is found thus far only on three-panel vessels with scroll decorative elements. The space is filled with a series of small circles that are made by scribing away the space in between them. Circular design elements created by scraping or scribing away the surrounding surface are diagnostic elements in Late Caddo Hodges Engraved type fine ware bowls and bottles found in this area after Friendship Engraved carinated bowls ceased to be made. Summing the examples in the sample by finishing element, it is evident that ticked lines are the most common choice (n = 143) and account for more than 57 percent of the total number of vessels in the study. The single line or “ladder-like” segments are the second most common (n = 48) at 19 percent (see fig. 2-4). Voids in the Table: Unacceptable Choices or Sampling Error? There are a number of empty cells in the figure 2-4 display of complete decorative options for Friendship Engraved. Some reflect limitations in the physical characteristics of design elements. Columns three and four show options for vessels with horizontal line decorative elements. Since the line does not create a closed space within the primary design structure, the options available for embellishing closed spaces cannot be used. 40

Early

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

2-7. Examples of “nongrammatical” vessels, each displaying two finishing decorative choices instead of only one. The vessels demonstrate the contemporaneity of these choices.

Some other empty cells, however, could be a result of sampling error, and yet to be discovered vessels may have the combination of panels and design elements that would fill these voids. This presumes that all of the options in each step of the decoration process were appropriate choices, and they were all circulating within the community of potters at the same time. One line of evidence demonstrates the contemporaneity of several choices from the suite of appropriate finishing options. Figure 2-7 shows three examples of “incorrectly,” although expertly crafted, vessels that violate the basic procedures of decoration by combining pairs of finishing options in unorthodox ways instead of using only one. In each of these three cases, the decorator has chosen two designs from the same group, alternating them around the vessel, instead of using only one design. In two of the three examples, two scroll designs were used. On the third, there are two options taken from the oval-within-an-oval group. The process of design construction was appropriate, however, as was the selection from a proper group of design elements. Singular Case or Broad Pattern Are the design elements or the decorative process exhibited by variety Freeman bowls unique to this one vessel category, time, or place? There are some other well documented fine wares from this river valley that are decorated with the same suite of finishing elements, indicating that the same principles guiding decoration on Freeman vessels were applied to other vessel categories. Form and Structure in Caddo Pottery Design

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

41

2-8. Friendship Engraved, var. Meador bowl.

Copyright © 2012. Nebraska Paperback. All rights reserved.

For instance, Friendship Engraved variety Meador consists of open bowls with a panel of engraved decoration on their rims (Early 1993:90; Wood 1981:fig. 16t). Instead of vertically separated horizontal panels, the decorative elements consist of stylized “S” elements that extend from the upper to the lower boundary of the panel and are nested against each other. The “S” elements are finished with the same choices used on variety Freeman bowls: ticking, equal spaced segments, undecorated areas framed with cross-hatch filler, and segments set out with multiple vertical line elements. Adjoining elements are set off from each other with cross-hatching and/or undecorated surfaces (fig. 2-8). Another example from the same region is a collection of bottles with globular bodies and relatively long necks. The design consists of concentric circle units repeated three times around the circumference of the vessel. Small circular or triangular elements are set in the blank areas between the upper and/or lower edges of the concentric units. The concentric circles are scribed with alternating types of lines. The options are essentially the same as the finishing elements on the Friendship bowls: simple single lines, multiple lines, lines with ticks, or doubled lines divided into segments. One specimen comes from the same grave at the Hardman site as the two-panel examples of variety Freeman mentioned previously, but numerous additional examples exist in local collections. The group comprises the Manchester variety of Belcher Engraved (Early 1993:86–87 and fig. 53c).

42

Early

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Decorating Principles In looking at the results of this case study, a number of decorative principles appear to have been at work that may be common to a class of pottery vessels, to a drainage basin, or to the Caddo ceramic tradition as a whole. One principle is the orderly and hierarchical nature of the decision-making process itself. A second is the subdivision of the decorative field into vertical and horizontal units. These directional dimensions were commonly expressed in Caddo designs, particularly in the vertical arrangement of several fields on a vessel, or what I call stacking designs. Stacking fields as a way of organizing space is found on several other Caddo pottery types, as well as on the vessels examined here. This can apply both to the design field created on one part of a vessel, such as a rim or body, and to the vertically stacked arrangements of decorations on body parts. In some fine wares, more than one design field was created on a single body part. This is particularly true of body decorations on bottles, such as the examples illustrated as Friendship Engraved in the original type definition (Suhm and Jelks 1962:46 and plate 23o–q), where two separate decorative panels were placed on the vessel, one above the other. The stacking concept was extended, at least in some instances, to the physical arrangement of vessels in space, usually bottle and bowl combinations, and to the fusion of two or more vessel types into a single object that simulated a nested set of separate items. Precise descriptions of the spatial relationships among complete pottery vessels in graves and other features are relatively rare in the literature, but one example indicates that this may be a fruitful source of information. Burial 18 at the Hardman site contained the remains of three men and a large number of grave offerings. All three skulls were missing, and in their place were pairs of vessels, a bottle and carinated bowl each, with the bowl overturned on top of the bottle (Early 1993:figs. 40–41). Fused vessels represent an extension and reiteration of decorative principles. These objects were made by assembling appropriate pairs or trios of fine ware vessels into a vertical composite made up most frequently of a bottle surmounting at least one carinated bowl. The bottles are actually half-vessels, because only the upper half of the vessel body was shaped before fusing it to the upper margin of the companion

Form and Structure in Caddo Pottery Design

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

43

Copyright © 2012. Nebraska Paperback. All rights reserved.

carinated bowl. Each vessel segment was then decorated with the appropriate design found on numerous complete and independent vessels matching each type and variety represented in the composite. Figure 2-9 shows three fused vessels. Figure 2-9a shows an Avery Engraved bottle, itself a compound figure with two decorative fields on the body, fused to a tall rimmed Means Engraved carinated bowl. The vessel, illustrated in Suhm and Jelks (1962:plate 53h), comes from the same part of the Ouachita River valley as the Friendship Engraved variety Freeman examples in this case study. Figure 2-9b is composed of portions of three vessels: the upper portion is a Keno Trailed variety Curtis bottle (Early 1988:77–79, 1993:94), the middle portion is the rim design from a Means Engraved carinated bowl, tilted outward, and the lower portion is a deep open bowl with a punctated lip. This bowl shape is also found as independent vessels with engraving below the lip, now defined as Hardman Engraved (Early 1993:91). Figure 2-9c is a Keno Trailed bottle seated in a shallow, incised rim bowl. The vessel was found at the Cedar Grove site (3LA97), a protohistoric Caddo settlement in the Great Bend region of the Red River. A number of vessels with the same pair of fused parts have been recovered elsewhere, including the Glendora site in the Ouachita River valley of north central Louisiana (Suhm and Jelks 1962:plate 44; Schambach and Miller 1984:123–124 and fig. 11-1; Moore 1909). Additional principles evident from this study are concerns for symmetry and opposition that are expressed in several ways. Hatched or cross-hatched backgrounds alternate with “negative” or blank blocks, both horizontally and vertically. Design elements have regularly placed attributes (i.e., ticks, segments, and rows of “negative” circles) and are arranged symmetrically within design fields. Unblemished polished surfaces are juxtaposed against heavy engraving that exposes matte core colors, thereby fashioning color and texture contrasts in fine ware vessels. Wet paste decorative techniques juxtaposed against each other in utilitarian wares—punctated fields against incising, for instance—likewise display oppositional textures. This concern with symmetry and opposition holds true throughout the ceramic tradition until ceramic making disappears in this part of the Caddo area in the eighteenth century. Beyond Ceramics If these rules of design construction are evident in Caddo ceramics produced in other communities and drainage basins within the greater 44

Early

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

a.

b.

Copyright © 2012. Nebraska Paperback. All rights reserved.

c. 2-9. Examples of stacked or fused vessels: a. HSU.3935; b. Means Engraved–Keno Trailed, var. Curtis; c. Keno Trailed. (Reproduced with permission of Pictures of Record, Inc.)

Caddo Area, are they concepts unique to the technological sphere of Caddo culture? Are they manifestations of broadly held concepts that were shared by members of Caddo societies as a whole, and may they be reflected in other aspects of material culture or social behaviors? George Sabo’s study of the structural underpinnings of historic period leadership among the Hasinai Caddo of East Texas offers a set of cultural categories and structural principles within several cultural domains that are independent of ceramics and suitable for comparison to the information derived in this study of Caddo pottery decoration. Form and Structure in Caddo Pottery Design

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

45

Copyright © 2012. Nebraska Paperback. All rights reserved.

Sabo proposes that symbolic representations evident in principles and behaviors in traditional kinship relations and in mytho-historical traditions “point to a more general structural feature, which is that the relationships among cultural categories comprising pairs or contrast sets (e.g.,human:superhuman, phenomenal:numinous, junior:senior, etc.) were structured according to a principle of hierarchy rather than opposition” (Sabo 1998:167). Sabo evaluated the categories and principles he proposed against descriptions of the material realm of Caddo houses and settlements and the behavioral realm of Caddo rituals. He concluded that several of the categories and principles that he delineated were expressed in the organization and use of space in houses and settlements and in ritual behavior. In particular, categories and their associated spatial distinctions related to hierarchical relations that were expressed in a number of ways, including vertical and horizontal orientations (Sabo 1998:172). The organizational principles that seem to have guided the manufacture of ceramics in this case study are compatible with some of Sabo’s structural principles. The design process appears hierarchical and frequently involves “either-or” or oppositional choices as part of the trajectory toward completion. Dimensionality in arranging design elements, and in the case of fused vessels the complex components of the vessel whole, is another operating principle. Design fields are laid out and filled with attention to both vertical and horizontal orientations. In the finished products, color and texture are used to enhance the oppositional effect of heavily scribed areas against polished backgrounds. This case study of Friendship Engraved variety Freeman bowls indicates that structural principles underpinning elements of historic Hasinai Caddo society may have deep roots in Caddo history and may have been shared among widely scattered pre-contact Caddo communities. It would be useful to examine additional groups of ceramics, and different vessel categories, in collections from different drainage basins, to see if the principles applicable to this one small sample are evident across the larger Caddo world.

46

Early

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

3. At the House of the Priest Faunal Remains from the Crenshaw Site (3MI6), Southwest Arkansas h. edwin jackson, susan l . sco tt, and frank f. schambach

Copyright © 2012. Nebraska Paperback. All rights reserved.

Introduction Zooarchaeologists have become more sensitive to the effects of ritual, status differences, and feasting in producing faunal assemblages that diverge from more ordinary cultural patterns. For instance, at Cahokia, in Illinois, analyses of the Sub-Mound 51 borrow pit contents have provided a detailed look at the residues of ceremonial feasting and ritual activities on the main plaza in the early occupation of the site (Kelly 2001; Pauketat et al. 2002). The general pattern in the American Bottoms indicates greater access to meat and in particular to choice cuts of venison by members of the elite stratum of society (Kelly 2001). Jackson and Scott (2003) examined the faunal remains associated with elite dwellings atop two mounds at Moundville, in Alabama. The assemblages demonstrated the proprietary access that elites had to certain faunal resources, including exotic taxa as well as symbolically charged species, producing a significantly more varied meat diet than others enjoyed in Moundville society. Moundville elite faunal use corresponds well with expectations for identifying “luxury” foods available to the elite (van der Veen 2003). Researchers working in Mesoamerica (e.g., Emery 2003b) share a parallel interest in the differential access to faunal resources by the elite. Another trend in zooarchaeological research, as well as in ethnographic and ethnoarchaeological research, is consideration of the importance and multidimensional quality of feasting in a broad range of prehistoric, historic, and modern societies. The effects of feasting can be seen not only in food remains but in other aspects of the material cultural assemblages (e.g., Dietler and Hayden 2001; Weissner and Schiefenhovel 1996). 47

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Our interest in how faunal assemblages might reflect social differences, feasting contexts, and ritual aspects of Mississippian societies includes how political economic relations result in particular assemblage characteristics (Jackson and Scott 1995b; Scott 1983), how fauna may be reflective of the system of symbols employed in defining rank differences, and the relations of the elite with cosmological understandings (Jackson and Scott 1995a). The unique characteristics of the fauna from the floor and refuse fan of an unusual building at the Crenshaw site, arguably the domicile of an individual with both elite social status and religious responsibilities, are a case in point. Although the faunal assemblage from this building has been published (Scott and Jackson 1998), it is worthwhile to reexamine it in light of the more recent considerations of feasting, status, and ritual, and of recent discoveries at Crenshaw. The unusually well preserved faunal assemblage from the floor of the building provides a sense of the ways in which patterns of faunal procurement and consumption transcend economic considerations, incorporating individual social and political status, their implications for preferential access, and the role of symbolism in creating, maintaining, and demonstrating the supernatural relationships that both legitimize and support the status quo. The ultimate result is a faunal assemblage that is a considerable departure from what might be characterized as “average” Caddo dietary patterns. Factors Conditioning Elite Faunal Assemblages

Copyright © 2012. Nebraska Paperback. All rights reserved.

Elsewhere Jackson and Scott (1995a) have laid out what they believe to be the factors conditioning the character of elite faunal assemblages in the aboriginal Southeastern United States. Not only do economic relations, expressed as occasional tribute and in some cases possibly regular provisioning, play a role, but there is also an important underlying symbolic system that conditions the range of taxa that comprise chiefly fare. The assumption that the zooarchaeological record is affected by, and may thus reflect, these cultural beliefs and practices is predicated on the cross-cultural generality that food is a powerful and malleable source of symbols (Douglas 1971, 1982; Fieldhouse 1986; Goody 1982). What is included in a meal and how it is prepared may mark differences in ethnicity, social class, and ritual purity. Feasting is an integral part of many public rituals and is a complex social phenomenon providing an opportunity to communicate the structure of the collective whole rep48

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

resented by the participants, the specific goals of the participants, and the subtle or not-so-subtle social distinctions that exist within the group (e.g., Hayden 2001). Meat has powerful symbolic potential, particularly in societies where meat is provided by hunting, but also in societies that raise livestock (Emery 2003a; Hayden 2003). Susan Kent’s 1989 cross-cultural examination of perceptions of meat indicates that it is the most highly valued food, regardless of its actual quantitative contribution to the diet. Hunted game is valued over meat produced by collecting activities such as shellfishing and fishing. Particularly in societies lacking domesticated livestock, ideological systems define a close relationship between humans and animals, perceiving both to occupy a conceptual supra-category of “intellectual beings.” Beliefs and myths often allude to the descent of humans from animal ancestors, and attribute a host of anthropocentric characteristics to animals, including powers or qualities to which humans aspire. Animals are often responsible in myths for defining or illustrating cosmic relationships. Thus the closeness between humans and animals, disclosed in myths and demonstrated in the rituals surrounding hunting, processing, and consuming game, give ample support to the idea of the powerful symbolic potential of meat consumption. A defining characteristic of emerging complexity is differential access to resources (Arnold 1996). However, the source and extent of the authority of one segment of a society to appropriate labor or other resources is variable. As Helms (1992) points out, in ethnographically documented chiefdoms there is an intimate dependence of chiefly authority on religious sanctification. It is the religious charter of the society that provides the chief his authority, and it is his closer link to the supernatural that enables and justifies his role as the decision maker for the society. The chief employs a host of symbols demonstrating this supernatural connection as a means of legitimizing his position in society, and among these we can expect food, and in particular meat consumption and other uses of animal products, to play a role (Jackson and Scott 1995a:105; see also Douglas 1971, 1982; Fieldhouse 1986; Goody 1982). Thus it may be the symbolic “load” of certain taxa that ultimately determines the restriction of their consumption to the elite. If access is distributed unequally among those of differing rank, so too are ritual responsibilities necessary for the welfare of the society. At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

49

Chiefs, based on their closer relationship with supernatural forces, can be responsible for maintaining the well-being of society at large through performance of required rituals. While hints of the use of particular animals in these rituals have been noted for Mississippian societies, for instance the wings of swans at Cahokia (e.g., Pauketat et al. 2002), or squirrels at Lake Providence Mounds in northeast Louisiana (Scott 2005), clear evidence of particular categories of animals for use in offerings can be found in the Mayan zooarchaeological record (e.g., Carr 1996; Emery 2003a; Pohl 1983). It is likely that with increased attention, the dimensions of Caddo and Mississippian ritual animal use will be forthcoming. Thus the success of the chief and the loyalty of his/her followers may be closely tied to the welfare of the society at large. Chiefly authority, particularly in the initial stages of social complexity, may be precariously based on an ongoing redefinition of the relationship between chief and follower (Barker 1992:62). Full storehouses, plentiful game, and sumptuous feasts all attest to the ability of the chief to mobilize supernatural forces for the continued well-being of the society. The Crenshaw Site

Copyright © 2012. Nebraska Paperback. All rights reserved.

Crenshaw (3MI6) is a large, complex, late Fourche Maline and early Caddo site located in the Great Bend region of the Red River valley in southwestern Arkansas (figs. 3-1 and 3-2). It has a long history of scientific excavation, beginning with the work of C. B. Moore in 1912, and an equally long history of looting by relic hunters (Davis 1962; Dickinson 1936; Durham 1976; Durham and Davis 1975; Durham and Kizzia 1964; Lemley 1936; Moore 1912; Powell 1977; Schambach 1972,; 1982c, 1996, 1997; Wood 1963a, 1963b). The evidence accumulated since 1912 indicates that the occupation of Crenshaw began between A.D. 600 and 700 with the development of a Fourche Maline village, burial mound, and cemetery complex. By around A.D. 900, Crenshaw was a sizable Fourche Maline village with at least four and possibly six or more burial mounds. Soon after that, most of the village population apparently dispersed into the countryside— the extensive midden areas at Crenshaw appear to be almost entirely Fourche Maline as opposed to Early Caddo—establishing the dispersed farmstead and nearly vacant ceremonial center settlement pattern that seems to have prevailed among the Caddo of the Great Bend region 50

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

3-1. Location of the Crenshaw site. (From Scott and Jackson [1998:fig. 1] in Arkansas Archaeologist 37, courtesy of the Arkansas Archeological Society.)

down to the historic period. However, for centuries after the dispersal occurred, the population continued to bury at least some of their dead in the cemeteries and mounds at Crenshaw (seven have been identified so far) and in the mounds, and perhaps even constructing Mounds A and E, neither of which has been excavated or even significantly disturbed by looters (Schambach 1982c:150, 1996a, 2002:111–112). An evidently small population of individuals whose specialty was ritual activity lived at the site during the Early Caddo period, perhaps At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

51

3-2. The Crenshaw site. (Map by Michael Evans, Jared Pebworth, and David Jeane of the Arkansas Archeological Survey, assisted by Jeff Girard and Joe Saunders of the Louisiana Division of Archaeology. © 2006.)

Copyright © 2012. Nebraska Paperback. All rights reserved.

year-round. For the most part, it would seem, the buildings in which these people lived and practiced were not on the mounds, the Caddo pattern later on. There is no reported evidence of buildings, neither “temples” nor domiciles, from the four mounds that have been completely or nearly completely excavated (Mounds B, C, D, and F), not even in the large, flat-topped Mound C (Wood 1963b:4). But such evidence could be still concealed in Mounds A and E. The “House of the Priest” The one Early Caddo special-purpose building of this type found thus far is the “House of the Priest,” the subject of this chapter. It stood close to the southern edge of the site (fig. 3-3, the “house floor”; see also fig. 3-2) overlooking what is now a relict oxbow lake called “Second Old River Cutoff.” Three ams dates (Beta-284671-3) on deer bone from the 52

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

3-3. Simplified plan view of the 1969 and 1983 Arkansas Archeological Survey Excavations at Crenshaw, showing the locations of feature 6 (the “house floor”), feature 1 (the “antler pile”), and the skull and mandible burial area (the “West Skull Area”). (From Schambach [1996:fig. 5.5], courtesy of Arkansas Archeological Survey. See Schambach for a detailed plan of the 1969 excavations and photos of the antler pile and skull and mandible clusters from the West Skull Area.)

floor of this house were obtained from Beta Analytic Inc. in 2010 and analyzed and calibrated using the calib 6.0.2 program and the IntCal09 curve (Reimer et al. 2009; Stuiver and Reimer 1993). These indicate a two sigma range for occupancy between A.D. 1161–1254. Thus this house would have been in use near the end of the Lost Prairie phase, the earliest phase of Caddo culture recognized in the Great Bend region of the Red River Valley, or at the beginning of the Haley phase (Weinstein et al. 2003:13 and fig. 5). At that time, Second Old River was probably either a recently formed oxbow lake or a segment of the active channel of the Red River. Significantly, for reasons that will shortly become apparent, this house also stood close to the southeastern edge of a large, special-purpose cemetery in which the severed full-fleshed heads (with mandibles) At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

53

Copyright © 2012. Nebraska Paperback. All rights reserved.

of many people, and the detached mandibles of still others, were buried in small, shallow holes dug in a now relict land surface, all without offerings (Schambach 1996). In 1983, salvage excavations by Schambach, Ann Early, and David Jeane within what appears to be only a part of this cemetery (the “West Skull Area” shown in fig. 3-3) uncovered the skulls, all with mandibles in articulation, of 127 people, some buried singly, others in groups of up to 11. The mandibles of 237 more people were also found. Some were buried singly, but most were in clusters, with the three largest containing 108, 60 or so, and 38 mandibles, respectively. These human remains are currently the subject of a Native American Graves Protection and Repatriation Act–supported study by the Arkansas Archeological Survey, a survey project sanctioned by the Caddo Nation of Oklahoma. Both the “House of the Priest” and the skull and mandible cemetery to the west and north of it occupied the lowest ground on the site (see fig. 3-2). This low ground had not been used by the Fourche Maline occupants, who kept to the higher elevations to the north and east. A stratigraphic test trench dug during the 1983 excavations demonstrated that the “House of the Priest” stood on the same relict land surface in which the skulls and mandible were buried (see fig. 3-3). Eventually the site of the “House of the Priest” and the skull and mandible cemetery were covered—and to some extent preserved—by a continuous stratum of grayish gumbo clay ranging from about 20 cm to 80 cm or more in thickness. This clay is the result of sedimentation in a slack water lake or pond that inundated these features sometime after A.D. 1200 as the “Great Raft” formed in the Red River valley, plugging the river bed with fallen timber and creating vast “raft lakes” that were still in existence in the nineteenth century when the raft was finally removed (Bowman 1911; McCall 1983). The remains of the “House of the Priest” were excavated by Schambach in 1969, with the support of National Science Foundation Grant GS-2684. These consisted of two adjacent features (see fig. 3-3): feature 1, a large deposit of antlers, and feature 6, a probable house floor marked by a large, more or less rectangular deposit of ash, replete with artifacts and exceptionally well-preserved bone (Schambach 1972, 1996). Feature 1, which abutted the ash deposit on its southwest corner, was 4–1 m long, north-south, 3–8 m wide, and 35 to 40 cm deep at its thickest part, near the middle. It was originally rectangular, judging 54

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

from the fairly straight west and north edges (Schambach 1996:fig. 5-4; Scott and Jackson 1998:fig. 3). The irregular south and east edges were the result of potholes dug by the landowner out of curiosity, after his plowing and disking began turning up masses of decaying antler. This feature consisted of the much decayed and compacted remains—mostly the proximal ends—of 2,042 antlers that had been placed in pairs, many with pieces of skull attached. These represent a minimum of 1,021 male white-tailed deer of all sizes; there was no apparent selection for trophysized racks. Schambach estimates that about one-third of the antlers deposited originally had been lost to decay, plowing, disking, and random shoveling by the landowner. The antlers had been piled in a shallow (10 cm deep) depression in the same relict land surface in which the skull and mandible burials to the west of this feature had been buried. Subsequently the antlers—and feature 6 to the north of them—were covered with a 20 to 40 cm thick stratum of nearly sterile sand. This deposit was thickest around the outside edges of the antler deposit and only 10 to 20 cm thick over the top of it and over feature 6. At the time of Schambach’s 1969 excavation, the highest part of the pile, near the center, had been damaged by plowing and disking. There were plow scars across parts of feature 6 as well. The same slack water gumbo clay deposit that covered the skull and mandible burials to the west covered this sand stratum, where it was confined to the plow zone and the first few centimeters beneath it. Since the sand stratum was confined to the area over and around features 1 and 6, it is apparent that it was put down by the Caddo to cover these features. This is the earliest recorded example of what would become the common Caddo practice of covering the sites of important buildings with mounds. Although this low mound and the overlying slack water gumbo deposit had mostly preserved both features from plowing, what really accounts for their survival down to 1969 is the fact that this part of the Crenshaw site was not cleared for the first time until 1965, and perhaps also the fact that both features were still partly covered by a substantial pile of bulldozed trees, tree roots, and earth when excavation began in 1969. They had more or less survived the single “break” plowing and the several subsequent diskings they had received, but it is not likely that they would have survived much longer, at least not reasonably intact. At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

55

Copyright © 2012. Nebraska Paperback. All rights reserved.

Feature 6, the ash deposit, was 9.5 m long, up to 8.5 m wide, and an average of 30 cm thick. No post holes were identified along the perimeters of this deposit, despite careful searching. Nevertheless, Schambach interprets it as the ash-laden floor of a large, rectangular house, his considerations being its size, shape, and orientation, the pattern of refuse distribution around it, and its distinctive contents. Except for a puzzling bulge on the west side, it resembles the post hole patterns of two buildings (Houses 1 and 3) that appear to be associated with an Early Caddo component of similar age at the Sam Kaufman (or Roitsch) site (41RR16) farther up the river in Red River County, Texas (Skinner et al. 1969:16–21). The precise north-south orientation is indicative of a building rather than a dump of some kind. There was a pronounced patterning of debris around the perimeters of this feature of the sort that would be expected around a building. Dark-stained midden rich in artifacts, food refuse bone, and other habitation debris was concentrated along the east and southeast edges, suggesting an outdoor activity area, while the western and northern perimeters were devoid of refuse and soil staining. Although the abundant domestic refuse from feature 6 is essentially identical ceramically and in the range of ordinary materials represented with the refuse constituting the midden to the east and south, feature 6 also yielded an assemblage of extraordinary materials, described below, which, with one significant exception, were not found in the midden. Schambach hypothesizes that the absence of post holes around feature 6 is more apparent than real and that it is attributable to the slack water gumbo clay deposit that eventually covered it. Soils scientists working with similar deposits at Cahokia (Young and Fowler 2000:224– 225) have learned that the sandy alluvial soils beneath these slack water gumbo clay deposits have been rendered virtually colorless by centuries of percolation. Such soils appear to be featureless, due to the loss of soil color, yet probably do contain evidence of post holes and pits detectable by remote sensing electronic technologies such as magnetometry, gradiometry, and resistivity that do not rely on soil coloration. Unfortunately, a recent attempt to evaluate this proposition with remote sensing at Crenshaw was thwarted by modern trash that has accumulated in the spot in the decades since the excavations described here (Samuelsen 2009). One other puzzling attribute of feature 6 is that although it appears to 56

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

have been a house floor defined by an extraordinary deposit of ash, there was no evidence of a hearth inside it. This could be accounted for by the historically documented practice of the western, Texas branch of the Caddo, the Hasinai, who heaped “glowing coals” under their beds during cold weather (Griffith 1954:102). There is also the Hasinai “medicine man’s” practice of piling hot coals under a patient’s bed as part of their cure (Swanton 1942:220). Given the elaborate ritual interest in fire recorded for the historic Caddo, it is conceivable that their ancestors had types of buildings in which live fire was not allowed, while hot coals for warmth were permitted (Griffith 1954:132–133). The nondomestic materials from feature 6 confirm that, as the antler deposit and the nearby skull and mandible cemetery indicate, it was the locus of ceremonial as well as mundane activity—a priest’s house, if you will. Most striking were the human remains scattered throughout the feature. These consisted primarily of 73 unmodified adult teeth and several skull fragments. There were no postcranial bones. Several of the teeth were imbedded in small pieces of jaw bone, so they, at least, had not been extracted from, or lost by, living people. At least 10 adults are represented by the loose teeth, based on the evidence of 10 maxillary first molars, and only 13 teeth were mandibular (Powell 1977). Thus it would appear that most of the teeth were from fleshless, somewhat desiccated, and perhaps mostly jawless skulls that were kept and handled in this building. If the mandibles were really missing, it is probable that they were deposited in the skull and mandible cemetery to the west and north of this feature. From evidence of six left maxillary canines from plow damaged and scattered remains found in and just beneath the plow zone in the vicinity of the southern and eastern edges of the antler deposit (Schambach 1996:fig. 5-3), at least six additional human skulls were buried in the low sand mound that covered both features. One of them was found above the eastern edge of the antler deposit. Of the 35 teeth collected from this locus, only four are mandibular, an “unexpectedly low proportion” (Powell 1977:112). It is probable that these skulls, all of which also lacked mandibles according to Schambach’s field observations, were placed in the mound covering the remains of the “House of the Priest” while its location was still known. The exotic artifacts from feature 6 consisted mostly of small numbers of freshwater pearl beads (not obtainable from the Red River, which At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

57

is too salty for freshwater shellfish), small copper beads and bangles, barrel-shaped marine shell beads made from the whorls of whelk shells, and small, well-made, polished bone pins. Schambach thinks that all of these artifacts were elements of the elaborate male headdresses often depicted in Mississippian shell engravings (Phillips and Brown 1978). If so, they could have been lost from the headdresses of elaborately coiffured skulls handled and kept in and around this building, whether those of ancestors or enemies. Or they could have come from the headdresses of the handlers themselves. The Cultural Context of the Excavated Sample

Copyright © 2012. Nebraska Paperback. All rights reserved.

The contents of features 1 and 6 indicate the presence of an important religious specialist whose duties probably included curating human heads and processing them for display or burial—it stands to reason that some, at least, of the many heads and mandibles buried in the large cemetery to the west and north of this building were processed though it—and the performance of rituals pertinent to hunting whitetail deer. Rituals involving human heads and others involving deer and deer antlers were observed by Spaniards traveling among the Hasinai of East Texas during the late seventeenth and early eighteenth centuries. It is likely that these were the lineal descendants of the rituals performed at Crenshaw, and surely other early Caddo ceremonial centers as well, during the twelfth century. However, Riley’s (2005:15–16) observation that early Spanish observers in the Southwest “misunderstood most or all of what they saw in Pueblo ceremonials” is a reminder that we should not put too much credence in their interpretations of these rituals. References to the taking, temporary curation, and disposal of human heads are found in Spanish accounts of the Hasinai Caddo. “Human heads” said to have been “taken by the Hasinai in war” were hung in a tree outside the door to the house of the Gran Xinesi (the paramount chief of the Hasinai) and later transferred, with fanfare, to an ash heap that accumulated outside the door of the Hasinai “fire temple” (Griffith 1954:132–133). Obviously, there are similarities here to what might have been going on at Crenshaw some 500 years earlier. Furthermore, there would seem to be an obvious conceptual relationship between taking trophy heads in war and taking deer heads and antlers in the hunt. Such a relationship did exist in Pueblo culture (Parsons 1939:134; Riley 2005:95). Nonetheless, pending the results of the Arkansas Archeologi58

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

cal Survey’s study of the large population of human heads and mandibles from this site, and perhaps not even then, we do not know that the heads manipulated, and perhaps stored temporarily, in the House of the Priest were “trophies.” They could be the remains of revered ancestors (Schambach 1996:40), or some combination of ancestral relics and trophies. All we can say is that around A.D. 1200 the Red River Caddo at Crenshaw, who were among the probable ancestors of the Kadohadacho of the historic era, may have had a burial program that was archaeologically and ethnographically unique in the Southeast; or they may have been engaged in the taking of trophy heads and mandibles on a scale that would seem to exceed anything now known archaeologically or ethnographically for the Southeast. Whatever they were doing, they probably were not doing it in isolation from the rest of the Southeast. Brown and Dye (2007:278) note that “severed trophy heads play a prominent role in Mississippian art” and they consider “the severed agnathous humanlike head of the Braden style . . . perhaps the distinctive iconographic trophy motif ” in Mississippian art. Thus the handling and, possibly, display of apparently coiffured and possibly agnathous human skulls in the “House of the Priest” would appear to be evidence of familiarity with, and some degree of participation in, the ideational system of the Mississippian world. The copper beads and bangles and the whelk shell beads are more tangible evidence of participation in the Mississippian interaction sphere by whoever owned them. A large, perfect cube of galena from the associated midden southeast of feature 6 is additional evidence of wider contacts. Assuming it is from the Potosi deposits of southeast Missouri, as is the case with much of the galena from Mississippian sites (Walthall 1981), this specimen bespeaks contacts with Cahokia, perhaps directly or perhaps indirectly via Spiro phase Spiroan traders based in the Arkansas River valley (Schambach 2000b:8–10). Clearly, the ranking resident of this house had considerable standing locally, and apparently in the wider Mississippian World as well, as did his or her approximate contemporaries at the Gahagan site farther south in the Red River Valley in Louisiana (Brown 1983:fig. 2 and table 3; Emerson and Hughes 2000:83; Emerson and Girard 2004). There are a few indications that whatever was going on with the human heads and mandibles at Crenshaw also went on at other important Red River Caddo sites. More than 50 years ago Harris (1953:63) reported At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

59

a pit full of human skulls at the Sam Kaufman site: “Just east of Salt Well Slough, and about the middle of the village area, a pit was found. This pit contained about 20 skulls, but no other bone of the bodies. There were male, female, and small children’s skulls in the pit.” More recently, Perino (1983:7) noted that “a group of about ten trophy skulls were reported found many years ago by the farmer” at the Bob Williams site and commented that: “Today one can still find teeth and skull fragments on the slope . . . on the northeastern corner of the terrace along Salt Well Slough.” The Sam Kaufman and Bob Williams sites are part of a large Caddo site complex, which also includes the Roden site, about 16 km downstream from the confluence of the Kiamichi River with the Red River (Perino 1983:1), so it is possible that the Harris and Perino reports are descriptions of the same find. On the other hand, the Kaufman skull pit was found in “about the middle of the village area,” according to Harris, while the skull fragments and teeth observed by Perino were eroding out of the slope of the terrace along Salt Well Slough. This suggests two deposits. A few reports of supernumerary mandibles in graves at later Caddo sites in the Great Bend region indicate that the collection and curation of mandibles was not confined to Crenshaw or to the Lost Prairie phase. In Burial 2, a Haley phase burial at the Haley site, Moore (1912:532) found:

Copyright © 2012. Nebraska Paperback. All rights reserved.

a human lower jaw having an end . . . of [a] ceremonial axe . . . lying upon it. This jaw, which is in an excellent state of preservation, far better than were the bones of the skeleton, had no connection anatomically with the bones of the skeleton, which had a mandible in place. The detached jaw bears numerous scratches and marks of scraping, which seemingly indicate that the flesh was removed . . . with the aid of some implement preparatory to preservation. The mandible, moreover, is smooth, even almost polished in places as if it had been in possession of its secondary owner for a considerable period of time. Its excellent state of preservation calls to mind that of many bone tools which have thoroughly hardened before inhumation, and whose state of preservation usually excels that of the skeleton with which the tools are found. This human jaw was evidently a valued possession of the occupant of the grave, when alive, and, found under these conditions, it opens a wide field for speculation. Moore (1912:567) also found in one of the four low mounds east of the big mound at the Battle site (3LA1) a burial which consisted of “the skull of an adult with the lower jaw, and in the neighborhood were 60

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

two other mandibles.” (And two of the five burials in the same mound contained supernumerary skulls.) It looks as if another supernumerary mandible was found during Lemley’s (1938) unpublished but well documented excavations at the Battle site. The drawings of Lemley Burial B-1, Plot 1, show an apparently plow-disturbed burial with the skull and mandible represented only by one skull fragment and some loose teeth. However, the drawing also clearly shows a supernumerary set of teeth, apparently the remains of a mandible, lying just to the left of the lower left leg. There were six pottery vessels with this burial. As for the antler deposit, documentation on the Kadohadacho who occupied the Great Bend region in historic times (Williams 1964) is so sparse that it is fair to say we do not know whether they were practicing any kind of deer ceremonialism when sustained European contact began in the late seventeenth and early eighteenth centuries. But the betterdocumented Hasinai were, according to Espinosa (Swanton 1942:136), who recorded rituals that involved antlers and a building: Before they go out to hunt deer, they put on a post in their thatched hut the dry head of a deer including the neck and horns, while they pray to their caddi ayo that he will put the prey into their hands, while at intervals they throw pinches of the tobacco that has been provided into the fire. When they have performed this ceremony—which lasts more than an hour—they put its head at the door of the hut and with another just like it they go out in the woods to hunt, covering their naked bodies with white dirt. When they have killed a deer, they divide the animal. For some time they talk into its ear but I do not know the meaning of this puzzle. They load themselves with it for the return trip. They throw it down at the door and the cooks cut it up. They take pains to see that the one who killed it does not eat of it unless the others invite him and that he does not take anything else to satisfy his hunger. The Faunal Collection Three-fourths of the faunal collection described here derives from feature 6, with smaller samples of bone from the five 2 x 2 m units excavated into the midden fan east of features 1 and 6, and a still smaller, and much more limited, sample from a small “control pit” square opened in the top of feature 1 in the process of relocating it at the beginning of the 1969 excavations. While it is tempting to dichotomize the structure and midden fan samples, in reality they are likely closely interrelated. Refuse accumulating in the structure may ultimately have been discarded in the At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

61

Copyright © 2012. Nebraska Paperback. All rights reserved.

midden. However, part of the midden refuse may reflect in situ activities in that location: different stages of butchering are represented by the midden and structure samples, implying spatially segregated but interrelated steps in the process. What is clear is that there are differences in the content of these two deposits. Some of these differences are taphonomic in nature; others, we will argue, are a result of the sacred space created by the structure on the one hand and on the other, are a result of partial processing for public observation, and possibly public consumption, outside the structure. As luck would have it, Stuart Struever’s 1968 “Flotation Techniques for the Recovery of Small-Scale Archaeological Remains” was published the year before Schambach excavated the “House of the Priest” and its environs, and—as detailed in his NSF proposal—he went to the field fully equipped and prepared to use those techniques at Crenshaw. So the excavation of feature 6 was state of the art for 1969. Basically, all of feature 6 was first troweled to record major objects in situ, then dryscreened through 1/4-inch mesh, then water screened through 1/8-inch mesh and floated. The floated residue and the residual heavy fraction from the 1/8-inch screening were saved for hand sorting in the lab. More than 21,000 bones were identified for this analysis, not including the largely unquantified contents of the antler deposit. Because unidentifiable bone was weighed rather than counted, it is impossible to give a precise number for all the fragments sorted, but the sample from the 1/8-inch screening at the site probably yielded in excess of 30,000 fragments. The majority of the analyzed bone, more than 14,000 fragments representing at least 53 species, came from feature 6, the structure floor. Fine screen (1/8-inch) recovery is estimated to have yielded an additional 35,000 bones based on extrapolation from several small samples of weighed and counted fragments. Bone from feature 6 was remarkably well preserved and no doubt was made even more durable by the fine-grained alkaline ash in association. More than 4,700 identifiable bone fragments were recovered from the refuse fan and from other small features in and around feature 6, adding another six taxa to the species list. In addition to the samples presented here are another 71 buck frontals from whitetail deer, these recovered from feature 6, from the initial “control pit” dug into feature 1, and from the refuse fan. Data from each of these three contexts are presented in table 3-1. Taphonomic issues are considered elsewhere in some depth (Scott 62

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Common Name Opossum Least Shrew Eastern Mole Eastern Cottontail Swamp Rabbit Unid. Rabbit Pine Vole Cotton Rat Rice Rat Wood Rat Hispid Pocket Mouse Rat Mouse Rat/Mouse Pocket Gopher Eastern Gray Squirrel Eastern Fox Squirrel Squirrel sp. Beaver Raccoon Mink Striped skunk Cougar

Scientific Name

Didelphis virginiana Cryptotis parva Scalopus aquaticus Sylvilagus floridanus Sylvilagus aquaticus Sylvilagus sp. Microtus pinetorum Sigmodon hispidus Oryzomys palustris Neotoma floridana Perognathus hispidus Cricetidae Cricetidae Cricetidae Geomys bursarius Sciurus carolinensis Sciurus niger Sciurus sp. Castor canadensis Procyon lotor Mustela vison Mephitis mephitis Felis concolor 4 20 13 102 305 321 21 22 3 2 1 32 90 21 51 137 83 254 1 1 1 1 1

6.4 0.1 1.0 43.8 217.4 102.0 0.4 0.8 0.1 1.5 0.1 1.5 1.2 3 .2 9.8 37.4 33.4 26.1 0.3 2.4 0.3 0.3 10.2

Feature 6 Count Weight

Table 3-1. Counts and weights of identified taxa from Caddo I contexts

Copyright © 2012. Nebraska Paperback. All rights reserved.

1.1

0.3

1

1

Feature 1 Count Weight

0.2

2.1 5.8 2.4 0.5 10.6

6.1

7 10 10 1 4

2

28.6 60.4 10.1

6.7

2

54 62 37

11

(continued)

Midden Count Weight

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Black Bear Unid Large Carnivore Whitetail Deer (Antler) Very large Mammal Large Mammal Small/Med. Mammal Bird/Small Mammal Canada Goose Mallard/Black Duck Duck Turkey Long Eared Owl Woodpecker Yellow-Billed Cuckoo Cuckoo Blue Jay Bobwhite Yellow Rail cf. Grackle cf. Red Wing Blackbird Blackbird Family Mockingbird Family Passenger Pigeon

Ursus americanus Carnivora Odocoileus virginianus

Branta canadensis Anas platyrhynchos/rubripes Anatidae Meleagris gallopavo Asio otus Picidae Coccyzus erythropthalmus Coccyzus sp. Cyanocitta cristata Colinus virginianus Coturnicops noveboracensis cf. Quiscalus sp. cf. Agelaius phoeniceus Icteridae Mimidae Ectopistes migratorius

Common Name

Scientific Name

3167.2 72.7 185.1

4440 457 1079 1.5 0.3 275.6 1.4 0.1 0.1 0.1 0.4 0.1 0.1 0.1 0.1 0.1 0.1 17.3

6530.1 2284.8

1168 4563

2 1 87 1 2 1 1 4 1 1 1 1 1 2 70

1.6

1

Feature 6 Count Weight

Table 3-1. Counts and weights of identified taxa from Caddo I contexts (continued)

Copyright © 2012. Nebraska Paperback. All rights reserved.

2

91 147 1 66 327

8.6

906.1 1351.1 3.4 129.9 158.5

Feature 1 Count Weight

200.1

1.7

5

12.2

1135.8

12 2.6 8324.3 673.1

68

1

2500

3 1 1012 266

(continued)

Midden Count Weight

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Unid. Medium Songbird Unid. Small Songbird Unid. Large Bird Unid. Med/Sm Bird Snapping Turtle Alligator Snapping Turtle Pond Turtle Map Turtle Pond/Map Box Turtle Mud Turtle Mud/Musk Turtle Softshell Turtle Unid. Turtle Viper Rattlesnake King/Rat Snake cf. Milk Snake Mud Snake Unid. Non-poisonous Snake Colubridae Frog Frog/Toad Amphibian

Passerine Passerine

Rana sp. Rana/Bufo sp.

Chelydra serpentina Macroclemys temmincki Pseudemys/Trachemys/Chrysemys Graptemys Emydidae (aquatic) Terrapene carolina Kinosternon subrubrum Kinosternidae Trionychidae Testudines Viperidae Crotalus horridus Elaphe/Lampropeltis cf. Lampropeltis triangulum Farancia abacura

Common Name

Scientific Name 0.1 0.1 102.8 6.3 3.5 0.6 0.8 69.3 22.0 2.0 2.9 6.4 51.5 0.6 1.2 0.1 0.2 0.1 0.2 0.5 0.9 1.0

6 1 164 54 6 1 1 36 35 2 10 7 212 1 6 1 2 2 6 1 9 1

Feature 6 Count Weight

Table 3-1. Counts and weights of identified taxa from Caddo I contexts (continued)

Copyright © 2012. Nebraska Paperback. All rights reserved.

2

5.2

Feature 1 Count Weight

1.1

18.7 1.8 1.8

27 2 1

2

0.3

2

15.2

6

93.1 4.7

69.2 15.7 0.9

91 56 1

55 2

0.1

1

(continued)

Midden Count Weight

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Bowfin Longnosed Gar Gar Scales Gizzard shad Smallmouth Buffalo Unid Buffalo Sucker Blue Catfish Channel Catfish Blue/Channel Catfish Brown/Yellow Bullhead Catfish Largemouth Bass Bass Crappie Sunfish Unid. Sunfish Freshwater Drum Spiny Ray Fish Unid Fish

Amia calva Lepisosteus osseus Lepisosteidae

Total Identified Specimens

Dorosoma cepedianum Ictiobus bubalus Ictiobus sp. Catostomidae Ictalurus furcatus Ictalurus punctatus I.furcatus/punctatus Ameiurus nebulosus/natalis Ictaluridae Micropterus salmoides Micropterus sp. Pomoxis sp. Lepomis sp. Centrarchidae Aplodinotus grunniens Perciformes Pisces

Common Name

Scientific Name

13402.8

8.4 2.4 0.1 1.3 6.1 5.3 2.6 9.0 11.0 0.6 0.9 0.4 7.0 0.5 0.1 1.4 22 0.7 22.5

58 20 1 3 5 15 1 1 16 3 5 2 10 4 1 14 35 1 114 14268

3.3

29

Feature 6 Count Weight

Table 3-1. Counts and weights of identified taxa from Caddo I contexts (continued)

Copyright © 2012. Nebraska Paperback. All rights reserved.

312

2404.4

Feature 1 Count Weight

1.3 0.4

3.0 0.7

2 3

10 1

10737.2

7.5

2

4418

1.1 0.2

1.8 1.2

1 1

1 1

Midden Count Weight

and Jackson 1998). It should be noted that the sources of the two largest samples of Caddo animal bone at Crenshaw are from contexts likely to differ significantly in composition due to culturally based taphonomic bias: a structure versus an open midden. Ethnoarchaeological studies (Gifford 1977, 1980; Gifford and Behrensmeyer 1977; Yellen 1977) and experimentation (Schiffer 1978) together have shown that dwellingfloor deposits are more likely to yield small bone fragments and bones from smaller taxa. This differential representation of smaller pieces is a consequence of the human propensity to overlook small items when removing refuse for discard, combined with an unusual geologic agent, human traffic. The culturally based taphonomic bias in fragment size is apparent in comparing “inside” versus “outside” bone assemblages in the Crenshaw sample. Fragments from the structure were, on average, significantly smaller than fragments from the adjacent midden deposits, exhibiting a mean weight of 0.94 g and 2.78 g, respectively (both samples were screened through 1/8-inch mesh). Not surprisingly, the proportional representation of small taxa in the structure was much higher, a point of some importance in assessing inter-site comparisons (below).

Copyright © 2012. Nebraska Paperback. All rights reserved.

Analysis The analysis of the Crenshaw bone examined patterns of selection, processing, and consumption represented in the domicile (feature 6), the midden and associated features, and to a lesser extent the control pit sample from feature 1. Of particular interest is how deer element distributions may be indicative of both processing activities and differential patterns of consumption. Also of interest are those taxa not normally considered to be part of Caddo fare, which may be indicators of the wider range of “special” foodstuffs available to the elite, foods for provisioning feasts, or animals needed for particular ritual uses. Both patterns of deer utilization and rare taxa are interpreted to reflect the specialized meat consumption and animal use patterns associated with the chief ’s house/antler deposit complex. As suggested earlier, these deposits are intricately interlinked through the cultural practices that created them; no simple dichotomy (e.g., commoner versus elite or public feasting versus private consumption) can confidently be relied upon. Nonetheless there are differences between the two contexts that are reflected by differences in faunal samples. That they together represent patterns of At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

67

meat consumption distinctly different from those represented at nonceremonial sites is best appreciated by comparison with other Caddo faunal samples, which is presented in a subsequent section. For economy of presentation, the data on which the following discussion is based have been omitted or pared down, but can be found in the original analysis (Scott and Jackson 1998). Animals Other than Deer

Copyright © 2012. Nebraska Paperback. All rights reserved.

We begin with taxa other than deer, of which 56 were identified. Nondeer mammalian cultural introductions include rabbits, squirrels, opossum, beaver, raccoon, striped skunk, mink, cougar, and bear. Aside from rabbits and squirrels, other mammals were quite rare in the assemblage, with none yielding more than 15 identifiable bones. Commensal mammals include a variety of rodents, moles, and voles, which exhibit a low incidence of charred bone (5.0 percent compared to 29.7 and 37.5 percent for squirrels and rabbits, respectively). All but three of the 246 rodent bones were recovered from the structure floor. These represent 2.4 percent of the mammal number of identifiable specimens (nisp) in this sample, compared with 0.05 percent of mammal nisp from the midden. At least 15 species of birds are represented, the most frequent being wild turkey (n = 157) and passenger pigeon (n=75). Twenty-two percent of the bird bones were charred (137 of 625). Interestingly, turkey bones are nearly evenly distributed between the structure and the refuse fan (56 percent vs. 44 percent), whereas passenger pigeon remains are quite clearly concentrated in the structure (70 of 75 or 93 percent). While turkey was a common food throughout the Mississippian world and is often associated with feasting as well as household refuse, we have suggested elsewhere that passenger pigeon was a species differentially accorded to Mississippian elite, based on its relatively higher frequency in elite contexts (Jackson and Scott 1995a). In a recent evaluation of the changing biogeographical range of passenger pigeon in the Southeast (Jackson 2005), there is evidence pointing to a southward and westward extension of their range during the Late Woodland period (ca. A.D. 700–1000) that continues into the Caddo era (ca. A.D. 1000–1500), likely reaching East Texas sometime shortly after A.D. 1000. Passenger pigeons do seem to occur in ritual contexts, particularly in the Caddo area, presumably because they were a novel food item and, when present, available in 68

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

large numbers as attested to by several very large deposits of their remains. Their exclusive association with elite contexts is, however, not so clear. Several of the other identified birds probably pertain to symbolic uses in ritual or ceremonial contexts, including woodpecker, cuckoo, blue jay, mockingbird or thrasher, blackbirds, and other unidentified small perching birds. All were part of the structure sample. Two species were recovered in the antler deposit (feature 1 control pit) sample, turkey and a long-eared owl, to be discussed more extensively below. Reptilian and amphibious taxa, comprising only 2 percent of the assemblage nisp, included seven species of turtles, most belonging to the family Emydidae. All except box turtle could have been easily captured in the Red River. Like rabbits and squirrels, 30 percent of the turtle bones were charred (111 of 366). Only 18 snake vertebrae were recovered, all but three from nonpoisonous species. Sixteen percent of these vertebrae were charred. Eleven amphibian remains (frog or toad) were identified, nearly half (n=5) charred and all from the structure floor. Except for bullfrogs, these represent very little meat, yet the high percentage of charring is intriguing. It is tempting to suggest ritual consumption, although another ritual use such as offerings might be responsible for the charring. Mayan ritual deposits often include reptiles or amphibians (Carr 1984). Fish remains contribute only 2.6 percent to the nisp and 0.5 percent by weight to the Early Caddo sample analyzed here. This is somewhat surprising given the proximity of the site to the Red River, although “expectable” patterns of procurement should not be expected to supply the sustenance of this (assumed) important personage. The structure produced the majority of fish remains (93 percent of the fish nisp), as would be expected by depositional circumstances. Modal fish length is between 20 and 40 cm standard length (SL), comprising 54 percent of fish Minimum Number of Individuals (mni). A few individuals exceed 60 cm SL (19 percent) up to approximately 90 cm SL in size, including gar and a very large blue catfish. Figure 3-4 shows a breakdown of the identified remains by family and standard length. Sucker, catfish, sunfish, and drum clearly predominate numerically. If a preferred species is present, it would be catfish, represented in every length category and producing 20 percent of the mni. Sunfish (basses, crappies, and small sunfish taxa such as bluegill) produced 24 percent of the mni, but never exceed 45 cm in standard length. At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

69

3-4. Size distribution of fish remains from Crenshaw, distributed according to family.

Whitetail Deer

Copyright © 2012. Nebraska Paperback. All rights reserved.

Age structure, seasonality evidence, and patterns of burning and fragmentation provide insights into the manner in which deer were procured, processed, and consumed by the Crenshaw site residents and visitors. First, age structure for the deer was estimated based on teeth and an assessment of epiphyseal fusion (Scott and Jackson 1998:15– 16). In comparison with age profiles presented in Elder’s now classic 1965 study of prehistoric and historic aboriginal deer assemblages, the Crenshaw site exhibits a population distribution that is well within the range of other prehistoric assemblages, suggesting that there was not excessive hunting pressure of deer, despite what feature 1 may lead one to expect. The proportion of “prime aged” (2.5–6.5 year old) deer is slightly lower than in the prehistoric assemblages analyzed by Elder, but this small difference can be accounted for. Under normal conditions younger and older individuals may have been ignored in favor of prime age individuals, but in the ceremonial hunts where the goal was to furnish sufficient meat for large gatherings, prey were probably taken on a first encounter basis. Although there is no information in this regard for the Caddo, at contact the Chickasaw are reported to have contributed the first prey killed to the chief, regardless of condition and quality (Moore 1988:34), and such rules may have applied at Crenshaw as well. A second observation about age determination of specific anatomical units, based on epiphyseal closure (a very rough indication of minimum age of an individual), is interesting as well. Elements representing the 70

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

forequarters are on the whole younger than those representing hindquarters. Moreover, forequarters represented by elements in the structure are from younger individuals than those recovered from the midden fan. We suspect that hindquarters were the most favored venison, since the greater amounts of fascia found in cuts from the forelimb render them less palatable. Those forelimb cuts that were consumed by the structure’s resident appear to have been from younger deer, possibly to compensate for the difference in palatability. Age data collected from maxillae and mandibles do not convincingly demonstrate any time-specific hunting pattern. The presence of a fawn less than a month old and at least two individuals between 13 and 17 months of age indicates warm weather hunting. The presence of at least three individuals aged 6 to 9 months indicates cold weather procurement. Likewise, antler at the site was found to exhibit both complete and incomplete ossification suggesting year-round harvesting. Most of the antler recovered was fully ossified, but cultural selection and taphonomic factors could easily explain their dominance, since they no doubt were more likely to preserve. Determining the customary cooking techniques for various taxa and parts thereof is almost impossible with archaeological remains alone. Cooking methods used by Southeastern tribes would not necessarily have left any unequivocal distinguishing marks on bone, and if they did, the marks might subsequently be obscured if incineration occurred during discard. Consequently, the two processes are considered together. We know from ethnohistoric accounts that Southeastern Indians boiled, roasted, and smoked meat, the latter process compensating for a lack of alternative methods of storing meat in the characteristically mild climate. Of the three methods, only the direct, close application of heat during roasting would leave distinctive marks on bone, and even then, charring would occur only if the bone were exposed. If roasting was among the customary methods used to cook a particular species, the charring present should be limited in extent to areas with minimal external tissue likely to be exposed to fire. In some cases, this charring might affect only a portion of an otherwise unburned fragment. Very few bones in the sample from the Crenshaw site correspond to this statement; a single distal scapula and proximal humerus in feature 6 were partially charred, probably indicating roasting of at least one complete upper forelimb. At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

71

Table 3-2. Charring of Early Caddo deer postcranial elements by body part Element Group

Number Burned

Percent Burned

NISP

72 208 32 24 75 13

11 26 9 6 35 46

653 813 340 425 212 28

Vertebrae Ribs, Costal Cartilage, Sternum Upper Forelimb Upper Hindlimb* Feet Patella *Including sacrum and pelvis, excluding patella.

Copyright © 2012. Nebraska Paperback. All rights reserved.

All other charred bone was thoroughly burned, possibly for disposal rather than consumption. However, it is possible to state that the deer bones most frequently charred are those of the distal extremities, which would have been exposed to fire if entire limbs were roasted. In feature 6 there is a relatively low incidence of charring overall, but of the burned bones, those exhibiting higher frequencies of charring include the distal ulna, distal tibia, and elements such as carpals, tarsals, metapodials, and phalanges. Interestingly, 13 of 28 patellae (46 percent) were burned. These small bones, embedded in connective tissue between the distal femur and proximal tibia, would have been charred relatively frequently if the hind limbs were roasted as an articulated unit. In combining all data on charring by major body part (table 3-2), the patellae, lower limbs and feet, and ribs show higher relative frequencies than upper limbs and vertebrae. According to ethnoarchaeological observations of the Alaskan Nunamiut, the vertebral column is usually boiled to more efficiently remove the meat adhering to those irregularly shaped bones (Binford 1978:145). The pattern observed here would seem to support a similar technique among the Early Caddo inhabitants of Crenshaw. These data also suggest that rib slabs and entire limbs were at least occasionally roasted rather than boiled. The relative degree of fragmentation of deer postcranial elements reflects the degree of bone processing reflected in an assemblage. For the Crenshaw sample, most of the larger marrow or bone-grease yielding bones in the assemblage are fractured, suggesting marrow removal or their eventual use in soups or stews. However, numerous articular ends are relatively complete, indicating that more intensive processing 72

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

for bone grease was not common. Such a pattern suggests that meat resources were plentiful, which should not be surprising given the implications of the antler deposit as well as its age structure.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Effects of Provisioning, Ceremonial Processing, and Curation on the Deer Sample Evidence that provisioning occurred at Crenshaw can be found in the near absence of bones of the lower limbs. Carpals, tarsals, metapodials, phalanges, and sesamoids are very rare in this assemblage (table 3-3). Figure 3-5 shows the relative anatomical representation of elements by food utility (Metcalfe and Jones 1988) in samples taken from the structure and the refuse fan outside it. Skull fragments, usually considered to be primary butchering refuse, are abundant, but no doubt are a consequence of skull preparation for display in the antler deposit. Consistently more than half of the skull remains occur inside the structure, whereas consistently more than half of the neck vertebrae are found in the midden outside. This pattern suggests that the head was removed outside, perhaps ceremonially, and then was taken inside the structure for further processing. There appears to be an important parallel between the preparation of deer skulls and the preparation of human skulls by the occupant(s) of the house. In the medium food utility group, three patterns are worthy of mention. First, the sacrum, although poorly represented in the sample overall, is better represented in the midden than the structure. Based on our analysis of butchering techniques presented elsewhere (Scott and Jackson 1998), the sacrum’s underrepresentation may be because excessive fragmentation rendered it unrecognizable. The greater representation of the sacrum in the midden outside the structure again suggests that butchering occurred there rather than inside. Second, lumbar vertebrae are very well represented both inside and outside the structure. Meat from the “tenderloin area” (lumbar vertebrae) may have been more highly prized than our current “objective” measures suggest. Finally, the relatively higher representation of forelimbs in the midden outside is notable. Not only are they better represented there than inside the structure, but as noted previously, more are from mature animals, a pattern possibly reflecting preferential access to the better (younger individuals) cuts by the priest. Finally, figure 3-5 clearly shows that anatomical portions of highest At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

73

Table 3-3. nisp, mne, and mni for whitetail deer from all analyzed Early Caddo deposits

Copyright © 2012. Nebraska Paperback. All rights reserved.

Element Antler Skull2 Maxilla Mandible Atlas Axis Cervical Vert Thoracic Vert Lumbar Vert Indeterm. Vertebrae Ribs Sternebrae Sacrum Scapula Humerus Radius Ulna Carpals Metacarpal Pelvis Femur Patella Tibia Tarsals Astragulus Calcaneum Metatarsal Phalanx 1 Phalanx 2 Phalanx 3 Phalanx 1/2 Lateral Phalanx Metapodial Carpal/Tarsal Caudal Vert. Unidentified Teeth Stylohyoid Sesamoid

NISP 4980 (3) 200 (10) 158 (3) 167 (1) 16 11 120 (11) 120 (17) 186 (25) 169 (19) 779 (208) 3 13 80 (5) 106 (7) 95 (9) 59 (11) 17 (6) 19 (7) 71 177 (9) 28 (13) 164 (15) 21 (5) 14 (3) 14 (2) 16 (1) 37 (16) 20 (8) 10 (5) 8 (5) 5 (5) 16 (5) 6 (2) 2 (1) 123 (10) 2 9

MNE

MNI

71 + 39 28 11 11 34 62 91

36 + 20 19 11 11 7 5 19

3 7 46 46 32 38

1 7 24 25 21 23

12 33 42 24 57

8 19 29 14 31

14 13 6 16 14 10

9 11 3 2 2 2

2

1

2 9

1 1/4-inch), and figures for McLelland, Cedar Grove, 78

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

3-8. Comparison of relative bone weight contributions by family to the Crenshaw and McLelland faunal assemblages.

and Taddlock include at least some micro-fauna, a 10 percent sample (by weight) of fine screen material from the structure floor was roughsorted to determine if fish could have been, in fact, a much more important resource than the macro-fauna at Crenshaw suggests. Extrapolation of this 10 percent sample of fine screen material from the structure indicates that all fish remains, including unidentifiable fish, from the structure floor could not have contributed more than 22 percent of the total nisp from the structure alone. Furthermore, turtle remains were almost nonexistent in the fine screen material (only seven fragments of unidentifiable turtle were found in the 10 percent sample), but bird and small mammal bone was, again, common. Since fish remains should be better represented in the well-preserved structure floor relative to the midden outside, we feel confident that such a low figure from this kind of depositional context clearly indicates differential consumption of nonaquatic taxa at the Crenshaw site. With the addition of the small sample of materials from the midden outside the structure, the percent nisp for fish would move back into a zone lower than any of the other four nonceremonial sites. With a larger sample from the midden, the differences presumably would be even greater. Given the depositional bias toward smaller taxa built into the structure sample, the argument that the massive quantity of deer in the Crenshaw refuse is unique for the area (regardless of differences in screen size) is quite clear. Obviously, considering the ceremonial significance of the antler deposit associated with this structure, any other result would be counterintuitive. However, Crenshaw is also unique in that so At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

79

Copyright © 2012. Nebraska Paperback. All rights reserved.

much of the assemblage is terrestrial fauna of all kinds. Rabbits, squirrels, turkey, and passenger pigeon all are exceptionally well represented in the assemblage. It is possible that the ceremonial significance of this structure largely (although not completely) precluded the introduction of animals from the “underworld,” a cosmological category distinct from “this world” (represented by four-footed terrestrial fauna) and the “upper world” (represented mainly by birds) in Southeastern Indian cosmology (Hudson 1976:122–132). It should also be kept in mind, however, that at the time the “House of the Priest” was in use, Crenshaw is believed to have been a “vacant” ceremonial center, permanently inhabited by very few people during most of the year, and that those fortunate enough to be permanent inhabitants were provisioned to a large extent. Under such circumstances, one would expect small game, particularly small mammals, to be abundant near the ceremonial center, both because hunting pressure within the immediate site catchment was relatively low (due to the local presence of only a few well-fed humans), and because food and cover requirements were better than what might be considered average for human-occupied areas (due at least in part to the probability that provisioning was occasionally excessive, and that some of the groups’ food stores were located on-site). With locally high populations of squirrels and rabbits, on those occasions that shortfalls in meat provisions did occur, this small game may have provided a substitute easily procured by the site inhabitants. Finally, with respect to provisioning patterns, although the relative frequencies of deer elements at other sites in the area have not been published in sufficient detail to provide a firm basis for comparison, qualitative differences are apparent. Byrd (1980:250–255) noted that the Early Caddo Hanna site produced a disproportionate number of fragments from the lower limbs and feet of deer. Clearly Hanna is too far afield to have fallen within the Crenshaw sphere of influence, but it nonetheless appears to be an example of a Caddo site on the production end of this system. A Late Caddo farmstead, the McLelland site (Weinand and Reitz 1994:178 and fig. 10-1), also appears to have produced at least enough carpals, tarsals, or metapodials to account for higher utility parts. Thus, by comparison, the poor representation at Crenshaw of such primary butchering debris does appear to be a consequence of provisioning the ceremonial center. 80

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Elite Midden Deposits: Crenshaw and Beyond Although the most abundant animals at Crenshaw appear to have been deer, rabbits, and squirrels, the sample includes more than 50 taxa. And although the aquatic component is relatively depauperate, the terrestrial component includes a number of taxa that, at least in our experience, do not usually occur in abundance in an “average” midden. Particularly notable inclusions are cougar and bear, a diverse and plentiful assortment of mice and rats (which we tentatively attribute to their commensal relationship predicated on abundant food stores within the structure), an abundant sample of passenger pigeon, and a number of birds more likely to be used as ritual paraphernalia than food. The unusual birds identified include a long-eared owl, yellow-billed cuckoo, blue jay, mockingbird, red-wing blackbird, grackle, and two smaller songbirds, one roughly sparrow-sized and the other about the size of a wren. Another possible reflection of the role of ritual in the production of this sample is one vertebra from a huge viper, buried in a small pit (feature 8) outside the structure proper. Although the presence of any one or even several of these taxa in other assemblages from the Caddo area is not unusual, the constellation of creatures associated with this deposit is unique for the area. It most closely resembles an assemblage from the mound at the Lubbub Creek archaeological locality in western Alabama (Scott 1983). The mound deposit at Lubbub produced a number of unusual birds including a merlin, passenger pigeon, Carolina parakeet, blue jay, cardinal, mockingbird, and crow. It also produced a bobcat, and a disproportionate share of the mouse and rat bones from the site as a whole. Given these similarities, the composition of a third “elite” assemblage was reviewed to see if the pattern held true for another site. The sample used for comparison was collected from structure 3 at the Toqua site in east Tennessee, analyzed by Arthur Bogan (1980). Structure 3 was interpreted by the excavators (Polhemus 1987) to be a special-purpose building, possibly a men’s house. Located adjacent to a mound and yielding a large sample of vertebrate remains, which were tabulated separately, structure 3 appears to share many of the traits deemed unique for Lubbub and Crenshaw. Animals dangerous to pursue are present (these include bear, cougar, and bobcat) along with a very large sample of passenger pigeon (more than three times the nisp for turkey in the At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

81

Table 3-5. Unusual taxa recovered from elite contexts at Crenshaw, Lubbub Creek, and Toqua sites

Bear Cougar Bobcat Passenger Pigeon Raptors Songbirds Rats and Mice

Crenshaw Structure

Lubbub Creek Mound

Toqua Structure 3

present present — abundant present abundant abundant

present — present present present abundant abundant

abundant present present abundant present present abundant

Copyright © 2012. Nebraska Paperback. All rights reserved.

same deposit), red-tailed and red-shouldered hawks, a cardinal, and a number of unidentified songbirds. Songbirds made a relatively poor showing in structure 3 at Toqua compared with the other two “elite” assemblages, being only slightly more common than was “average” for the site as a whole. Rats and mice appeared to be more than amply represented, collectively ranking fourth by percent nisp, identical to their ranking at Crenshaw using the same methods to judge frequency overall (table 3-5). The fact that rats and mice are so abundant in each of these “elite” deposits is unusual. Their presence and abundance suggests environmental conditions more than merely conducive to their survival. These commensal rodents apparently thrived in proximity to revered aboriginal leaders, presumably because provisioned elites occasionally garnered food supplies beyond their subsistence needs, but perhaps also because surplus maize or other foods were stored in proximity to these structures. Finally, the significance of birds in Native North American religions cannot be overemphasized. The only raptor recovered from Crenshaw is a long-eared owl, which, of all North American birds, most resembles an antlered deer in silhouette. The single element, a humerus, was recovered in the “control pit” excavation unit in the top of the antler deposit. Considering that this deposit produced few bones other than deer frontals, its presence must be considered to be non-random. At Lubbub, raptors occur only in elite deposits. Analysis currently under way of bone from elite middens at Moundville supports this statement. A peregrine falcon, originator of the “forked eye” motif common in Mississippian 82

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

iconography, was recovered from Mound G context at Moundville, as were several hawk elements from Mounds G and Q (Jackson and Scott 2003). Deliberate symbolism appears to be behind these inclusions. At Crenshaw, luckily, the symbolism appears obvious. In sum, assemblages from elite contexts including the structure at Crenshaw suggest that a constellation of taxa are good indicators of status. The markers identified to date include a diverse and relatively abundant assortment of birds, particularly passenger pigeon and songbirds, in ceremonial contexts, raptors, and other animals which due to their size or ferocity were difficult or dangerous to procure. In addition, assemblages yielding an abundance of rats and mice may well be associated with such elite status, not because they were used as food sources but because they were more concentrated in areas with both plentiful food and sufficiently intermittent human traffic that fatal encounters with humans were rare.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Conclusions The Crenshaw faunal assemblage is atypical of prehistoric Caddo faunas, an observation that is of little surprise given the unique nature of the site and the specific contexts from which the sample was recovered. It does, however, fit well in terms of general character with samples collected from elite contexts at other Mississippian sites, and thus provides another example of the ways in which such samples may mirror the unique positions and roles of the leaders of these prehistoric Southeastern societies. What is clear from the similarity of these samples and their distinctively different character vis-à-vis nonelite samples is that, although faunal assemblages do reflect something about the ecological and economic relationships that shape culture, they are also shaped by a myriad of other cultural rules that extend beyond these factors. In particular, along with other material expressions of status, food can be used to convey encoded messages expressing one’s position in society. At every level of food procurement or production, processing, and consumption, there is present opportunity for variation that expresses the different statuses of members of society (Douglas 1971; Goody 1982). These differences become more pronounced with the emergence of complex societies, one of the defining characteristics of which is the differential access to resources (Arnold 1996). In addition to symbolic demonstrations of social status conveyed in the foods that are eaten and how they At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

83

Copyright © 2012. Nebraska Paperback. All rights reserved.

are prepared, food may also convey messages related to the contexts in which they are consumed. Meals are rituals in and of themselves and are often integral aspects of a wide range of social, political, and religious ceremonial activities. Choices about the content of meals are governed by often complex sets of rules related to the purposes about which people gather to eat as well as the social statuses and interrelationships of the participants. The analysis of the Crenshaw assemblage brings into sharp focus the important social and ritual dimensions of food procurement, preparation, and consumption, particularly in complex societies (Gumerman 1997). To try to understand food refuse from archaeological sites as merely a reflection of nutritional or economic exigencies fails to account for the enormous symbolic content that food consumption has, including a complex set of reflections and strategic utilizations of social status, political influence, group composition, and more esoteric connections with the supernatural underpinnings that legitimize such relations and maintain the culturally defined notions of worldly order. Deer at Crenshaw played this central role in facilitating and reinforcing critical cosmological relations, defining social, political, and religious inequality, and reinforcing social solidarity. The accumulation of deer remains within and adjacent to feature 6 certainly reflects the preferential access by the politico-religious elite to certain cuts of meat (as well as certain other non-deer taxa), and also reflects the rituals performed to maintain the successful relationship between this critical species and the Caddo people. The central, although probably not the only, participant in all of these processes was the inhabitant of the structure, feature 6. If the outlying midden represents food remains in addition to those of the priest, it is not a far stretch to see a public side of his responsibilities, possibly serving as host or focal point of public ritual feasting. What is certainly clear is that accorded prime venison, performing the ritual butchery necessary to increase and maintain the antler deposit, and manipulating key animal symbols as part of early Caddo ritual, underlie the divergence of the Crenshaw assemblage from those from Caddo domestic contexts. Although the meanings behind the faunal patterning can only be faintly glimpsed in the ethnohistoric record of Caddo ritual hunting, it is clear that the assemblage cannot be fathomed in purely economic terms. 84

Jackson, Scott, and Schambach

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Acknowledgments

Copyright © 2012. Nebraska Paperback. All rights reserved.

We thank Michael Evans of the Arkansas Archeological Survey for preparing figure 3-2 and Deborah Sabo of the Arkansas Archeological Survey for assistance with figures 3-1 and 3-3. Our thanks to the Arkansas Archeological Society and the Arkansas Archeological Survey for permission to use these figures.

At the House of the Priest

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

85

4. Bioarchaeological Evidence of Subsistence Strategies among the East Texas Caddo diane wilson

Introduction

Copyright © 2012. Nebraska Paperback. All rights reserved.

Food is important symbolically because it is one of the few cultural components necessary to sustain life. As omnivores, humans pursue a vast array of diets, chosen at least as much for environmental and economic reasons as for history and identity (Sahlins 1972). Archaeological studies of subsistence in the New World have largely focused on the origins of agriculture and have used the presence of agriculture as a way to highlight diversity among cultural groups. Agriculture requires a higher degree of sedentism than hunting and gathering pursuits, and therefore often results in a more complex social system. Agriculture affects a human population’s health not only by changes in diet and nutrition but also through changes in workloads, settlement patterns, and overall mobility. These changes in turn result in changes in stress, disease, and pathogen loads. But what happens when a culture does not practice a subsistence strategy uniformly, as may have been the case among the Caddo in prehistoric and early historic times? Caddo populations are to a significant degree defined traditionally by having a maize-based agricultural subsistence strategy. However, archaeologists and bioarchaeologists have recently noted not only a diversity in the amount of maize consumed by different Caddo groups living in different parts of the Caddo archaeological area, but have argued that in some regions maize was never a key component of the Caddo diet (Burnett 1990, 1999; Rose et al. 1998; Perttula 2008). This chapter examines the incorporation and importance of maize in the Caddo diet by reviewing subsistence data from studies of human remains in northeast Texas. Specifically, dental data and stable isotope analyses from 86

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

numerous sites are reviewed, since this information provides the most direct evidence of diet. Secondary to this, dental wear is examined as it provides evidence of diet and processing strategies. Evidence of irondeficient anemia in the form of porotic hyperostosis and cribra orbitalia, which may be the result of a maize-based diet deficient in protein and/ or parasitic infection, are rare among the Texas Caddo and are therefore not included in this analysis. I show that there were differing degrees to which maize was consumed and that protein resources changed through time. Within the Caddo archaeological area, maize agriculture must be viewed within the context of existing social systems; maize agriculture functioned in relationship to economic food needs and trade networks, as well as in civic/religious situations. Ethnographically and ethnohistorically the Caddo were known to be maize agriculturalists, and symbolically maize was a very important part of the Caddo diet (Newkumet and Meredith 1988). For example, a corn gruel drink was described as the custom to be served at a gathering of elders and the chief (Carter 1995:102). While planting and harvesting were communal events, it was the xinesi who was responsible for the first sowing. In 1687, Joutel described the harvesting of maize as a communal event in which many Caddo gathered at a farmstead to help in the harvest and then they were treated to a feast (Swanton 1942:127). Such events both solidified communal relationships and sustained communities. For the Caddo, maize may have been an important component of the diet symbolically, but archaeological evidence is showing that it varied in consumption through time and across the region (Perttula 1993, 2008). Jackson et al. (chap. 3, this volume) discuss the symbolic nature of meat use in the Caddo diet. While historically maize is described as the base of the Caddo diet, meat may have had a more varied importance (Swanton 1942:135). Staple sources of animal protein were deer, duck, turkey, fish, and rabbit. Maize Agriculture and the Caddo The evolution of maize agriculture among the Caddo was clearly a complex process, likely influenced by political, social, economic, and environmental factors. Archaeological evidence shows that there was diversity in the use of maize throughout the Caddo area. Paleobotanical evidence shows that corn came into the general area as early as A.D. Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

87

Copyright © 2012. Nebraska Paperback. All rights reserved.

400, following domesticated weedy annuals that remained an important component of the diet until they were replaced by maize ca. A.D. 1300. However, both Perttula’s (2008) study and this chapter show that in some locations maize was an important component of the diet earlier than the Middle Caddo period (ca. A.D. 1200–1400). The development of agriculture among the Caddo has been described as “the product of local circumstance” by Harmon and Rose (1989:47). The largest regional summary of the development of agriculture based on bioarchaeological analysis largely agreed with this characterization. In Burnett’s summary (1990, 1999), she subdivided the Southern Caddo region into river basins and vegetative regions. She hypothesized that the relationship between maize and the natural environment was stronger than a simple increase in maize consumption through time. Burnett and Rose et al. (1998) described a higher dietary contribution of maize among the aboriginal inhabitants of the Post Oak Savannah and Blackland Prairie than in the Pineywoods. This study focuses on the portion of the Southern Caddo region south and west of the Red River. The area includes the Post Oak Savannah, Pineywoods, and Blackland Prairies vegetational areas of northeast Texas (Diggs et al. 2006:figs. 1–4). Unfortunately, bioarchaeological data relevant to the study of Caddo subsistence has been somewhat limited in this region. This is due largely to the poor preservation conditions in the acidic sandy deposits characteristic of this part of Texas. Often graves are identified within an archaeological site only by stains in the soil and/or by associated funerary objects (i.e., whole pottery vessels) in a burial-sized pit feature. Nevertheless, it is important to examine the preserved human remains from Caddo sites for evidence of diet as it is the most direct means of determining what a person ate and, in the case of the Caddo, because Rose et al. (1998:116) have shown that the paleobotanical record does not always match the bioarchaeological record for Caddo populations. The sites on which this analysis is based are from single and multiple mound centers, cemeteries, villages, farmsteads, and hamlets. They range from as early as the Formative Caddo period (ca. A.D. 800–1000) to as late as the Historic Caddo period (postdating A.D. 1680). The data reviewed herein are from previously published bioarchaeological studies as well as data I have collected from recent bioarchaeological analyses of Caddo human remains in several locales within the region. 88

Wilson

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Dental Data The analysis of dental remains provides a wealth of information about diet. Food items consumed and food preparation techniques leave evidence in the form of dental caries, attrition, abscess, and antemortem tooth loss. Throughout the Americas, dental disorders, especially dental caries, increased with the adoption of maize agriculture. Reliance on maize provides a sticky, carbohydrate-rich dietary staple high in sucrose that is favorable to microbial attack. Although Lee (1999) called into question a direct genetic link between Fourche Maline populations and the Caddo, the Woodland period populations who preceded the Caddo and had a much greater reliance on hunting and gathering in the region had fewer dental caries (Rose et al. 1984; Powell 1985; Burnett 1999) than later Caddo populations. Stable isotope results generally agree with the interpretation of these results (Rose et al. 1998; Tine and Tieszen 1994). In this chapter I propose to test Rose et al.’s 1998 hypothesis that carbohydrate consumption among Caddo groups in northeast Texas was variable until ca. A.D. 1100, when it became uniformly high, as well as Burnett’s hypothesis that maize consumption, based on the frequency of dental caries, was highest in the western portion of the region and decreased in an easterly direction across the Texas Caddo region. The Caddo site populations reviewed here do support Rose et al.’s 1998 hypothesis but do not support Burnett’s hypothesis.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Dental Attrition Dental attrition is the result of occlusal abrasion and a diet containing unprocessed vegetal materials and grit. Teeth respond to abrasion through the loss of enamel. Abrasion in the diet can be the direct result of hard and tough dietary items consumed, grit from the environment, and grit introduced from stone grinding implements. Rose et al. (1998) demonstrated a substantial decrease in abrasives in the diet between pre–A.D. 800 Fourche Maline and post-A.D. 800 Caddo time periods. Using micro-wear patterns on teeth, Rose et al. and Harmon and Rose (1989) linked the decrease in abrasives to a change in the consumption or processing of hickory nuts and a change from stone to wooden grinding implements. In this chapter, I relied on studies using standardized techniques for estimating dental wear summarized by Smith (1984) and Scott (1979). Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

89

Table 4-1. Average dental attrition scores for sites in northeast Texas Average wear using Smith (1984)

Average wear using Scott (1979)

41DT1 Manton Miller 41DT6 Tick 41DT16 Spike 41DT80 Thomas 41DT124 Doctors Creek 41HP78 Lawson 41HP102 Arnold 41HP105 Cox 41HP106 Hurricane Hill

7.8 7 4.3 6.9 4.6 — 6.3 6.7 5

39 30.5 — 34.7 15.1 21.7 24.6 35.8 27.9

Average

6.1

28.7

41RR16 Roitsch (Derrick et al. 2008) 41LR2 Sanders (Wilson 1997b) 41MR63 Pleasure Point (Wilson and Steele 1996)

4.2 — 4

14.4 16.5 —

Average

4.1

15.4

41CP71 Shelby Mound (Wilson 2004) 41HS269 C.D. Marsh (Wilson 2006) 41NA231 Tallow Grove (Wilson 2008b) 41BW4 Mitchell (Wilson 1997b) 41AN38 Lang Pasture (Wilson 2011) 41AN14 Isibell-Gene Donnell (Wilson 2011) 41AN21 Emma Owens Farm (Wilson 2011) 41AN32 Fred McKee (Wilson 2011) 41CE3 Solon Stanley (Wilson 2011) 41CE4 J. W. Blackburn (Wilson 2011) 41CE6 E. W. Hackney (Wilson 2011) 41CE12 Jim Allen (Wilson 2011) 41CE17 E. W. Henry (Wilson 2011) 41CE19 George C. Davis (Wilson 2011) 41CE290 Forest Mound (Wilson 2011)

4 — 2.9 — 3.8 3.0 4.0 3.2 4.7 3.4 4.8 4.5 5.2 4.0 3.3

16.8 18 9.9 14.4 19.1 24 20.2 16.8 16.9 20.2 20.4 11.9 14.9 17.5 17.8

Average

3.9

17.2

Site Blackland Prairie (Wilson and Steele 1997)

Post Oak Belt

Copyright © 2012. Nebraska Paperback. All rights reserved.

Pineywoods

Note: Wear scores for anterior teeth from Smith (1984) and molar scores follow Scott (1979). A significant difference in dental wear was found among the regions.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Since dental disorders, particularly wear, are age cumulative, it is important to review the age of the populations in this summary: only adult dental data are included. Adults are defined roughly as individuals more than 20 years of age at the time of death. Average adult age at death is the same for populations that lived in the Blackland Prairie, Post Oak Savannah, and Pineywoods and is not considered to have been causal in any differences in wear patterns. In their study of the Roitsch site (41RR16) populations, Derrick et al. (2008) found higher rates of anterior dental wear than molar wear. They suggest that this may be the result of the use of teeth as tools for hide processing or using teeth to hold objects, leading to the high rate of first molar antemortem loss. This, combined with the larger surface area over which molar teeth grind objects for digestion, means that molar wear patterns may be more reflective of grit in the diet than those seen on anterior teeth. Table 4-1 presents the average dental wear for Caddo sites in the northeast Texas region. It clearly shows that Blackland Prairie populations had more grit in their diet. In fact, Kruskal-Wallis one-way anova analysis on the molar and anterior teeth show that the difference is significant at the p < .01 level. The same tests demonstrate highly significant differences (p < .01) in dental wear through time. The significant difference is between populations from the Formative/Early Caddo periods (ca. A.D. 800–1200) and more recent Caddo populations. Molar wear from the Formative Caddo period averaged 27.9, but 31.8 during the Formative to Early Caddo periods, and then only 21.1 during the Early Caddo period. After that, molar wear scores dropped to a mean of 14.2 and 16.9 during the Middle and Late Caddo periods, respectively. When comparing earlier populations from the Blackland Prairie and Pineywoods, sites from the Pineywoods had lower mean dental wear. This suggests that the location of Caddo populations in a particular biotic region was important in the temporal trend of food items and processing techniques used by the Caddo. Analysis by river basin also shows that the Blackland Prairie Caddo stand out. When dental attrition is examined by river basin, there is also a statistically significant difference (p < .01) among the five drainages for which data was available. When the Cooper Lake populations from the Sulphur River are omitted (Blackland Prairie sites), there is no statistically significant difference in a Kruskal-Wallis one-way anova test. Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

91

Caddo cultures clearly had much in common with their Mississippian and Southeastern contemporaries, including obvious status differentiation. In an attempt to roughly examine status, sites were divided into mound and non-mound contexts. Although there is higher average tooth wear at mound sites, statistical significance was not seen in dental wear. This suggests that social status was not a factor in Caddo diet and/ or food processing strategies. Future studies with more mound populations will be needed to test this finding. Taken together, changes in dental wear were both temporal and regional in the Texas portion of the Caddo region. Dental attrition shows that food processing and/or diet was coarser in the Formative Caddo period and remained so in the Blackland Prairie region and Sulphur River drainages into the Early Caddo period, while populations in the Pineywoods had a softer diet and/or used wooden grinding tools in the Early Caddo period. Dental Caries

Copyright © 2012. Nebraska Paperback. All rights reserved.

Dental pathologies are generally regarded as age progressive; however, this is not always the case with archaeological populations and dental caries (Hillson 1979). Caries are not age progressive among the Caddo from this region (Wilson 1997a, 2008b, 2009; Wilson and Steele 1997). Rose et al. (1984) suggested that the incidence of caries in a population covaries with the amount of maize that was consumed. More recently, Rose et al. (1998) have suggested that it is the increased proportion of dietary carbohydrates in the diet rather than maize that results in a relatively high frequency of caries. Although maize should not be regarded as the sole source of sucrose in the Caddo diet, maize is higher in sucrose than other dietary staples, followed by squash and sunflower seeds. Lactobacillus acidophilus and Streptococcus mutans bacterium in the mouth convert sucrose into an acid that destroys enamel and dentin. Frequencies of carious teeth that range from 8 to 25 percent are considered to be indicative of maize agriculture (Cohen and Armelagos 1984; Powell 1985; Larsen et al. 1991), while Turner (1979) gave a larger range of 2.3 to 26.9 percent carious teeth to indicate an agricultural diet. Others have used a caries rate (the caries rate is equal to the number of caries divided by the number of individuals in the population) to differentiate between high and low carbohydrate diets. Rose and Marks (1985) used a caries rate of 2.0 to distinguish high carbohydrate diets. 92

Wilson

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Biotic Region The frequency of caries in northeast Texas Caddo sites indicates that maize was a significant part of the diet in all but the Blackland Prairie biotic region (table 4-2). In contrast, Burnett (1990, 1999) found the highest caries rate in the Blackland Prairie and the lowest caries rate in the Pineywoods. Here, I separate Burnett’s data and present her findings but do not use her data, because her summary data has in many instances been reexamined in new bioarchaeological studies. A one-way anova test reveals that the difference among the regions is not statistically significant (p < .01). In this study a minimum of 26 individuals from the Blackland Prairie, 290 from the Post Oak Savannah, and 116 from the Pineywoods had teeth that could be observed for caries. The caries rate from Wilson and Steele (1997) for the Blackland Prairie sites is lower than the threshold for a high carbohydrate diet. The caries rate for the Caddo living in the Pineywoods region had considerable diversity. Five (Tallow Grove, Beech Ridge, Forest Mound, Lang Pasture, and C. D. Marsh) of the 19 bioarchaeological populations from Pineywoods Caddo sites have caries rates indicative of a low carbohydrate, presumably maize, diet. The Tallow Grove and C. D. Marsh site populations each lacked caries, but the populations were small. From the C. D. Marsh site, there were only two teeth present from one Historic Caddo period individual, and neither were molars; molars are more commonly affected by caries. Tallow Grove, a Middle Caddo period site from Lake Naconiche (Wilson 2008b) in the Attoyac Bayou basin, had two individuals with a total of 27 teeth between them. The two large populations with at least 10 adults with dental remains from the Pineywoods are from Jim Allen and George C. Davis, with a mean caries rate of 5.5 between the two sites. Time For this study burials were assigned to time periods following Story’s 1990 categories based on radiocarbon dates when available and archaeological context and associations. In some cases individuals could not be assigned to specific time periods; most of these were omitted from temporal analysis. However, if an individual came from a site that had a burial population that predominantly dated to one period and an individual could not positively be placed in that period, it was assumed to Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

93

Table 4-2. Caries data for sites discussed in the Blackland Prairie, Post Oak Savannah, and Pineywoods of northeast Texas Site

Caries rate (caries/individual)

Teeth with caries (%)

Individuals with caries (%)

3.8 1.7 1.8 2.8

— 10.5 10.5 —

— 26.0 26.0 —

2.8 — — — 6.5 7.4 — 4.5 3.3 3.0 5.4 4.6

— 5.6 10.1 15.3 56.0 — 14.1 23.0 19.0 — 20.0 —

— — — — — — — 91.0 79.0 — 79.0 —

0.5

—

—

2.7 6.8 2.0 2.2 0 1.0 5.0

12.3 27.1 50.0 — 0 — —

67 100 100 — 0 100 —

2.0 3.0 6.5 — 3.0 13.0 4.5

50.0 17.6 27.6 47.4 27.0 58.8 25.6

100 100 100 100 100 100 100

Blackland Prairie Burnett’s (1990, 1999) summary for Blackland Prairie Cooper Lake (Wilson and Steele 1997) mean including Burnett’s data mean including Burnett’s data Post Oak Savannah Burnett’s (1990, 1999) summary for Post Oak Savanna 41RR77 Rowland Clark (Loveland 1986) 41RR11 Holdeman (Loveland 1986) 41RR16 Kaufman-Williams (Loveland 1986) 41RR16 Sam Kaufman (Colby 1997) 41RR16 Roitsch (Derrick et al. 2008) Bentsen-Clark (Butler 1969) 41BW4 Mitchell (Wilson 1997b) 41LR2 Sanders (Wilson 1997b) 41CP71 Shelby Mound (Wilson 2004) mean including Burnett’s data mean including Burnett’s data

Copyright © 2012. Nebraska Paperback. All rights reserved.

Pineywoods Burnett’s (1990, 1999) summary data for the Pineywoods 41NA49 Washington Square Mound (Franciscus 2008) 41HS15 Pine Tree Mound (Wilson 2008a) 41RK170 Nawi haia ina (Wilson 2003) Henry Spencer (Wilson 2006) 41NA231 Tallow Grove (Wilson 2008b) 41NA2242 Beech Ridge (Wilson 2008b) 41NA285 Boyette (Wilson 2008b) 41AN14 Isibell-Gene Donnell Site (Wilson 2011) 41AN21 Emma Owens Farm (Wilson 2011) 41AN32 Fred McKee (Wilson 2011) 41CE3 Solon Stanley Farm (Wilson 2011) 41CE4 J. W. Blackburn (Wilson 2011) 41CE6 E. W. Hackney (Wilson 2011) 41CE12 Jim Allen (Wilson 2011)

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Site

Caries rate (caries/individual)

Teeth with caries (%)

Individuals with caries (%)

22.7 10.8 24.2 16.7 3.7 0 24.8 —

100 — 93.3 100 17.0 0 81.0 —

Pineywoods 41CE15 A. H. Reagor (Wilson 2011) 41CE17 E. W. Henry (Wilson 2011) 41CE19 George C. Davis (Wilson 2011) 41CE290 Forest Mound (Wilson 2011) 41AN38 Lang Pasture (Wilson 2011) 41HS269 C. D. Marsh (Wilson 2006) mean including Burnett’s data mean excluding Burnett’s data

— 4.5 6.5 1.0 0.2 0 3.8 3.4

Copyright © 2012. Nebraska Paperback. All rights reserved.

Note: Burnett’s (1990, 1999) data were segregated because several of the individuals included in her data are included here in other reports.

fall into that period. For example, burial 13 from the Hurricane Hill site was dated to the Early through Middle Caddo periods. Rather than omit burial 13 from the temporal analysis, it was lumped with Early Caddo period burials, since the majority of the burial population from this site dated to the Early Caddo period. When examining the biotic regions in northeast Texas where the bioarchaeological populations were recovered, one obvious difference between the Blackland Prairie region bioarchaeological samples and those from the other two regions is in the temporal composition of the samples. The majority of individuals from the Cooper Lake study date from the Early Caddo period (Wilson and Steele 1997). The majority of individuals from the Pineywoods and Post Oak Savannah date to the Late Caddo period. Although not emphasized in her bioarchaeological study, Burnett’s 1990 and 1999 data does suggest that there was a general but small increase in caries rate through time in northeast Texas. Her best data is from the Red River where caries rates increase from 1.9 in the Middle Caddo period to 3.4 in the Late Caddo period. This agrees with Perttula’s 2008 analysis, in which maize is described as being less important as a dietary staple until the latter part of the Middle Caddo period, around A.D. 1300. Caries rates in this study do not show a significant temporal trend. Mean caries rates for the Early, Middle, and Late Caddo periods are 1.7, Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

95

2.0, and 5.0. The only Historic Caddo site included in this study, the Jim Allen site, has a caries rate of 4.5. The means are biased by small population sizes. When site populations are examined that have at least 10 individuals, the lack of temporal trends in the caries data is clear. The predominantly Early Caddo period George C. Davis site has a caries rate of 6.5, followed by the Middle Caddo period Sanders site population with a caries rate of 3.3, and the three predominantly Late Caddo populations from the Mitchell and Roitsch/Sam Kaufmann sites are 4.5, 7.4, and 6.5, respectively (see table 4-2). River Basin

Copyright © 2012. Nebraska Paperback. All rights reserved.

Burnett’s (1990, 1999) bioarchaeological data suggested that only populations living along the Sulphur and Trinity River drainages ate significant amounts of maize. For the current analysis, new bioarchaeological data was not available from the Trinity River basin. But data presented here from upper Sulphur River Caddo sites does not bear out Burnett’s findings. Wilson and Steele’s 1997 data from Cooper Lake contradict Burnett’s data from the Sulphur River in that they determined there was a caries rate of 1.7 for the Cooper Lake population as a whole and 1.3 for the individuals who could be definitively attributed to prehistoric Caddo contexts; this compares to Burnett’s analysis that the caries rate was 4.8. Perttula’s (2008:table 2) examination of the paleobotanical record adds a temporal dimension, with maize ubiquity peaking during the Early Caddo period in the upper Sulphur River basin and later in other river basins as well as in one Late Caddo site (Ear Spool, 41TT653) downstream in the Sulphur River basin. While the Middle Caddo period individuals from the Hurricane Hill site (41HP106) had a caries rate of 0, a larger sample size is needed to evaluate the possibility that maize use declined after ca. A.D. 1200 in the Sulphur River basin. When means for the river basins are examined, only the Sulphur and Cypress Creek populations had average caries rates below the 2.0 high carbohydrate level. Statistical analysis using the Kruskal-Wallis one-way anova test showed no significant difference in caries rates among the river basins. Sample size was likely the main contributing problem. The Cypress Creek data came from three sites, only one of which had a caries rate below 2.0. The average caries rate is highest in the Sabine River basin with a mean of 4.4. But here only two sites were represented: the Nawi haia ina (41RK170) and Pine Tree Mound (41HS15) sites. Other 96

Wilson

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

than in the Sabine River basin, there was diversity within each river basin in this caries rate measure. In the Red and Sabine River drainage basins, paleobotanical data from Caddo sites indicate that maize use peaked sometime after the Middle Caddo period, while in Caddo paleobotanical assemblages in the Cypress Creek, Neches, and Angelina River basins maize use peaked sometime during and after the Late Caddo period (Perttula 2008:table 2). Dental data can confirm Perttula’s findings in the Red River basin but not in the Neches River basin; more bioarchaeological data are needed from Caddo sites in the Sabine River basin. In the Red River basin, the highest caries rates are from the Roitsch site and the lowest are from the Sanders site. Caries rates in the Neches River drainage do not appear to peak during the Late Caddo, with sites such as the predominantly Early to Middle Caddo period George C. Davis mound center population having a caries rate of 6.5 and the predominantly ca. early fourteenth and early fifteenth century Lang Pasture habitation site having a caries rate of only 0.2. Clearly, more well-dated assemblages will provide a clearer answer to the apparent discrepancies in caries rates through time.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Site Type Dental data show a difference in diet between mound and non-mound Caddo burial populations that almost reaches statistical significance (p = 0.06). When individuals buried at sites with mounds are examined apart from individuals buried in villages and hamlets, caries rates are nearly always high in mound contexts, suggesting a higher carbohydrate diet may have been consumed by Caddo residents at these civic-ceremonial sites. This is the opposite pattern from that described by Ambrose et al. (2003) for the consumption of maize at Cahokia. Sites not associated with mounds have caries rates that are more diverse than mound context burials. The average caries rate from non-mound context is 2.5, while the average caries rate from mound contexts is 4.6. Non-mound Caddo sites in the region have caries rates that range from 0 to 13, indicating significant variability in maize consumption, more so than at the mound centers, while mound sites have caries rates that ranged from 1.0 to 7.4. Stable Isotope Data Both the available paleobotanical record and the bioarchaeological record from Arkansas and Louisiana indicate that maize became an important component of the Caddo diet later than caries rates might Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

97

Copyright © 2012. Nebraska Paperback. All rights reserved.

suggest. Rose et al. (1998), in an early examination of stable carbon isotope results for Caddo populations from Arkansas and Louisiana, argued that increased caries from the Woodland (pre–A.D. 800) and Formative Caddo periods to the Early Caddo period was the result of the increased consumption of carbohydrates, not necessarily the consumption of maize. Perttula (2008) found that by the Early Caddo period maize was present throughout east Texas but was likely less important to the Caddo diet than starchy annuals as well as wild plants and meat in most Early Caddo sites. If caries rates indicate an increase in carbohydrate consumption but cannot distinguish between maize and other carbohydrate sources as to which carbohydrates were being consumed, then stable isotope analysis provides the only means of determining when and where maize became an important component of the Caddo diet. The summary of stable isotope results from the region indicates that there was a general increase through time in maize consumption, but with some notable exceptions. Table 4-3 summarizes the available stable isotope data from the east Texas region. Stable isotope data has been compiled from 77 Caddo individuals from 29 sites. Results are available for carbon on collagen and apatite and for nitrogen on collagen. The majority of the stable isotope results are for stable carbon isotopes on collagen. Stable isotope analysis is used to examine the trophic levels of organisms and the amount of plant matter in the diet from different photosynthetic pathways: C3, C4, and CAM. In the Caddo region of east Texas, the environment is characterized by plants that use the C3 photosynthetic pathway. Maize can be recognized in stable isotope signatures from the region because it is a C4 plant. The collagen enrichment factor in stable carbon isotope analysis is the difference between the dietary and bone signature for carbon and is approximately 5‰. C3 plants have a delta 13C gelatin signature of -26.5‰ on average, while bone from a C3 plant consumer would have a delta 13C gelatin value of -21.5‰. C4 plants have a higher delta 13C gelatin signature of -12.5‰ on average, and bone from a pure C4 consumer would have a delta 13C gelatin value of -7.5‰. Carbon collagen is more commonly sampled than apatite in isotope analyses. Collagen values reflect the isotopic composition of the organic fraction of bone, while apatite reflects the mineral portion. The two likely show different amounts of individual life span and signal different 98

Wilson

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

41HP106 Hurricane Hill 1 41HP106 Hurricane Hill 2 41HP106 Hurricane Hill 4a 41HP106 Hurricane Hill 4b 41HP106 Hurricane Hill 13 41HP106 Hurricane Hill 15 41HP106 Hurricane Hill 18a 41LR2 Sanders 17-1 41LR2 Sanders 17-2 41LR2 Sanders 17-3 41TT13 Alex Justiss 3 41TT13 Alex Justiss 12 41TT13 Alex Justiss 7 41TT13 Alex Justiss 5 41TT13 Alex Justiss 6 41TT13 Alex Justiss 13 41TT13 Alex Justiss 9 41TT13 Alex Justiss 19 41TT13 Alex Justiss 28 41TT13 Alex Justiss 23 41CP10 Harold Williams 41RR11 Holdemann 11

Site / burial no. Early Caddo Early Caddo Early Caddo Early Caddo Early–Middle Caddo probably Formative Caddo probably Formative Caddo early Middle Caddo early Middle Caddo early Middle Caddo Late Caddo/Titus phase Late Caddo/Titus phase Late Caddo/Titus phase Late Caddo/Titus phase Late Caddo/Titus phase Late Caddo/Titus phase Late Caddo/Titus phase Late Caddo/Titus phase Late Caddo/Titus phase Late Caddo/Titus phase Late Caddo/Titus phase Late Caddo

Time

Table 4-3. Stable isotope data from the study region

Copyright © 2012. Nebraska Paperback. All rights reserved.

13

-16.5 -15.3 -15.3 -17.6 -15.2 -15.0 -15.0 -10.0 -12.9 -11.9 -20.9 -21.6 -15.4 -21.6 -15.9 -22.7 -13.2 -14.9 -13.4 -22.0 -14.0 -16.9

C gelatin

13

-9.6 -7.4 -8.8 -10.2 -7.3 -8.7

C apatite 6.9 7.9 6.5 7.4 7.9 6.3

Difference

15

8.9 8.9 9.0 8.4 9.2 9.1

N gelatin Blackland Prairie Blackland Prairie Blackland Prairie Blackland Prairie Blackland Prairie Blackland Prairie Blackland Prairie Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah

Biotic zone

(continued)

Sulphur Sulphur Sulphur Sulphur Sulphur Sulphur Sulphur Red Red Red Cypress Creek Cypress Creek Cypress Creek Cypress Creek Cypress Creek Cypress Creek Cypress Creek Cypress Creek Cypress Creek Cypress Creek Cypress Creek Red

River basin

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

41RR11 Holdemann 14 41RR11 Holdemann 23 41RR11 Holdemann 21 41RR16 Roitsch/Sam Kaufman 15 41RR16 Roitsch/Sam Kaufman 17 41RR16 Roitsch/Sam Kaufman 19 41RR16 Roitsch/Sam Kaufman 20 41RR77 Rowland Clark 2 41RR77 Rowland Clark 7 41RR77 Rowland Clark 21 41RR77 Rowland Clark 33 41SM300 Lindsey Park 41CP245 Polk Estates 41CS1 Coker Mound 41RK170 Nawi haia ina 1 41RK170 Nawi hai ina 2 41NA231 Tallow Grove 45 41NA231 Tallow Grove 65 41NA231 Tallow Grove 77 41NA242 Beech Ridge 1 41NA285 Boyette 1/10 41NA285 Boyette 13A 41AN38 Lang Pasture 76D

Site / burial no. Late Caddo Early Caddo Middle Caddo Late Caddo Late Caddo Late Caddo Late Caddo Late Caddo Late Caddo Late Caddo Late Caddo Late Caddo Early Caddo Middle Caddo early Late Caddo Middle Caddo Middle Caddo Middle Caddo Middle Caddo Middle Caddo early Middle Caddo Middle Caddo late Middle Caddo

Time

-17.0 -14.5 -14.0 -11.5 -21.8 -21.2 -14.6 -15.1 -15.4 -15.3 -15.3 -9.1 -13.4 -10.0 -16.3 -13.9 -16.3 -18.1

-14.5 -20.2 -15.5

C gelatin

13

Table 4-3. Stable isotope data from the study region (continued)

Copyright © 2012. Nebraska Paperback. All rights reserved.

13

9

8.4 4.7 8.8

-7.9 -11.6 -9.3

Difference

-18.1

-8.0 -8.2 -8.2 -8.0

C apatite

6.3

8.7

10.8

N gelatin

15

Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Post Oak Savannah Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods

Biotic zone

Red Red Red Red Red Red Red Red Red Red Red Neches Cypress Creek Sulphur Sabine Sabine Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina

River basin

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

41AN38 Lang Pasture 92 41AN38 Lang Pasture 93B1 41AN38 Lang Pasture 76C 41AN38 Lang Pasture 93B2 41AN38 Lang Pasture 8 41AN38 Lang Pasture 88 41AN38 Lang Pasture 76B 41CE19 George C. Davis 119S1 41CE19 George C. Davis 155 41CE19 George C. Davis 130 41CE12 Jim Allen AH-3 41CE12 Jim Allen AH-11 41CE12 Jim Allen AH-10 41CE12 Jim Allen AH-6 41CE12 Jim Allen AH-12 41CE12 Jim Allen AH-5 41CE12 Jim Allen AH-7 41AN21 (Emma) Owens Farm G-1 41CE6 E. W. Hackney AN-1 41CE6 E. W. Hackney AN-2 41AN56 Lee Ellis Farm 41CE4 J. W. Blackburn AI-3 41CE4 J. W. Blackburn AI-2 41AN54 O. L. Ellis Farm 41AN54 O. L. Ellis Farm AE-2-B

Middle-Late Caddo early Late Caddo late Middle Caddo early Late Caddo early Late Caddo middle Late Caddo Middle-Late Caddo Early-Middle Caddo Formative Caddo Early-Middle Caddo Historic Caddo Historic Caddo Historic Caddo Historic Caddo Historic Caddo Historic Caddo Historic Caddo Late Caddo Late Caddo Late Caddo Early-Late Caddo Late Caddo Late Caddo Late Caddo Late Caddo

Copyright © 2012. Nebraska Paperback. All rights reserved.

-9.7

-21.0 -18.8 -18.7 -15.2 -15.6 -19.5 -17.6 -15.6 -16.6 -19.7 -19.6 -17.1 -17.4 -18.9 -21.8 -18.4 -14.5 -13.1 -13.7 -13.3 -14.8 -13.9 -12.9 -13.9 13.1 -12.8

-6.4 -10.2 -10.8 -9.7 -9.2 -8.7 -10.3 -11.7 -11.1 -8.7 -7.3 -7.8 -5.7 -8.3 -9.9 -6.1 -6.8 -6.8 -6.7 -7.4 -7.6 -7.7 -7.6 -8.1 -7.4 2.1

14.6 8.5 4.4 9.8 6.4 11 6.8 7.2 10.7 9.7 7.2 5.3 8 5 4.9 7.8 6.1 7.1 6.1

13.7 12.4

4.7 9.3 2.9 11.0 1.6 4.2 12.7 6.3 2.8 8.9 12.4

7.2 8.7

9.7

6.7

Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods

(continued)

Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Time

C gelatin

-19.3 -19.9 -15.9 -13.3 -12.2 -12.0 -14.9 -13.3 -14.8 -13.8

13

13

4.8 5 1.2 7.4 6.6 5.7

-6.7 -9.1

Difference

-14.6 -10.9 -12.1 -4.8

C apatite

13.2

10.4 10.8

10.9 11.5

N gelatin

15

Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods Pineywoods

Biotic zone

Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina Neches/Angelina

River basin

Hurricane Hill data from Gill-King (1999); Sanders data from Wilson and Cargill (1993); Alex Justiss data from Rogers et al. (2003); Harold Williams data from Turner et al. (2002); Holdeman and Rowland Clark from Perttula (1997); Roitsch/Sam Kaufman are from Derrick et al. (2008); Coker Mound is from Derrick (1997); 41CP245 is from Nelson and Perttula (2006); 41SM300 is from Perttula and Nelson (2003a) and 41RK170 is from Perttula and Nelson (2003b). Where two values are given for collagen, the latter is from Beta Analytic radiocarbon data. There is some evidence of differences in results from different laboratories using different machines and processing techniques, and this was found by Rose et al. (1998) as well.

41CE3 Solon Stanley Farm AK1,2,3+4 Early Caddo 41CE290 Forest Mound Middle Caddo 41CE17 E. W. Henry Farm AL-2 Late Caddo 41AN32 Fred McKee Late Caddo 41CE15 A.H. Reagor Farm AJ-1 Late Caddo 41CE15 A.H. Reagor Farm AJ-2 Late Caddo 41AN34 Pierce Freeman Farm D-2 Late Caddo

Site / burial no.

Table 4-3. Stable isotope data from the study region (continued)

Copyright © 2012. Nebraska Paperback. All rights reserved.

aspects of an individual’s diet. Collagen signatures are biased by protein sources, particularly in low protein diets as has been suggested for Native American maize agriculturalists (Ambrose 1993). Ambrose and Norr (1993) and Tieszen and Fagre (1993) have proposed that delta 13C values from collagen vary in source representation based on the amount of protein in the diet of the consumer. For example, in populations of aboriginal North American maize agriculturalists, it has been shown that the carbohydrate portion of the diet is reflected in delta 13C collagen values because the amount of protein in the diet was relatively low. Because it is not biased by protein source, apatite should be better able to show a small increase in the amount of maize in a given diet. The difference between collagen and apatite isotopic values can be used to determine trophic level. Ambrose (1993) suggested that where the majority of carbohydrates are from a C4 source, such as maize, and the protein is a C3 source in a low protein diet, the difference between delta 13C values of bone collagen and apatite will be large. Nitrogen values are ultimately a result of bacterial fixation or soil nitrates that are passed up the food chain with a 2–3‰ increase per trophic level. Humans in terrestrial-based food webs typically have delta 15N values of 6–10‰, whereas consumers of fish may have delta 15N values that range as high as 15–20‰ (Schoeninger and DeNiro 1983). Unlike carbon isotopes, nitrogen stable isotopes also reflect temperature and humidity in the natural environment.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Time C4-enriched isotope values indicate that maize was present and a significant part of the Texas Caddo diet in both Formative Caddo and Early Caddo diets. There is a general increase in maize consumption through time. Second, considerable variability exists in the composition of the diet, particularly during the Middle and Late Caddo periods. This variability is a key component of the east Texas prehistoric Caddo diet (figs. 4-1 and 4-2). Some of this may be explained by a subsistence strategy that utilized different types of maize itself with variable isotopic signatures (see Perttula 2008). Delta 13C collagen data suggest that maize was a more important component of the Caddo diet in east Texas before that was the case in other Caddo regions outside the state. By the Middle Caddo period, other Caddo regions were consuming similar amounts of maize. Delta Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

103

4-1. The relationship between apatite delta 13C and delta 15N collagen values through time. Delta 13C apatite values increase through time while delta 15N values show increasing variability through time. 13

Copyright © 2012. Nebraska Paperback. All rights reserved.

C collagen values indicate more maize was consumed by the Formative Caddo populations from Texas than contemporaneous Mississippian populations from the eastern Ozark region (see Lynott et al. 1986). During the Formative Caddo period, delta 13C collagen values do not differ from values in the middle Mississippi valley (Buikstra and Milner 1991; Schurr 1992). After the Formative Caddo period, delta 13C collagen values are, in general, lower than Mississippian values for the same time period (see Schurr 1992 for a comprehensive table of Mississippian isotope values). After the Formative Caddo period, the variation in delta 13C collagen values within sites is also somewhat greater among the Texas Caddo (mean standard deviation = 1.63, adjusted standard deviation = 1.94) than Mississippian populations. Perttula (2008) showed the large degree of variability in delta 13C values for maize (standard deviation 1.32) in east Texas sites, which probably accounts for some of the variability seen among the Caddo. In a discussion of dietary variability in Late Prehistoric maize agriculturalists from the middle Ohio valley, Greenlee (2006) clearly demonstrated both inter- and intra-site differences, but overall Caddo variability is greater than that seen in the middle Ohio valley (mean standard deviation of 1.30, which is the same 104

Wilson

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

4-2. The relationship between apatite delta 13C and delta 15N collagen values through time. Early Caddo delta 13C collagen values show maize consumption but are higher (less maize intensive) than more recent periods.

as Schurr gives for Mississippian populations). However, there is almost no difference in the four delta 13C values for maize from the middle Ohio valley (Greenlee 2006). Formative Caddo stable carbon isotope values range from -21.8‰ at the George C. Davis site to a high of -15.0‰ in burial 15 at the Hurricane Hill site (see table 4-3). Unfortunately, dental remains were not present for burial 15 from the Hurricane Hill site (Gill-King 1999), but the caries rate from the other Caddo individuals with teeth from Hurricane Hill was very low at 0.75. The stable carbon isotope value on collagen was broadly similar in all individuals sampled from the Hurricane Hill site (-15.0 to -17.6‰). In this case, caries data indicate a low carbohydrate diet while stable isotope values show that maize was an important component. Fourche Maline populations from southeastern Oklahoma and southwest Arkansas from the Mahaffey, Old Martin Place, and Crenshaw sites averaged -21.21‰, -21.73‰, and -21.20‰, respectively (Rose et al. 1998) and are similar to the value of -21.8‰ from the Formative Caddo period individual from the George C. Davis site (see table 4-3). This suggests that maize was more important in the diet of Formative Caddo populations living in the Sulphur River part of the Blackland Prairie in the western part of the region than in populations to the north and east. Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

105

Table 4-4. Mean isotopic values through time alongside caries rates and dental wear scores using the Scott (1979) system for posterior dentition

Time Period Woodland to Formative Caddo Early Caddo Middle Caddo Late Caddo Historic Caddo

Denta wear

Caries rate

Delta C collagen

Delta C apatite

Delta 13C apatite-collagen spacing

27.9 21.1 14.2 17.3 11.9

2.8 1.7 2.0 5.0 4.5

-18.40 -17.77 -15.86 -15.90 -13.74

-9.90 -9.79 -10.19 -8.10 -7.41

8.50 7.29 8.08 6.09 6.33

13

13

Delta 15N Collagen

9.10 9.14 8.53 9.80 6.63

Copyright © 2012. Nebraska Paperback. All rights reserved.

Similarly, Rose et al. (1998) had obtained Early Caddo period average delta 13C collagen values of -19.78‰ for Crenshaw, -20.12‰ for Miller’s Crossing, and -20.17‰ for the Hanna site; these three sites are in the Red River basin. The data they present from northwest Louisiana and southwest Arkansas show that maize was more important in the Early Caddo Texas population diet. Early Caddo period stable carbon isotope values from east Texas range from -20.2‰ at the Dan Holdeman site (41RR11) on the Red River to -15.3‰ at the Hurricane Hill site (see table 4-3). Again these values show that maize was an important part of the Caddo diet for Caddo groups living in the western portion of the region in Texas before it became an important component of the diet in other Caddo communities. Middle Caddo stable carbon isotope values on collagen range from -21.0‰ from the fourteenth-century Lang Pasture site (feature 92 as tested by Geochron, -18.8‰ from Beta Analytic) to -10.0‰ from the Sanders site (feature 17-1) on the Red River. Despite the range of values, the mean from the Middle Caddo period (table 4-4) is slightly higher than the mean for the Early Caddo (or Caddo II) period from Arkansas of -13.56‰ (Rose et al. 1998). Late Caddo period stable carbon collagen values have the most variability, ranging from -22.0‰ at the Titus phase Alex Justiss site (feature 23) to -9.7‰ at the Frankston phase J. W. Blackburn site (feature AI-2) in the Neches River basin. While this is also the time of greatest variability in maize isotopic values (see Perttula 2008), the differences are greater than can be explained by varieties of maize alone. This suggests that there was a tremendous variability in the amount of maize being 106

Wilson

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

consumed during the Late Caddo period by different Caddo individuals and populations. Mean delta 13C collagen values from the Late Caddo period (see table 4-4) are nearly identical to the mean from Caddo burials at the Belcher Mound site (-15.88‰) on the Red River in northwest Louisiana. The Historic Caddo mean from the Cedar Grove site (3LA97) (-14.11‰, Rose et al. 1998) is also nearly the same as the mean from Historic Caddo sites in east Texas (see table 4-4). In this study all Historic Caddo remains are from the Jim Allen site in the upper Neches River basin. Differences in stable isotope values in the east Texas Caddo bioarchaeological record through time are statistically significant only for delta 13C apatite values (p = 0.009). Apatite is a more sensitive indicator of small changes in maize consumption. Maize consumption remained at a consistent level from the Formative Caddo through Middle Caddo periods. There is a large increase in maize consumption between the Middle and Late Caddo periods. There is also a considerable increase in the caries rate between the Middle and Late Caddo periods. In fact, caries rates increase the most between the Middle and Late Caddo periods. The Historic Caddo values for apatite represent a further increase in maize consumption. While the lowest mean delta 13C apatite value occurs in the Middle Caddo period, ranges in values were also the largest during this time. In the Middle Caddo period, delta 13C apatite values ranged from a high of -13.9‰ at the Boyette site (feature 1/10) to -6.4‰ at the Lang Pasture site (feature 92). Apatite values were most similar during the Formative Caddo period, followed by those detected in the Historic Caddo period bioarchaeological samples. Differences between delta 13C collagen and apatite values also show considerable variability within time periods. This suggests that the amount of protein consumed by different Caddo groups may have been influenced by factors other than time. The greatest single difference in bone collagen and apatite, -14.6‰, was in an individual from the Lang Pasture site (feature 92), which suggests a low protein diet for this particular individual. The smallest difference in collagen and apatite values for one individual during the Middle Caddo period was also in an individual from the Lang Pasture site (feature 76C): namely -4.4‰. The smallest difference value, -1.2‰, between apatite and collagen isotope Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

107

values, and thus suggesting the highest protein diet, was from a burial (AL-2) excavated at the Late Caddo period E. W. Henry Farm site. The decrease in delta 15N mean values seen in the Historic Caddo period likely represents an increase in the consumption of beans. Delta 15N values show that the difference in protein amount and source increases through time. Prior to the Late and Historic Caddo periods, all delta 15N values are above 6.0‰, and most are below 10.0‰; these values are consistent with protein obtained by the Caddo from terrestrial-based ecosystems. Variability in delta 15N values more than doubles from the Middle Caddo period to the Late Caddo period. During the Late Caddo through Historic Caddo periods, delta 15N values range from a low of 1.6‰ at the Historic Caddo Jim Allen site to 13.7‰ at the Late Caddo O. L. Ellis Farm site; both of these sites are in the upper Neches River basin. These values are the highest and lowest delta 15N values for all of the Caddo time periods in the east Texas Caddo bioarchaeological record. These values also represent a large range of protein intake and sources, from beans to herbivore to fish species. The nitrogen isotope data suggests that the variability in protein source and intake increases through time within sites as well as between them. For example, delta 15N values from the earlier Hurricane Hill site range from 8.4‰ to 9.0‰, while delta 15N values from the Historic Caddo Jim Allen site range from 1.6‰ to 12.7‰. This indicates that through time the Caddo had a greater collection of dietary protein sources or that dramatically different amounts of protein were eaten.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Biotic Region Burnett (1990, 1999) and Rose et al. (1998) argue that maize consumption, based on caries rates, increased from a low in the Pineywoods to a high in the Blackland Prairie. The caries data presented here, however, suggests just the opposite. Stable isotopes show no significant differences among the regions in terms of maize and protein consumption (table 4-5). Delta 13C apatite and collagen values show that maize was a significant part of the Caddo diet in all three regions. Mean delta 15N values are consistent with the consumption of inland terrestrial protein sources. Unfortunately, bioarchaeological data was not available for the Post Oak Savannah Caddo populations for delta 13C difference values or delta 15N values. 108

Wilson

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 4-5. Stable isotope means in the Blackland Prairie, Post Oak Savannah, and Pineywoods

Biotic Region Blackland Prairie Post Oak Savannah Pineywoods

Denta wear

Delta 13C collagen

Delta 13C apatite

Delta 13C apatite-collagen spacing

Delta 15N Collagen

28.7 16.8 17.2

-16.07 -16.36 -15.95

-8.67 -8.10 -8.98

7.15 8.26 6.78

8.92 — 8.71

Copyright © 2012. Nebraska Paperback. All rights reserved.

Note: Dental attrition means are on posterior teeth following Scott (1979).

Despite the consistency in biotic region stable carbon isotope value means, there is considerable diversity in individual and site values within all regions except for the Blackland Prairie, where isotopic data are all from the Hurricane Hill site (Gill-King 1999). Isotopic signatures are most variable in the Pineywoods region, where most of the stable carbon isotope testing has been done to date. When regions are divided by time, there are important differences in the Early Caddo period, but the sample sizes are reduced considerably. In the American Bottom, the pattern of maize adoption is one of peripheral site adoption followed by large-scale adoption (Reber 2006). The significantly greater amount of maize and lower amount of protein consumed by the Early Caddo populations living on the Blackland Prairie may be an indication that the Caddo had a similar pattern of maize adoption. The two Early Caddo individuals from the Pineywoods have an average difference of collagen and apatite of 4.7 compared to 7.2 in the four individuals from the Hurricane Hill site (Blackland Prairie), indicating a lower protein diet on the Blackland Prairie. Nitrogen stable isotope values also differed, with the one individual from the Solon Stanley Farm (Pineywoods) having a value of 10.9 compared to the average of 8.8 for the four individuals from Hurricane Hill. Carbon collagen results were more numerous than apatite. Carbon collagen results were identical from the Post Oak Savannah and Pineywoods (-20.2‰) compared to the mean value of -16.5‰ from the Blackland Prairie sites. River Basin Table 4-6 illustrates that in terms of mean values for the river basins, only delta 13C collagen values suggest any difference in diet, but these proved statistically insignificant. Based on collagen results, more maize Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

109

Table 4-6. Stable isotope means in the east Texas river basins

River basin Cypress Creek Neches/Angelina Red Sabine Sulphur

Caries rate

Delta 13C collagen

Delta 13C apatite

Delta 13C apatite-collagen spacing

δ15N collagen

1.7 4.0 3.7 4.4 1.7

-18.07 -16.06 -14.55 -15.20 -15.89

— -8.98 -8.10 — -8.67

— -7.01 — — -7.15

— 8.71 — — 8.92

Copyright © 2012. Nebraska Paperback. All rights reserved.

was consumed by Caddo groups living in the Red River basin than in the Cypress Creek basin. Caries rates are also lower in the Cypress Creek drainage basin than in the Red River basin. Both the Red and Cypress Creek drainages had a similar range of delta 13C collagen values. For example, in the Cypress Creek drainage, individuals from the Late Caddo period Alex Justiss site had delta 13C values that ranged from -13.2‰ (burial 9) to -22.7‰ (burial 13) and in the Red River basin values ranged from -10.0‰ at the early Middle Caddo period Sanders site (burial 17-1) to -20.2‰ in an Early Caddo period burial (burial 23) from the Holdeman site. The range of delta 13C collagen results within the Caddo bioarchaeological population at the Alex Justiss site is unique among east Texas Caddo sites, and these results were interpreted to mean either that part of the population originated in locations where maize was not consumed or that there were gender or status differences between individuals buried at the site (Rogers et al. 2003:142). Stable isotope results are available from only one Caddo site in the Sabine River drainage: the Nawi haia ina site (41RK170). Here two individuals were tested for delta 13C collagen and none for delta 13C apatite and delta 15N. Stable isotope results are the most numerous from Caddo burials in the Neches/Angelina River basin. In terms of apatite results, there is a large range in the 38 available burial samples. These results ranged from -14.6‰ at the Early Caddo Solon Stanley site to -4.8‰ at the Late Caddo Fred McKee site. Within this range of apatite isotope values, 29 percent of the low maize values cannot be explained by time period and 17 percent of the enriched maize values occur earlier than the Late Caddo period in the Neches River basin. 110

Wilson

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 4-7. Stable isotope means for Caddo mound and non-mound sites

Type of site

Dental attrition

Caries rate

Delta 13C collagen

Delta 13C apatite

Delta 13C apatite-collagen spacing

Delta 13C collagen

Mound Non-mound

21.3 16.1

4.6 2.5

-15.55 -16.19

-9.35 -8.78

-8.15 -6.89

10.1 8.66

Copyright © 2012. Nebraska Paperback. All rights reserved.

Site Type For Mississippian populations, bioarchaeological studies indicate that social status sometimes meant differential access to meat and maize (Buikstra and Milner 1991; Schurr 1992; Wilson 1997b; Ambrose et al. 2003). In this volume, Jackson et al. (chap. 3) show that elite members of Caddo society had greater access to preferred cuts of meat and more exotics in their diet. In a cursory examination of Caddo burial locations in mound and non-mound sites as an indicator of social status, I discerned no apparent difference between mound and non-mound contexts in terms of the diet of the deceased Caddo individuals (table 4-7). If the Caddo elite had access to different types and cuts of meat, this study suggests they did not consume significantly more meat than nonelite. While caries rates indicate a higher consumption of sucrose-rich carbohydrates in individuals buried in mound contexts, stable isotope data does not show a difference in maize consumption between burials from mound and non-mound contexts. This differs from Buikstra and Milner’s 1991 finding of greater maize consumption at civic-ceremonial sites in the central Mississippi valley and less maize consumption at mound sites in the central Illinois valley. It is clear, then, that this difference between caries rates and stable isotopes among the Caddo needs further investigation. Few individuals from mound contexts have been tested for stable isotopes, and thus the results are at best provisional. Only 8 individuals from Early and Middle Caddo mound contexts have delta 13C apatite and collagen results. Only 2 individuals from mound contexts had results for delta 15N compared with 29 from non-mound contexts. Delta 15N values and the difference between delta 13C collagen and apatite values suggest that only very slightly more protein may have been consumed by Caddo individuals interred at mound sites. When broken down by time, the Early Caddo mound burial has the same delta 15N value as the Early Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

111

Caddo period average, but the Middle Caddo mound burial has a delta 15 N value 2.4‰ higher than the Middle Caddo average. Summary

Copyright © 2012. Nebraska Paperback. All rights reserved.

For the Caddo, maize has been an important component of their cultural identity in both the archaeological and ethnohistorical records. Food is both a cultural symbol and a necessity for life. Humans create and pass on food lore, creating stories and myths about food’s meaning, origin, and taste. We celebrate rites of passage and religious beliefs with distinctive food dishes (Gabaccia 2002) and often through feasting. There are numerous examples of the important role that maize played in Caddo culture. Sagamite, a cornmeal mush, was served when people were called to assembly, and extensive rituals took place surrounding the ripening of the first corn (Carter 1995). This study has tried to examine to what extent the Caddo actually ate this symbol-laden food. I have used three forms of bioarchaeological evidence to reconstruct the Caddo diet: dental wear, dental caries, and bone chemistry. Three conclusions can be made from this research. First, dental wear differed significantly among the Texas Caddo. Second, delta 13C apatite values show increases in maize consumption in the latter Caddo sequence. Third, variability was the norm for Caddo diet and subsistence strategies, both in terms of the isotopic signature of maize (presumably type) and the amount of maize consumed. Dental wear patterns indicate that there was a change in the coarseness and/or grit in the Caddo diet. Dental attrition reflects food processing and the coarseness of dietary items. Dental wear should decrease as the amount of coarse items such as hickory nuts and acorns decrease in the diet and as softer items such as maize become a more important part of the Caddo diet. There is significantly more dental wear in Caddo populations from the Blackland Prairie than those populations living in the Post Oak Savannah and Pineywoods of east Texas. Grinding stones were found sporadically in the Blackland Prairie Caddo sites examined here (Fields et al. 1997), so it does not appear that the difference is in food processing tools. This portion of the Caddo region of Texas had a relatively poor natural resource base compared with other physiographic zones, since it is dominated by large expanses of tall grass prairie (Diggs et al. 2006). The caries rate is the lowest in this region, a finding that suggests maize was less important. However, stable isotope data shows 112

Wilson

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

that maize was important in the diet of Blackland Prairie Caddo groups who lived in the Sulphur River basin during the Formative and Early Caddo periods. The explanation for the difference in Blackland Prairie wear scores is likely the result of the higher coarseness of the diet in the form of more seeds and nuts, not one associated with less maize in the diet. Changes in dental wear through time are significant as well. In fact, rates of dental wear drop significantly from the Formative Caddo to the Early Caddo period, followed by another large drop from the Early Caddo period, with wear scores decreasing 35 percent over these periods. The Historic Caddo period bioarchaeological samples are represented by the lowest wear scores (11.9), all of which come from burials at the Jim Allen (41CE12) site. In Caddo regions outside of Texas, Rose et al. 1984) also found a significant drop in dental wear from the Fourche Maline or Woodland period to later Caddo time periods. Stable isotope results indicate that the decreased dental wear from the Formative Caddo through Middle Caddo periods is not the result of an increase in maize consumption. Instead, coarser nuts may have been replaced by softer seeds or seed or nut processing increased during this time. This was coupled with an increase in bean consumption. Caries data was somewhat less informative. Using caries data, Burnett (1990, 1999) hypothesized that maize agriculture decreased in an easterly direction across the Caddo archaeological region of east Texas. Caries rates in this study do not corroborate Burnett’s findings, instead showing two or more times the caries rate of the Blackland Prairie Caddo in populations farther east. It is possible that the high rate of dental wear in the Blackland Prairie populations accounted for the lower caries rate documented in this study. In an examination of Caddo populations outside Texas, Rose et al. (1998) found that carbohydrate consumption, based on caries rates, varied until about A.D. 1100, when it was uniformly high. In this study, variability in caries rates per site continues throughout the entire temporal sequence. The greatest variability is in the Late Caddo period, which also has the largest number of sites (n = 14), when caries rates range from 0 to 13.0, followed by the Early Caddo period. Since variability increases with increasing numbers of sites, it appears that dietary variability characterizes the Caddo subsistence strategy in east Texas. Delta 13C collagen results indicate that maize was present among Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

113

Copyright © 2012. Nebraska Paperback. All rights reserved.

the east Texas Caddo earlier than was the case in southwest Arkansas, northwest Louisiana, and southeastern Oklahoma. Of the two Woodland or Formative Caddo period collagen results, one (George C. Davis, feature 155) shows that maize, if present, was an insignificant part of the diet. The other, burial 15 from the Hurricane Hill site, suggests that maize was an important part of the diet of this South Sulphur River Caddo group. Nearly all of the Early Caddo collagen results demonstrate that maize constituted some part of the diet. Adjusted for sample size, standard deviations on delta 13C collagen results within sites confirm that the Caddo had a varied diet. Only one site, Nawi haia ina, had an adjusted standard deviation (0.2) consistent with a homogenous diet (see DeNiro and Schoeninger 1983). Only three other sites had diets that would be characterized as minimally varied (George C. Davis, Hurricane Hill, and Jim Allen). An examination of delta 13C values on maize from east Texas sites demonstrates dietary variability, but not to the same extent as seen in the humans who were eating it. For example, delta 13C maize values were only minimally varied at the George C. Davis (41CE19) site (Perttula 2008). Overall mean deviation is greater among and within Caddo sites than Mississippian sites, suggesting a greater flexibility in the Caddo diet or a greater range in delta 13 C values for maize. Apatite more precisely depicts changes in the amount of maize consumed in the diet. Delta 13C apatite values increase after the Middle Caddo period and again in the Historic Caddo period. Apatite values also demonstrate that maize appears to have been present in the archaeological sequence earlier in east Texas than in other Caddo regions, but agricultural intensification occurred in east Texas later than it did in southwest Arkansas, northwest Louisiana, and southeast Oklahoma Caddo sites. Does classifying a group of people by subsistence mean all members should ingest that food to the same degree? The Caddo were a complex society, and with complexity came differences. In east Texas, stable isotope results show an increase in maize and bean consumption with intra- and inter-site variability. Among Mississippian populations, maize consumption has been shown to vary locally, regionally, by status, and diachronically. A similar temporal pattern is apparent in the Mississippian diet, but stable carbon isotope values are lower among the Caddo once maize became a significant part of the subsistence regime. This implies a greater reliance on maize among the 114

Wilson

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Mississippian populations or a difference in the isotopic signature of maize with Mississippians consuming a more C4-enriched variety. There is evidence from at least one Caddo site that different varieties of maize had different isotopic values (Perttula 2008). Environmental factors also influence isotopic values. For instance, there is a difference between maize values from the middle Ohio River valley that average -8.5‰ (Greenlee 2006) and maize from east Texas with an average delta 13C value of -10.9‰ (Perttula 2008). Delta 13C apatite results are most variable during the Early, Middle, and Late Caddo periods. During the Middle Caddo period, for instance, these values range from -13.9‰ to -6.4‰. Apatite results from the Middle Caddo period are from five sites in the Neches/Angelina River drainage (Pineywoods), showing that these differences are not related to differences in the local biotic environment. Delta 15N collagen values are most variable during the Historic and Late Caddo periods; the Historic Caddo period results are all from the Jim Allen site in the upper Neches River basin in the Pineywoods. When looking at the two best sampled sites for stable isotopes, Alex Justiss and Lang Pasture, it is apparent that there is considerable variability within Caddo sites as well as between them. While there is less difference in the stable isotope results from the Lang Pasture site compared to those from the Alex Justiss site, delta 13C apatite values nevertheless range from -10.8‰ to -6.4‰. This contrasts with results from the Hurricane Hill site, which are generally similar to one another. At the Jim Allen site, there is less difference in stable carbon isotope values than in delta 15N values. At the Alex Justiss site, Rogers et al. (2003) suggested that the great differences seen at that site in delta 13C collagen results are a product of status or gender-based differences in diet or population mixing. In this study, status based on burial location (mound vs. non-mound contexts) was not found to be a significant factor in the character of the Caddo diet; gender and population affinities were not examined as part of this study. In a study on population affinities, Lee (1999) found the Caddo to be a distinct population from, but related to, Central Texas hunter-gatherers and distinct from Mississippian populations. Since dietary differences do not appear to be macro-environmentally related after the Early Caddo period, they can be explained as either small-scale temporal variations in subsistence practices or personal choices. In her 2006 examination of dietary variability in the middle Ohio Evidence of Subsistence Strategies

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

115

valley, Greenlee ruled out short-term drought, soil conditions, and macro-environmental depletion in favor of micro-environmental proximity to resource-rich areas in order to explain why some individuals consumed less maize than others. A more thorough investigation of local environmental circumstance is warranted for Caddo bioarchaeological studies. Conclusions

Copyright © 2012. Nebraska Paperback. All rights reserved.

In his review of the Caddo paleobotanical record from east Texas, Perttula (2008) found geographic variability in the amount of maize consumed that could be attributed to rainfall and soil productivity and that maize is less common in east Texas Caddo sites than in those in Louisiana and Arkansas. In contrast to his results, stable isotope values indicate maize was consumed in greater quantities among the earlier Texas Caddo populations than in Caddo regions outside the state, followed by similar consumption patterns later in the archaeological record. Certainly more examination of bioarchaeological and paleobotanical data from Caddo sites is needed to address this issue. Clearly the dietary contribution and likely symbolic importance of maize in the Caddo diet increased through time. The power of food necessity and symbolism can be seen in the more recent history of the Caddo during U.S. control of Native American culture systems. With the deliberate destruction of traditional food ways, including bison, and other ecological destruction and land dispossession, there came the nutritional impacts of the United States’ relationship with Native Americans, which was seen in starvation and malnutrition (Miewald 2002). The impact on Native Americans can be seen today in the form of chronic debilitating diseases (including diabetes) and obesity as a result of forced dietary changes.

116

Wilson

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

5. Spiro Reconsidered Sacred Economy at the Western Frontier of the Eastern Woodlands james a . brown

Copyright © 2012. Nebraska Paperback. All rights reserved.

Introduction In the summer of 1935 an unparalleled collection of fancy goods and exotic materials was pulled from the large central feature of a hitherto obscure mound lying at the western edge of the Mississippian cultural world (Clements 1945; Orr 1946). The abundance of copper and marine shell objects with imagery well established from important sites located over 600 km to the east promoted Spiro as a hitherto unrecognized center of social and political importance during the Mississippian period (Waring and Holder 1945). The quantity of elite goods from the site far exceeded that known elsewhere. For instance, the number of distinct engraved marine shell cups on record from Spiro’s central mound feature is more than 100 times the total known from all contemporary centers (Muller 1987; Phillips and Brown 1978:24–25, 1984:xi, xvii [third section]). The abundance of other highly crafted goods carrying rarely found imagery lent credibility to the site’s importance over a broad area (Bell 1947a; Brown 1984a). This scale came to imply the existence of a chiefdom level of social and political organization (Brown 1996; Rogers 1983, 1996, 2006; Wyckoff 1980). Thirty years of research have been invested in the proposition that Spiro was the center of a chiefdom. But an updated review of the Great Mortuary and the central cone of the Craig mound in which it rests has disclosed this feature to be a starkly different kind of monument than heretofore contemplated. A picture has now been constructed that points in an entirely new direction: that this monument employed human remains and spectacular objects to create a “center of the universe.” Reanalysis shows that the entire main cone of the Craig mound was a 117

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

5-1. The Craig mound from the northwest during the winter of 1913–1914. (From Thoburn and Wright [1929:2:896]. Courtesy Oklahoma Historical Society.)

unified monument whose main purpose had nothing to do with the marking of the status of elite dead (fig. 5-1). Reanalysis makes clear that the kind of status-marking treatment of the dead is represented weakly, if at all, in the Great Mortuary. The massive deposits of exotic goods and elaborately treated skeletal remains in the Great Mortuary are not the indicator of elevated social status as once thought. Spiro as a Chiefdom

Copyright © 2012. Nebraska Paperback. All rights reserved.

The cultural connections invoked by the special imagery found on certain artifacts at Spiro thrust the site into the limelight and elevated what had been a minor site into a place of considerable social and cultural importance (Brown 1996; Clements 1945). The first response was to gravitate toward a narrative that tied the Spiro images to similar materials found in the elite burials at the distant Etowah and Moundville sites, in Georgia and Alabama, respectively, as well as sites as far removed from it as the east coast of Florida. In the compressed chronology of the 1940s a special “event” likened to the Ghost Dance and called the “Southern Cult” was devised to account for the sudden, seemingly abrupt, appearance of elite goods in highly exclusive locations. Very different regional histories were linked to this widespread event by the presence of this Southeastern Ceremonial Complex, none more different than the cultural and economic background of Spiro, an outlier in many other respects (Waring and Holder 1945). In the rush toward a narrative unifor118

Brown

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

mity Spiro’s distinctiveness was brushed aside despite Krieger’s (1945) insistence otherwise. The presence of look-alike objects among the participating centers of the “cult” suggested some sort of elite exchange, but in their defining article on the subject, Waring and Holder (1945) explicitly rejected trade as relevant. The social and economic contexts of Spiro were overlooked in favor of the dramatic appeal that a cultural revitalization model brought to the subject (Krieger 1945). In short, a spectacular collection required a commensurately dramatic rationale. The revitalization cult model required some reconsideration when closer examination of the archaeological contexts of Spiro found logical precursors to the flashy mortuary contexts of the central feature (Baerreis 1957). Trends toward increased attention to the status in the treatment of the dead led the way toward recognition that a chiefly evolutionary model might apply in the case of Spiro (Baerreis 1957; Brown 1971, 1975, 1981). Burial practices implied the kind of social complexity associated with political chiefdoms (Brown 1971, 1981; Phillips and Brown 1978:1–22). The central feature became recognized as an arbitrary section of a larger structure aptly called the Great Mortuary on account of the size of the level space given over to the placement of stored goods and burials (Brown 1966). The sphere of exchange implied by the large array of materials lent support to this political interpretation (Brown 1983; Wyckoff 1980). Likewise, settlement patterns and ethnographic analogy seemed to concur in favoring a chiefly political organization in the absence of much knowledge about the political economy (Brown et al. 1978; Wyckoff and Baugh 1980). Charnel House Model The substantive gap in the economic model was plugged by conceiving of the charnel house at the Great Mortuary as an economic institution based on its later role during early historical times (Brown 1990, 1991). A relevant model was introduced by Jonathan Friedman (Friedman and Rowlands 1978) in his prestige goods thesis: “The model makes use of long-distance exchange to anchor the top of the wealth pyramid in exotic artifacts. But those that do depend on the monopoly that a dominating elite have on extrasocietal exchange relations with distant peers” (Brown et al. 1990:255). That mortuary houses were used as storehouses of rare items was documented by Natchez and Powhatan ethnohistorical examples. These Spiro Reconsidered

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

119

Copyright © 2012. Nebraska Paperback. All rights reserved.

were drawn upon to posit a model of a political economy of the prestige goods type that was centered on the Great Mortuary ancestor charnel house: “A pattern of central concentration of widely distributed wealth items fits the expectations of a system in which strategically [the] elite have a decided advantage in manipulating the flows of exchange or in controlling the means of production and access to scarce resources” (Brown 1996:24). The Great Mortuary was visualized as a particularly large and rich ancestor shrine that was unique in the entombment of its contents in place. The sample of comparable deposits would have been more plentiful if, it was supposed, these deposits had not been removed to a separate location for placement in a mound dedicated for the purpose. The cycling concept was taken from Bell’s 1972 proposal that grave goods that ended up in the Harlan site accretive burial mounds were taken from their associated mortuary houses. Thus what ended up as grave goods started out as objects stored in a mortuary house: “The Spiro mortuary house served as a ‘cult strong house’ filled with wealth items and weaponry useful in exchanges, political rewards and prosecution of war. The marine shell cups and beads, together with the copper plates and ax heads, probably functioned as money-like reservoirs of value” (Brown 1996:23). The social standing of the dead was indexed according to the local value of the grave goods. The goods themselves were accumulated in the mortuary at a rate that indicated the political and economic power of the group claiming control over them. Consequently, the scarcity of the more precious of these valuables and most of the jewelry could be translated into a statement about social class and the political power of centralized leadership (Brown 1981, 1991). The merits of this model notwithstanding, the Spiro archaeological record falls short in fulfilling important expectations about the sociopolitical complexity of a chiefdom by questioning its centralizing function as a node in reciprocal long-distance trade. The local productive economy was relatively undeveloped by comparative standards, and no outflow of goods can be readily identified. These flaws can be summed up in the following terms. The abundance of high value goods in the complicated mortuary tableau of the Great Mortuary is not necessarily the product of reciprocal trade, and 120

Brown

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

the complexity of the mortuary program should not be confused with complexity in social relations. The Great Mortuary burial practices can no longer be used to support a model for social or political hierarchy. In its use of skeletal remains, this facility bears closer resemblance to the ossuaries of the Iroquois Feast of the Dead than to the hierarchical treatment of the dead exemplified by Natchez practices. The collaboration of many individuals inherent in the creation of a collective monument scrupulously laid out to exacting cosmological principles calls for some theoretical underpinning.

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Sacred Economy The “sacred economy” is a model that stands as an alternative to the prestige goods economy (Renfrew 2001). The Great Mortuary and the totality of the main cone of the Craig mound in this kind of economy becomes a ritual-architectural object of religious allurement (Jones 2000a, 2000b) or, in Renfrew’s terminology, a “Location of High Devotional Expression” (lhde). In the Americas such a monument is conceptualized as a “center of the universe.” The Spiro version of this sacred monument worked as a magnet for sacred objects and elite goods. It was designed to draw and integrate a scattered and ununified population. The accumulation of goods mimicked the accumulations of the prestige goods economy, but whereas valuables in the latter economy were given to subordinates to cement allegiance in the prestige goods economy, in the sacred economy they remained concentrated to fulfill the potential spiritual power resident in the monument. Construction of a “center of the universe” is an act whose significance goes far beyond the energetics of heaping earth at some designated location. What can make earthen construction transcend mere engineering achievement is the impact of acts of devotion (Renfrew 2001). Construction becomes a kind of ritual act in itself, and it is not surprising that mound construction has less to do with straight muscular effort than the devotional contributions of the multitude. In many cases the construction process of mound building has incorporated materials that can only be acquired through extravagant means, with distant transport, dangerous acquisition, and outsized talent and determination. One of the “payoffs” is that human labor can be converted into karmic merit through investment in permanent sacred architecture, whether with earth or more durable materials. Such merit, acquired either by laborious effort Spiro Reconsidered

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

121

or by difficult means, becomes the measure by which devotion is gauged (Jones 2000a, 2000b). Each version of a specific universe requires an appropriately determined center, an axis mundi. The creation of a culturally determined image of this universe, or imago mundi, stands to capture the attention, use, and allegiance of a scattered and dispersed population. In Lindsay Jones’s terms, the Great Mortuary represents a “center of religious allurement” (Jones 2000a:75). Eliade (1959) has elaborated on the appeal of such centers. To quote Jones: Accordingly, Eliade (and many who follow his lead) describes the appeal of certain places and structures in terms of a “nostalgia for Paradise,” an almost instinctual “desire to be always, effortlessly, at the heart of the world, of reality, of the sacred” [Eliade 1958:383], a thirst to transcend the human condition and regain a divine state of affairs like that which people had enjoyed prior to the Fall, in the “Time before time” [Eliade 1958, 1959:54–55]. Yet, because that exhilarating experience of transcendence can be effected, not everywhere, but only at those houses, temples, palaces, and cities that are conceived as being situated at the “centre of the world” (that is, at those hierophantic places where “the sacred” has made itself accessible), homo religious are inevitably drawn—or, in my term, “allured”—to those special places. . . . What I call ritual-architectural allurement is, then, from Eliade’s perspective, a consequence of the supernatural activity of the sacred rather than of human builders or ritual choreographers ( Jones 2000a:75)

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Center of the Universe embodies the mythical archetypical concepts of the axis mundi. Direct access to the sacred was achieved by creating a spiritual connection between the earth and the heavens. By inscribing the immediate environs materially, an image of the universe is created that replicates that world. Each cultural tradition has its ways of materially inscribing the center. One of the common resources for achieving this world-making is to use the remains of the dead, thereby securing the favorable attention of the ancestors or, more accurately, the group’s forebearers. The dead are thereby instrumental in creating a cosmogonic tableau (Parker Pearson 1999). Renfrew (2001:22) conceptualized the “sacred economy” as a special kind of transaction: The devotee comes to the special ritual center and observes or takes part in some special rituals. Thereby the pilgrim discharges an obligation and receives the benefits of spiritual enlightenment or religious experience, or perhaps simply entertainment. 122

Brown

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

The benefits are (in a general sense) reasonably clear. In straight economic terms the visitor has the benefit of a “performance” just as much as going to a theater or a cinema. Renfrew elaborates on the outcomes in which the center is integrated with a power system. An obligation of labor services, food, or gifts may be forwarded. When the sacred economy is not integrated, a payment to the religious specialist for services received is made or an offering to the deity is made, etc. Renfrew (2001:22) sees as a consequence that “a very substantial inflow of material goods is to be expected, generally including valuables and other offerings suitable to the presiding spirit at a LHDE.” He adds that the lhde “invites consideration of the ‘sacred economy’ as one in which a number of generally applicable economic assumptions may not hold. . . . For the incoming resource takes the form of material goods or of labor, while the counterpart of the exchange . . . is entirely nonmaterial. The counterbalancing payment is accepted in the form of religions experience and the accompanying social experience for the individual of participating in a great seasonal fair.”

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Spiro Site and Its Peculiarities The sacred economy entails turning on its head the proposition that social stratification and political centralization is indicated by the accumulation of valuables in the context of socially exclusive mortuary display (cf. Spielmann 2007). The splendor of the Great Mortuary is the central fact. One would not envision any political centralization at the Spiro site without the existence of the Great Mortuary. Why is that? The Spiro site is spread over a high terrace and low hills flanking the Arkansas River valley at the point in the valley where southeastern vegetation begins to be replaced by open grassland. Spiro was relatively small, and its unfavorable location at the margin of the wellwatered woodlands of the Southeast translates into reduced subsistence productivity as well (Brown 1996:29–32, 37–39). The wooded borderlands would have been vulnerable to variation in rainfall patterns and the resource shortfalls that would ensue. Such a climate must have exercised a drag on the local economy of food production. Spiro simply was not as well endowed as lands to the well-watered east (Brown 1984b; 1996: 1996:29–32, 37–39). Contemporary habitation debris is not abundant, and the density of occupation raises doubt as to whether the productive economy was Spiro Reconsidered

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

123

Copyright © 2012. Nebraska Paperback. All rights reserved.

sufficiently substantial to support a political hierarchy in the first place. Spiro was the largest settlement in the region, composed of a group of 11 mounds with a surrounding occupation of 33 hectares (ha). Occupation areas were distributed in relatively small areas of about 1.5 ha in size within a 3 km radius of the site. Works Progress Administration (WPA) excavations disclosed that each of these patches consisted of a cluster of structures. Substantial cemeteries were placed near these habitation areas (Rogers 1995; Rohrbaugh 1982, 1984). All told, the density of population appears to be rather small to support a political hierarchy in which leaders are honored by sumptuous mortuary rites. Although the distribution of settlement takes on a hierarchical form to some extent (Brown et al. 1978), this settlement pattern does not necessarily translate into political hierarchy. With the exception of sites in the Fort Coffee neighborhood of Spiro, the known site density is sparse (Brown 1996; Orr 1946; Rohrbaugh 1982, 1984). More damaging, there is no evidence for centralization of any economic activity. Of course, at any one instant in time the distribution of contemporary sites was probably even sparser. In sum, not only was the resource base weak and unstable, but a centralized political structure is difficult to support considering the general sparseness of settlement. The single possible exception to this portrayal takes place around the Spiro center during the Spiro phase. Cemeteries display greater complexity in disposal patterning than the patterning seen later during the Fort Coffee phase. In addition, specially designed buildings made during the time contributed toward greater complexity in settlement than was seen either earlier or later (Brown 1996; Rohrbaugh 1982, 1984). Their appearance in the vicinity of Spiro around 1400 suggests that the Great Mortuary event was not disconnected to other social and economic changes. As a sacred monument the Great Mortuary is likely to have been a significant stepping-stone to chiefly organization. But it was not a material representation of that complexity, which if the archaeological record of the Fort Coffee phase is indicative, such a polity does not seem to have endured more than a generation. The social changes that evidently took place after the Great Mortuary raise the likelihood that its construction was a prelude to launching a chiefdom. But as should be clear from the preceding discussion, the Great Mortuary does not constitute a map or other kind of representation of social hierarchy. 124

Brown

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Great Mortuary as a Sacred Monument The record has to be set straight by presenting all of the information that we have to date on the context and internal organization of the Great Mortuary. We have drawn on a site history created indirectly to make up for the absence of a systematically recorded stratigraphic record. The sample was made up of radiocarbon-dated burial and domestic assemblages from the immediate region. A sequence of four grave periods— with finer distinctions in two of them—provides the chronological framework herein. A grave lot sequence order was created that divided the graves into four periods from Spiro I through Spiro IV (with various subdivisions) was developed through a grave lot sequence order. The effect of the grave lot sequence has been to sharpen the cultural material distinctions between the grave periods advanced previously on less complete data (Brown 1996). This created a cultural periodization buttressed by a sequence of ordered grave lots and a limited number of radiocarbon dates (Brown 1996). A radiocarbon date age in the early fifteenth century corresponds with dates from village deposits in the Spiro vicinity (Brown and Rogers 1999; Rogers 2006). The Great Mortuary and earlier mortuary deposits in the Craig mound at Spiro occupied a space that was bracketed by a string of preexisting accretive mounds. One stood to the north and three stood to the south. Sharing this space was a large crematory basin of nearly 5 m in diameter. The accretive mounds were low structures built up through the addition of one layer of burials on top of another, each layer separated by sheets of cedar bark and earthen fill. Their construction took place in Spiro I, Spiro II, and perhaps Spiro III times as defined in chrono-stratigraphic terms. The accretive burial mound components in the Craig mound are of the type exemplified by the Groseclose mound (34Mi45), on the Canadian River, and the Harlan burial mound (Brown 1996:179–180). The Craig mound mortuary record departs from these examples by two major features: the bermed enclosure containing the Great Mortuary, and the giant Crematory Basin. Both probably sustained repeated use over a long time rather than at some specific period only. By the time the Great Mortuary tableau was constructed in the early fifteenth century, the mortuary display area combined new forms of burial treatment with preexisting ones. Layers of mortuary remains and Spiro Reconsidered

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

125

Copyright © 2012. Nebraska Paperback. All rights reserved.

precious goods rested on a well-constructed, prepared floor (Brown 2010). Within a rectangular enclosure of about 12 × 18 m in size, flooring of split cane was laid longitudinally to a northerly direction, oriented just west of true north. This floor lay within a space defined by 2 m high dirt embankments or berms, apparently not unlike ones found elsewhere in the Caddo area (Kay and Sabo 2006; Sabo 1998). Two layers of burials have been identified on the floor. The lower layer was an ossuary mass of disarticulated individuals thoroughly mixed with broken artifacts. Upon this ossuary was another distinct layer that consisted of either litter burials or partly disarticulated burials excarnated or stripped of flesh through exposure. The litters, which contained a very condensed set of human remains, were oriented in conformity with the direction of the underlying split cane flooring (Brown 1996). Deliberately laid within the upper burial layer were cult objects (mainly massive effigy pipes), many sacred bundles consisting of twillwork baskets (often containing copper plates), and piles of textiles and shell beads (Brown 1996, 2010). Most of the copper plates attracting the greatest museum attention were, in fact, ones that were housed in flat, rectangular, and lidded baskets. An important aspect of this layer is the deployment of complete objects rather than single fragments. They are whole or nearly whole shell cups, copper plates boxed within bundles, stacks of textiles carefully folded into individual packages, heaps of pristine shell beads of many shapes and sizes, chipped stone maces, and monolithic stone axes. The contrast between the two layers in completeness provides a glimpse into a distinctive cultural code. Not only were objects treated differently, but divergent means by which these objects were acquired and brought to Spiro also seems to be at work. The symbolic use of human remains does not end with the completion of the tableau on the floor. A vault-like cover appears to have been created from upright posts of cedar. Traces of this structure have been identified. Upright cedar posts planted in the deposit floor poked through a low dome of earth (Brown 2010; Duffield 1973). Posts held this earthen structure in place. The structure only covered the heart of the tableau. In the corners of the berms that framed this dome, human remains were deposited together with broken artifacts. Additional posts were planted as more heaps of earth and disarticulated human bone added height to an already substantial mound. At the close of this cycle, human remains were placed in discrete layers interleaved with fill to a 126

Brown

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

height of 6 m within the ascending mass (Spiro IVC through IVJ). (To continue with my stratigraphic nomenclature in the first Spiro site report, see Brown 1996.) These burial layers formed a cross with the broad short arm positioned across and over the south end of the tableau and the long sparsely marked arm oriented northward. A final cone of dirt completed the mound to a height of 10 m.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Orientation and the Axis Mundi The arrangement of sacred objects and human bone can be likened to a layer cake. The ossuary was the first layer, the second was the layer of litter burials and cult objects, and the third was the laying down of small amounts of disarticulated remains without included objects. The floor of entirely disarticulated human remains mixed with broken and crushed artifacts can be thought of as a votive prelude to the tableau created by litters, sacred box bundles, whole cult objects (effigy pipes and mortuary statuettes), and other complete objects. The broken objects and disarticulated remains placed above the Great Mortuary floor deposit repeat the offertory aspects of the first layer of the ossuary remains. A pole planted in the floor provided the initial axis mundi; after it was removed, this role was taken over by the lapped seating of cedar poles higher in the mound cone. To cement the continuity with the floor, disarticulated remains in the higher elevations were laid out in the form of a gigantic T with the long thin arm oriented north. Artifacts, however, are absent in the upper elevations. A third pair of sacred objects was placed at the edge of the mound near the stem tip of the T. The totality of the burials in the main cone of the Craig mound becomes a consolidated offering to spiritual powers from above. The component layers were united by an axis mundi. At the beginning of Great Mortuary construction, this axis was materialized by the erection of a single pole. A deep open pit (LfB80) marked its location at the north end of the tableau, just south of the well-known redstone “Big Boy” effigy pipe. Evidently it held a large upright pole that was planted and subsequently removed during the buildup of the lower burial layer. Fragments of engraved shell and other materials from the surrounding burials had spilled into the open pit (Brown 1996:704).1 After its removal, proxy poles were planted in the mound as it rose in height. The use of upright cedar poles created a spiritual conduit up into the summit of the mound. Complementary to the zenith direction was another Spiro Reconsidered

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

127

materially marked north-south orientation. This second axial marker is defined on the north by a pair of effigy pipes at the base of the upright pole. This attention to the north-south alignment was evidently continued by a third pair of effigy pipes that was stashed beneath the northfacing surface of the mound above the Great Mortuary floor. This pair ties into both the vertical and north-pointing axis. The earthwork was conceived and probably constructed as a single, unified project. The Great Mortuary and the entire display of mortuary remains stacked within the main cone was the grand finale at Spiro. The entire main cone that includes the Great Mortuary belongs to the end—or near end—of the mound sequence. The construction of the Great Mortuary tableau presumably coincided with social changes that ushered in the Fort Coffee phase (Rogers 2006; Rohrbaugh 1982, 1984). The Great Mortuary as Imago Mundi

Copyright © 2012. Nebraska Paperback. All rights reserved.

The erection and removal of a substantial pole that provided coherence to the rectilinear plat of litter burials and shrine objects assumes a key role in turning the burial patterning on the floor of the Great Mortuary into the basal tier of something like a sacred “mountain” with coded reference to the sacred universe within the middle layer. The disposition of very substantial, subject-specific, and color-coded objects, from a geographic spread of locations, testifies to cosmological principles involved in their choice rather than the product of economic exchange activities (fig. 5-2). Certain cult objects were aligned northsouth along the Great Mortuary center line. At the north end of this centerline were placed two stone pipes (the birdman effigy pipe and the rattlesnake-stemmed elbow pipe [Hamilton 1952:34–35, 39 and plates 9-10 and 22b]). From what can be gleaned from firsthand accounts, the birdman pipe was positioned directly north of a substantial, evacuated post pit large enough to be filled with materials from the lower burial layer. Evidently the post that once had stood erect at this location had been pulled by the time the tableau was sealed. Immediately to the south was a set of seven chipped and polished stone maces (Hamilton 1952:42 and plates 35–41). The number seven is a sacred number, and it is not coincidental that the mace club is the armament consistently associated with the image of the Birdman, and the presence nearby of this cache could well have been in recognition of a close conceptual association. Positioned next to the central stack of textiles were an additional two 128

Brown

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

5-2. The Great Mortuary floor showing the approximate location of the major cult figures and other furniture recovered by the relic hunters among the litter burial constructions and other features uncovered by the wpa. “S” represents the location of the cache of Olivella dama beads. (From Hamilton 1952:plate 5; floor plan from Brown 1996:1–30.)

massive effigy pipes, their relative positions being unrecorded. These were the earth mother pipe and the raptor-over-recumbent-human pipe (Hamilton 1952:34, 36 and plates 6 and 15). To the south lay three wooden human figurines (Hamilton 1952:40 and plates 25–26). The heads of an additional two figurines were recovered within a gravelot (LfB108) from the WPA operation (Brown 1996:531–532). These cult objects created a line of paired pipes and human effigy figures extending across the Great Mortuary in a north-south line. No other information is available because all of the objects from the hollow chamber were grabbed during the relic grubbing summer of 1933 with the barest of locational details. This centerline separated the eastern side objects from the western ones. To the east of this line lay many of the sacred basket boxes containing copper plates and other items. These were plotted as the “thirty copper axes” and the “ten baskets with sheet copper” (Hamilton 1952:Plate 5). The bag or basket containing the Olivella dama beads from the Gulf Spiro Reconsidered

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

129

Copyright © 2012. Nebraska Paperback. All rights reserved.

of California was an eastern deposit (Brown 1996:721, 2004; Kozuch 2002). The basket of small wooden masks was a western deposit. In sum, the outline of a ritual tableau emerges that was centered along a (nearly) north-south axis. The cult objects as effigy pipes referenced here are complete and prominent in size. Seemingly different in disposition are the large, engraved, marine shell cups that are whole or nearly so. They were sufficiently numerous to be reported as encircling the hollow chamber floor. The reported disposition of this mass of shell cups need not be treated as credible, but the number was sizeable. Philip Phillips counted about 100 reasonably complete cups in a corpus he estimated contained at least 1,000 cups with unique decoration (Phillips and Brown 1978:24–25, 1984:xi, xvii [third section]).2 The latest style phases (Braden C and Craig C) dominate; the next most numerous is the earlier Craig B; whole cups in other styles are present in small numbers (Phillips and Brown 1978, 1984). The objects themselves constitute an important part of this tableau. Bearing in mind that a complete inventory of the Great Mortuary objects cannot be assembled, certain patterns do emerge that help us understand the contents primarily as an assemblage deliberately chosen to activate sacred forces. The complete copper plates were key components of sacred bundles (Kuttruff 1993); the textiles were made freshly for the event; the massive stone sculptured pipes and wooden shrine figures were arranged pairwise in a north-south alignment; and chipped stone maces were displayed in front of the redstone pipe Morning Star figure (“Big Boy”) (Brown 2010). The imagery embodied in the pipes states a cosmological message created specifically for the Great Mortuary (Brown 2010). Four pipes were placed in pairs. At the north end birdman and snake imagery composed one pair (Hamilton 1952:Plate 5). Near the center an image of earth mother was paired with a raptor embracing a recumbent human (Hamilton 1952:Plate 22b). The choice of images was presumably important. The northern pair embodied the opposition between the giver of human life and the taker (Brown 2007b; Brown and Dye 2007). The central pair represented the source-of-fertility figure (Emerson 1989; Prentice 1986) and a spirit that could be interpreted as representing death and transformation. The pairing of effigy pipes was repeated later in the Craig mound sequence. High in the north face of the main cone (in a location, LfA9, now lost) a pipe-smoker figure and a seated effigy were cached together just beneath the mound surface. 130

Brown

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Each of these pipes encoded color symbolism: the pair in the mound cone are white, the northern Great Mortuary pair are bright red, and the centrally located pair are divided between snow white and black. These color patterns are surely informed by the cosmic significance of color. Patterning does not stop with the form and color of each of the images. Each of the shrine figures represents a distinctly different source area. The northern Great Mortuary pair represents the Cahokia area and the Republican River area. The central pair remains without fingerprints, but the white one is plausibly from the Ozark Uplift and the dark mate from another location. The pair high in the main cone are from distinct sources as well; the smoker pipe is arguably from the southern Appalachians. That leaves blanks in our knowledge about the direction of sources, but one can argue that the southern and western blanks are partially covered by the Olivella beads from the Gulf of California, the preserved swatch of southwestern cotton cloth, and even the obsidian scraper from the Pachuco formation in Hidalgo, Mexico. The beads were found on the floor of the Great Mortuary; the obsidian could be from a location a meter or so north of the Big Boy pipe; and the cotton cloth, which was preserved by copper plates, was from a sacred box bundle (A6) from an area outside of the Great Mortuary with strong connections with that sacred space through cross-mends (Brown 1996:72; King and Gardner 1981). The obsidian scraper of relatively uncertain archaeological context possibly belongs to the same timeline, if not to the same stratigraphic unit (Barker et al. 2002:104; Brown 2010; Brown and Rogers 1999). When combined with the references made by the images themselves, the result becomes a powerful statement respecting the sacred collectivity of the main mound cone as a whole. It is surprising that this information can be squeezed from isolated items, since we know so little about the contextual data surrounding most objects in the Great Mortuary. It is safe to suppose that the repeated deposition of paired image pipes employing color, gender, and spiritual personas is only a part of the cosmological design represented in the Great Mortuary tableau. Disposal of the Dead without Personhood Considering all that has been written in support of Spiro stratification, the reader might be surprised to learn that both the human bone and the precious objects were brought to the mound and arranged in a deposit Spiro Reconsidered

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

131

Copyright © 2012. Nebraska Paperback. All rights reserved.

created for a collective purpose, not to honor or memorialize anyone’s social standing (Brown 2010). Good reasons can be marshaled for thinking that cosmological principles guided the placement of the “graves” as well as the principal cult objects and the sinking of upright posts among them. In this respect the component layers of the Great Mortuary can be thought of as constituting a highly structured ritual tableau in which the tall conical mound surmounting the deposit represents the third dimension of this tableau. I have argued elsewhere that disarticulated funerary deposits are the product of collective rites when they are part of a single interment event (Brown 2003a, 2003b, 2006, 2010). Bones and items spread upon a single well-defined floor qualify amply as a single event. Collective rites make it difficult to argue that the objects interred adjacent to an individual interment “belong” in any meaningful way to that individual, assuming that they can be singled out objectively. Moreover, it is not possible to assert that “individuals” retain a social identity, or personhood, in the way that isolated individuals in pits or facilities can have. The individual extended interments with grave goods found at the Sanders site on the Red River, for instance, qualify as social individuals (Jackson et al. 2000). The Great Mortuary interments are doubtful. Although I have argued for personhood over the years, the case in support has eroded considerably (Brown 2003b, 2010; Gillespie 2001; Goldstein 2000). First, the cane basket burials become a status class only through courtesy of human bones being included in some—but not all—baskets. These I have come to recognize as sacred bundles (Brown 2010). The one or two teeth or bones in these bundles are simply one of the artifacts included, preeminently those being copper repoussé plates, copper-headed axes, and shell beads. Bones can be regarded as enhancements of bundle power rather than the marking of personhood. That leaves the burials in biers for consideration. The paltry representation of human remains in the biers has been an argument that high status is indicated thereby (Brown 1971, 1975, 1981). But the personhood standing of these interments is very weak. The density of human remains is arguably little different from the ossuary deposit beneath. At least one (LfB62) that we know rather well is actually composed of two individuals: one a cremated bone fused within a whelk shell fragment, and the other consisting of highly visible skull parts among a pair of associated ear spools. This is not a sterling example upon which to argue 132

Brown

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

for personhood. With these two burial types deprived of their former standing as documents for high status personages, we are left with the excarnated burials. The record for their standing as being accompanied by grave goods likewise is not stellar.

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Devotional Economy The splendor of the artifact accumulation in the Great Mortuary at Spiro testifies to strong outside connections whose presence customarily has been taken to indicate trade. Long distances implied great attractive strength, particularly where so many different source areas are involved. That attraction typically has been modeled on a control over a desirable resource or set of goods. Those who have preferential say in that control exercise stand to attain social deference that can be transferred to status display in death. Long-distance connections are implied by the marine shells, hypertrophic objects in exotic materials (e.g., Kaolin and Duck River cherts), and effigy pipes of exotic materials (copper, marine shell cups, see Banks 1990; Bell 1947a; Brown 1983, 1996; Emerson and Hughes 2000; Emerson et al. 2002, 2003; Phillips and Brown 1978, 1984; Walthall 1981) Many of these can be sourced through raw material or artistic style to distant locations (Bell 1947a; Brown 1983, 1996; Muller 1987). The Great Mortuary has produced exotic and specialized goods at such an outsized scale that it is difficult to distinguish site-wide patterning from the stronger one at the Great Mortuary itself. Furthermore, materials from graves outside of the Great Mortuary frequently do not derive from distant sources (Brown 1983). Hence the Great Mortuary poses some interpretive problems. When the Great Mortuary collection of objects is subtracted from Spiro grave goods all told, grave lot trends can be identified that have correspondence to other ones in the transMississippi South. To account for the exceptional accumulation in the Great Mortuary under the trade and exchange thesis, the volume of exchange would have to have been particularly intense during or immediately preceding the Great Mortuary. If one looks outside of the Great Mortuary, the end of the thirteenth century (Spiro III) would rate as the time when exchange was most intense. It is during that period at Spiro that the most shell beads (of all types) ended up as grave goods. The number of beads per total count of Spiro III burials measured this relationship. Other trade Spiro Reconsidered

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

133

Copyright © 2012. Nebraska Paperback. All rights reserved.

items can be attributed to this period, as much as 100 years before the Great Mortuary (Brown 2010). The peak in exchange intensity took place in Spiro III times, in contradiction to many assumptions about the importance of exchange relations during the Great Mortuary (e.g., Phillips and Brown 1978). The sheer quantity of valuables in the Great Mortuary clearly has posed a challenge for trade as a vehicle for bringing the engraved shell to Spiro (Muller 1987:18–19). A great disparity exists between the large number of shell cups associated with the Great Mortuary and the small numbers of cups in other Spiro graves, made all the more contrastive by the latter cups being exclusively unengraved. Two graves have yielded shell cups, in both cases only unengraved ones. From the Spiro II grave of BrB6 are 25 shell cups from a small litter burial, usefully thought of as a prototype for the much larger ones in the Great Mortuary (Brown 1996). A Spiro IV grave (BrB3/5) has 11 cups placed upon an excarnated extended burial of an adult female (Brown 1996). If for the sake of argument we suppose that the number of cups in the Great Mortuary is simply a multiple of the number found distributed in extra-mortuary graves, it would take 40 of the former small litter burials to achieve the number of 950 to 1,020 unique shell cups and gorgets estimated to have been in the Great Mortuary and 90 in the case of the latter excarnated extended burials. It stretches credulity that so many burials remain undetected at Spiro—or anywhere else for that matter. This calculation leads to the conclusion that the volume of shell cups flowing to the Great Mortuary is a result of processes that have little or nothing to do with the trickle of shell cups that occasionally made their way into extra-mortuary graves. Without going into detail, much the same picture of moderate exchange flow can be supported from the grave finds at the Sanders site, one of the few mound centers with anywhere near the volume of marine shell cups and gorgets reported from Spiro. Schambach (2000a) shows that this site is a much better candidate for social complexity than Spiro. For Spiro an economy of an entirely different kind is thereby indicated. Much the same conclusion can be drawn in the case of copper repoussé. The fact that both marine shell (largely whelks from the west coast of Florida) and copper (unsourced but largely tooled by Cahokian craftspeople) constitute highly visible types of exotic materials make the case for an extraordinary transport of these objects to a site at the western 134

Brown

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

edge of the eastern Woodlands. In certain respects, the record for these two materials falls in with the geographical pattern noted for the large effigy pipes, when these are seen as sacred objects. In the copper and pipe categories, a handful of local products, or ones from within a radius of 200–300 km, join the trove of ones from far outside that sphere. This pattern makes sense as being generated by gifts from afar simultaneously with those near at hand and those made on-site. The last named source has clearly not been considered, but inasmuch as the cloth mantles were thought to have been made for the occasion, why not other items? I am particularly impressed by the appearance of a Craig C head embroidered on an example of twined tapestry (King and Gardner 1981). Human figures with the same style of head figure prominently in the corpus of engraved shell cups. Copper plates do not display this pattern. Instead, their association with bundles suggests that it is the bundles that have moved, not the separate contents. Production for the event has not been clearly examined, whether it was accomplished on-site or at a more remote location. The sacred economy articulated by Renfrew (2001) clearly has a research agenda that remains untapped. It is commonly thought that an accumulation of exotic artifacts is due to trade or exchange. Exchange presupposes a set of objects or commodities that can be used in exchange. Little can be tracked to the Spiro locality (Muller 1987:18). In an effort to create such a commodity base, Schambach (1999, 2000a) has suggested that bison hides and Osage Orange bow wood may have held that role, although tangible traces have eluded detection. In a summary of Southern Plains exchange, Vehik (2002) described different patterns of exchange across this region, some of which were situational and mutualistic. Most took place across the latitude of the Red River where a more substantial density of settled agriculturists had been established since ca. 1000. The Sanders culture of this region has had exchange connections with Spiro, but with little to suggest that it derived much from the Spiro vicinity (Bruseth et al. 1995). There is reason to suppose that trade is not operative in the sense of an exchange of one object or commodity for another. Certainly, it does not apply at the scope and scale required to explain the quantity of exotics in the Great Mortuary and their presence from so many diverse sources. Down the line exchange probably had a place in the Spiro economy, but it is so small in scale as to be relatively insignificant. Schambach (2000a) has proposed that Spiro was an entrepôt funneling Plains area goods Spiro Reconsidered

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

135

Copyright © 2012. Nebraska Paperback. All rights reserved.

(primarily bow wood) into the Ouachitas and the Red River basin portion of the Caddo archaeological area. In his scenario, engraved marine shells flowed west to balance the eastward movement of bois d’arc or Osage Orange. Never mind the near invisibility of this wood species in the archaeological record, a more basic issue has to do with trade as exchange in the first place. The economic activity associated with the Great Mortuary tableau can be more appropriately regarded as a sacred economy borne by gifts toward a common enterprise. Sacred gift exchange is a name for a kind of “devotional” economy propelled by the attraction to a site of sacred goods with highly spiritual powers. Ordinary gift exchange is not a key feature of this economy, because the benefits of the gift are measured by the intangible return of a power that accrues from amassing gifts from many donors (Morehart and Butler 2010; Sanderson 2008). The result is a power that is greater than the sum of its parts. Hence if the value of an item is measured by its spiritual qualities, a merger with complementary qualities yields a power that is vastly more powerful than the components. This kind of exchange has as its objective the collaborative effort to convey sacred goods to a central location with the intention of benefiting from the higher-order power that it can emanate. Key is the inverse relationship of distance and spiritual aura that sets up prior conditions on value. Whereas one could argue for trade as applying to some Spiro contexts, the Great Mortuary stands apart in its accumulation of specific kinds of valuables and ones that plausibly possess high spiritual value. None of the distant source locations, however, have yielded anything during the period under consideration that could be pinpointed to Spiro or even attributed to the Caddo archaeological area. Summary The Great Mortuary at the Spiro site, which forms the entire central cone of the Craig mound, is not a closed grave record in the usual sense. Rather, it is a testimony to the deployment of sacred objects and human remains to create a sacred monument. These remains are divorced of personhood, and the objects that entered into the construction of the monument were chosen to stand for the known universe in its geographical and cosmological dimensions. Certain objects such as the engraved shell cups in the Craig C style may have been produced for the occasion; likewise, the textiles appear to be woven for the event. A quid 136

Brown

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

pro quo exchange does not dominate, nor does it explain the outsized accumulation of marine shell cups and copper repoussé plates. Central leadership is not warranted by the exchange, nor does it have any compelling basis in the special burials. The Great Mortuary burial displays and the outsized accumulation of exotic goods have a different social and cultural basis than heretofore argued. The mortuary tableau created makes use of a multitude of human remains to express collective and perhaps corporate identity, rather than using the medium of individually dedicated interments. The object of allurement is an artifact itself produced by a group intending to influence events and to affect social change. In this regard the productive capacity to fund the assembly of ritual specialists and labor that went into the construction of this monument had to be promoted by one or more powerful families under an energetic leadership. Yet the political security that could be translated into a funerary focus on a leader had not materialized. The Great Mortuary was a monument for alluring population to Spiro, not a statement of political structure that was yet to be.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Acknowledgments The author would like to express his indebtedness to Prof. F. Kent Reilly III, who first proffered the suggestion that the Great Mortuary layout might be productively viewed as a cosmogram. In line with this perspective, workshop panels convened on Mississippian iconography at Texas State University have enriched this perspective on the cosmological message contained in the remarkable Great Mortuary context. Additional ideas likewise have sharpened the perspective offered herein. I would like to thank my fellow workshop participants Alex Barker, Robert Hall, and George Sabo for prodding me toward the present interpretation. I look forward to additional insights that they and others will provide. On more than one occasion, discussions held during Reilly’s annual workshop on Mississippian iconography have helped to sharpen my perception of Spiro and the Great Mortuary. The views expressed herein are ones I take responsibility for. Notes 1. One of the pieces of engraved shell from B 80-5a is in the Craig C style (Brown 1996:figs. 2-138x and 2-148t; Phillips and Brown 1984:xiii [third section]). Note that the elevation of the floor of the pit (B 80) is in error (see comments in B 77 [Brown 1996:704]).

Spiro Reconsidered

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

137

Copyright © 2012. Nebraska Paperback. All rights reserved.

2. Phillips and Brown (1984:xi [third section]) have counted in different ways. “Of the resulting larger total, 1,445 artifacts and fragments” (“whole artifacts reasonably intact, matching fragments counted separately”) the University [collection] accounts to 40 percent. The contribution of the “hollow chamber” of the Great Mortuary is indicated by another breakdown: “the University [of Oklahoma] collections include 3 cups and 1 gorget compared with 95 cups and 10 gorgets attributable to the commercial diggers.” All except one (A11-7); but it is from a grave lot north of the Great Mortuary floor that seems to be involved in its setup (Brown 1996:674 and fig. 2-50b).

138

Brown

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

6. Viewshed Characteristics of Caddo Mounds in the Arkansas Basin gregory vogel

Copyright © 2012. Nebraska Paperback. All rights reserved.

Introduction Between 900 and 1400, Caddo populations in eastern Oklahoma, northwest Arkansas, and southwest Missouri constructed large earthworks that required the sustained and organized efforts of large groups of people to complete. Some of the earthworks served as burial mounds, some covered the remains of what appear to be structures for processing the bodies of the deceased, and others served as platforms for buildings in which local leaders resided. These mounds clearly held great significance to the people who built them as locations where important social, political, economic, and ritual gatherings took place. This chapter is a study in the visibility of these monuments: What could be seen from the mounds? From where on the landscape could the mounds be seen? What are the overall dynamics of how the mounds visually interacted with the landscape? The analysis begins with quantitative evaluations of the size of mound viewsheds, taking into account mound and site types and the differences between views to and from mound locations both before and after mound construction. The quantitative analysis continues with a Monte Carlo sampling approach to derive a statistical background for regional viewsheds: Are mound viewsheds larger or smaller than would be expected by chance alone? Qualitative aspects of the viewsheds are also evaluated for potential intervisibility between sites and other mound-to-landscape visual relationships. Primarily this study is an investigation into the visual characteristics of the mounds in relation to the overall landscape and a search for patterns within those characteristics. Rather than a “stand-alone” study with independent conclusions, I offer this analysis as novel lines of inference to be integrated into further site- and landscape-scale questions. 139

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

For example, elsewhere in this present volume, Brooks and Lockhart explore mound site relationships with environmental variables such as watersheds, topography, and soils. Walker and McKinnon focus on community structure within individual sites. While the geographic focus of each of these studies differs, they all share a common theme of spatial analysis and a basic premise that the constructed landscape, in the form of mounds and habitation features, is intimately linked with numerous aspects of the surrounding environment. I add viewsheds to this mix as a further environmental variable for consideration. And just as water courses, soils, vegetation, or any other environmental variable may simultaneously influence and be altered by human populations, viewsheds appear to have had at least some influence on mound locations, and the constructions of the mounds themselves appear to have altered mound site viewsheds in significant ways. Mound Sites in the Arkansas Basin

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Arkansas basin is topographically and environmentally diverse, including portions of the Ozark Plateaus, the Arkansas and Neosho Rivers and their tributaries, and the northern reaches of the Ouachita Mountains (fig. 6-1). Most mounds in this area can be classified into three types, after Brown et al. (1978): structural, accretional burial, and platform. Structural mounds are low, conical features erected over a previous structure. The structures buried at the base of the mounds appear to have been used for specialized ceremonial purposes. Many are hypothesized to have served as charnel houses or temporary burial locations. These mounds commonly contain few or no burials or artifacts except for remnants of the structures buried beneath them. Accretional burial mounds are elongated, multilobed, or round features that are the result of multiple episodes of mound building. Each building episode generally contains multiple burials, many of them secondary burials that may have been processed within the structures underlying nearby structural mounds. Grave goods and elaborate burial features are common in this type of mound. Accretional burial mounds and structural mounds may represent paired earthworks: structural mounds built over relatively short-term processing structures, and accretional burial mounds built as permanent locations for the burial remains and grave goods. Platform mounds are generally large rectangular and flat-topped 140

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

6-1. Relief map showing locations of sites employed in this study.

structures, occasionally with a capping layer that gives the final form of the mound a conical shape. The fill is commonly found to be devoid of artifacts. These generally contain no burials except for those of protohistoric or historic age, which are common. The Norman mound I-1 is typical of area platform mounds, and excavations have revealed a general building sequence typical of that reconstructed from other platform mounds (Vogel et al. 2005). First, a mound stage consisting of a few decimeters of soil with a flat surface was constructed, then numerous large pits were excavated into this stage, exposed for an unknown period of time, and refilled. The initial soils used for each mound stage and the soils used to refill pits appear to have come from many sources, some of them local and others likely from a significant distance away. After one or more mound stages were emplaced, the surface was capped by a compacted and burnt layer of clay, and another cycle of construction commenced. Subsequent pits in higher mound surfaces never penetrate previous capping layers. Evidence for mounded berms, large posts, and Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

141

Copyright © 2012. Nebraska Paperback. All rights reserved.

other structures exist within many of these platform mounds as well, but the basic unit of construction seems to be a cyclical sequence of filling, pitting, and capping. These three mound types are found either singly on the landscape or aggregated to form mound centers that can be classified into three distinct hierarchical echelons. First echelon centers are the simplest or least elaborate type, containing at least one burial mound (or a non-mounded burial area), and an associated mortuary structure, which may or may not be mounded. Second echelon centers contain an additional platform mound. Third echelon centers are essentially large second echelon centers with the addition of a fourth mound type or other large architectural element. The additional mound type or architectural element is generally unique to the site, suggesting an “organizational discontinuity with the lower-order centers” (Brown et al. 1978:189). The unique structures commonly contain burials with large associated caches of finely made artifacts, and may have served as burial places for the regional elite. The Spiro, Norman, and Harlan sites are the only third echelon centers recognized in the Arkansas Basin study area. This study employs a total of 24 sites; of the 72 mounds recorded in the Arkansas Basin study area (table 6-1), 41 were suitable for inclusion in this study. The remaining 31 mounds were not included either because their locations could not be ascertained with sufficient precision or because their inclusion would have produced information redundant with adjacent mounds of a similar type. Three sites (Goshen, Unauthorized, and Berry) are poorly understood and are included here only provisionally. Mounds and Visibility Studies in mound visibility are warranted by numerous lines of evidence that point to the importance of their visual characteristics. Concerning prehistoric monumental architecture in general, Bradley (2000:104) asks, “What do monuments do to the places they are built?” He explains that (1) they transform the entire way the locations are experienced; (2) they make particular places more visible and make certain spots more visible to one another; and (3) they can invest already significant natural places with additional layers of symbolism. Overall Bradley (2000:158) concludes that though monumental architecture may be reworked through time and adapted to new purposes, all monuments “were origi142

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

nally conceived with the aim of conveying ideas to a large audience.” A prime criterion for conveying ideas to large audiences is simple visibility. Arguing that platform mounds are highly communal in nature, Knight (1986:678) writes,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Despite their size and communal derivation, Mississippian platform mounds are no less sacra, that is, objects of sacred display, than are embossed copper plates, engraved shell cups, or monolithic axes. Periodic rebuilding of the mound surfaces by the addition of a new blanket of earth, the special characteristic of these mounds, demands to be seen as a purely expressive act. Pursell (2004) calls attention to the brightly colored soils that periodically covered many platform mounds. Instead of the neutral tan or grass-covered features commonly portrayed in mound site reconstruction drawings, they may more properly be depicted as covered in bright red, yellow, white, or gray soils, contrasting highly with the natural background. Visual elements of mounds in the Arkansas Basin have been noted before. For example, Morrow (2004) has speculated that the flat-topped and cone-shaped mounds at the Spiro site mimic prominent natural features, including both flat-topped and cone-shaped mountains that dominate the southern horizon from Spiro. In Morrow’s view, natural features may have helped shape the worldview of the mound builders, serving as both inspirations for cultural constructions and powerful places in their own right. Concerning views from the Collins, Huntsville, and Goforth-Saindon mound centers in the Ozark Plateaus, Kay et al. (1989:137) note that “the visual effect is that of being on stage of a large amphitheater, as one has a commanding view of the valley and flanking uplands from any of the mounds.” The “amphitheater effect” from mounds in the Ozark Plateaus is telling and is discussed in more detail below. Viewsheds This study employs Geographic Information System–based viewshed analysis to quantify viewshed sizes and to gain some insight into the qualitative aspects of views to and from mounds. A viewshed is the visual equivalent of a watershed: everything that can be seen from a particular spot on a landscape is within that location’s viewshed. Viewsheds are created in gis using a topographic model and the location and height of Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

143

Table 6.1. Recorded mounds in the Arkansas Basin, ordered by state site number

Copyright © 2012. Nebraska Paperback. All rights reserved.

Site

Site echelon

3BE245 Goforth-Saindon

2

3CN4 Point Remove

2

3FR1 Spinach Patch

1

3LO15 Page 3LO208 Logan Eddy

2 2

3MA22 Huntsville

2

3PP23 Scotia

1

3PP49 Cove Creek 3SB1 Guy Brittain 3SB3 Cavanaugh 3SC1835 Unauthorized 3WA1 Elkins

1 1 2 1 2

3WA1241 Goshen 3YE15 Bluffton

1 2

3YE21 Berry 3YE352 Carden Bottom Cem. 23MD46 Pineville 23SN42 Loftin 34AD11 Ewing Chapel Cem. 34CK6 Harlan

1 2 2 2 2 3

Mound name

Mound type

Mound 1 Mound 2 Mound 3 Mound 4

Platform Unknown Unknown Unknown Platform Unknown Unknown Burial Burial Burial Unknown Unknown Unknown Platform Unknown Unknown Unknown Unknown Unknown Burial Burial Platform Unknown Platform Unknown Unknown Unknown Unknown Unknown Platform Unknown Unknown Platform? Platform Platform Platform? Burial Unknown Platform Structure Platform

(several smaller mounds) West Mound East Mound Mound A Mound B Mound C Mound A Mound B Mound C Mound D Feature 1 Feature 2

Mound A Mound B Mound C Mound D Mound E South Mound North Mound

Unit 1 Unit 3 Unit 4 Unit 6 Unit 7

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Notes

Major references

Partially excavated

Kay 1986; Kay et al. 1989 Kay 1986; Kay et al. 1989 Kay 1986; Kay et al. 1989 Kay 1986; Kay et al. 1989 Davis 1967; Vogel 2005b:Appendix A Vogel 2005b:Appendix A Vogel 2005b:Appendix A Bond 1970 Bond 1970 AAS site files; Vogel 2005b:Appendix A AAS site files AAS site files AAS site files Kay et al. 1989; Sabo 1982, 1986 Kay et al. 1989; Sabo 1982, 1986 Kay et al. 1989; Sabo 1982, 1986 Kay et al. 1989; Sabo 1982, 1986 Caldwell 1958 Caldwell 1958 AAS site files AAS site files; Vogel 2005b:appendix A Vogel 2005a; Vogel 2005b:appendix A AAS site files Fritz 1979, 1986; Kay et al. 1989 Fritz 1979, 1986; Kay et al. 1989 Fritz 1979, 1986; Kay et al. 1989 Fritz 1979, 1986; Kay et al. 1989 Fritz 1979, 1986; Kay et al. 1989 AAS site files AAS site files; Vogel 2005b:appendix A AAS site files; Vogel 2005b:appendix A AAS site files AAS site files; Vogel 2005b:appendix A Conner 1998; Ray 1996 Bray 1983 OAS site files; Wyckoff 1980 Bell 1972; Cranford 2007 Bell 1972; Cranford 2007 Bell 1972; Cranford 2007 Bell 1972; Cranford 2007 Bell 1972; Cranford 2007

1967 AAS dig, Partially excavated Removed before 1967 Excavated Excavated Destroyed by looting

Partially excavated

Copyright © 2012. Nebraska Paperback. All rights reserved.

Partially excavated Destroyed by looting Exposed profile Considered provisionally Destroyed by looting

Destroyed by looting Considered provisionally Partially excavated Partially excavated Considered provisionally Partially excavated Excavated Triple mound, excavated Excavated Excavated Excavated Excavated

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

(continued)

Table 6.1. Recorded mounds in the Arkansas Basin, ordered by state site number (continued)

Site

Site echelon

34CK43 Brackett 34DL1 Reed 34DL4 34DL41 Lillie Creek 34LF6 Victor Area 34LF9 Holson Creek

2

34LF16 Sol Thompson 34LF37 Spiro

1 3

34LF70 Skidgel #2 34LF117 Borrow Pit

3 2

34MI45 Eufala 34MS4 Hughes #1

1 2

Mound name

Reed #1 Reed #4 2 1 1

Mound 1 Mound 2

Craig Brown Copple House Mounds 1 to 6 Ward Mounds 1 and 2 Mound 1 Mound 2

34MS175 Ft. Davis

Copyright © 2012. Nebraska Paperback. All rights reserved.

34MS52 Cat Smith 34SQ12 Parris 34WG2 Norman

Mound type Platform Platform Burial Platform Unknown Unknown Unknown Unknown Burial Platform Platform Structure Structure Platform Unknown Unknown Burial Platform

Unknown 1 2 3

Mound I Mound II Mound III Mound V Mound VII Mound VIII

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Burial Platform Platform Burial Structure Unknown Unknown Unknown

Copyright © 2012. Nebraska Paperback. All rights reserved.

Notes

Major references

Excavated Excavated Excavated Excavated Mostly destroyed by looting Partially excavated Partially excavated, may have one more mound Partially excavated, four mounds Looted and excavated, triple mound Looted and excavated Excavated Mostly excavated Mostly excavated Excavated, associated with Spiro Partially excavated, double mound Partially excavated Excavated Excavated. 34MS4 and 34MS175 designated as two sites; considered one site here

Bareis 1955; Howard 1940 Purrington 1971 Purrington 1971; Thoburn 1931 Purrington 1971; Thoburn 1931 Watson 1947; Wyckoff 1980 Bell 1947

Several smaller mounds possible. Destroyed by looting Partially excavated Excavated, double mound Excavated, double mound Excavated Excavated Excavated Excavated

Bell 1947 Brown 1996 Brown 1996 Brown 1996 Brown 1996 Brown 1996 Brown 1996 Brown 1996 OAS site files OAS site files Orr 1941, 1942

Brown 1996 OAS site files OAS site files; Vogel 2005b:appendix A OAS site files; Wyckoff and Barr 1967 Muto 1978; Muto et al. 1980 Albert 2000; Vogel et al. 2005; Cranford 2007 Albert 2000; Finklestein 1940; Cranford 2007 Albert 2000; Cranford 2007 Albert 2000; Cranford 2007 Albert 2000; Cranford 2007 Albert 2000; Cranford 2007

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

one or more viewpoints. For this study the topographic model is based on existing and modified U.S. Geological Survey digital elevation data, and the viewpoints are the locations of the mounds themselves: both the ground-level elevation, representing pre-mound views, and mound summit elevations, representing views from the constructed mounds. Viewsheds are generally reciprocal, meaning that they show not only what locations on a landscape are visible from a particular point but also the locations on the landscape from which that point itself is visible. This assumption of reciprocity is not always true (Wheatley and Gillings 2000:7–8), but given a relatively low viewer elevation compared to the overall viewing distance (as is the case with this study), the differences between views to a point and views from a point are minimal. The viewsheds generated here thus show locations that could be seen from the mounds and locations from which the mounds themselves could be seen. There are exceptions, however, with quite interesting implications, discussed below. gis viewshed studies cannot be taken literally as applied to past landscapes. Wheatley and Gillings (2000) elucidate numerous problems with accepting computer viewshed output uncritically. It would be impossible, for example, to model precise locations of individual tree branches that might have obstructed a particular view. But as Wheatley and Gillings (2000) and Lock (2003) demonstrate, even if such studies cannot exactly replicate views of the past, they can be usefully applied to answer specific questions. Thus the viewsheds represented in this study do not demonstrate exactly what could or could not have been seen, but represent maximum potential views to and from the mounds in order to test specific hypotheses, including: (1) Does viewshed size correspond to mound type or site echelon? (2) Are viewsheds from the mounds larger than would be expected by chance alone? (Specifically, were large viewsheds a consideration in mound site location?) (3) What potential intervisibility, if any, exists between mound centers? In addition, the morphology of viewsheds generated here—that is, not just how much could be seen from each mound but what could be seen—may offer interesting if cautious insights. 148

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

GIS Procedures

Idrisi Kilimanjaro version 14.02 was used for all gis analyses in this study. The base topographic model is drawn from U.S. Geological Survey digital elevation data with a resolution of 1-arc second of latitude and longitude, resulting in a raster-based elevation model with pixels approximately 30 m per side. Although higher resolution elevation models are available, the overall size of the study area (nearly 11 million hectares, or ha), and the necessity for very large elevation data sets in order to eliminate edge effects in the analyses, renders the use of such models unfeasible (given current data processing speeds) due to file size. Because the available elevation models are based on recent topographic information, numerous constructed reservoirs (primarily impoundments of the Arkansas River and its tributaries) are expressed as flat surfaces corresponding to average lake levels. This is problematic for viewshed generation where mound sites are near or within such reservoirs. The existing elevation data set was therefore modified by digitizing and incorporating topographic maps drawn prior to reservoir construction as “bridging” layers to model pre-reservoir topography. Several other clearly modern landscape modifications that interfered with viewshed outcomes (such as tall railway and road embankments) were similarly corrected (see Vogel 2005b:46–53 for details on these procedures). Approximately 80,000 ha of the elevation model were corrected in this way, primarily within four reservoirs in Oklahoma: Grand Lake o’ the Cherokees, Fort Gibson Lake, Robert S. Kerr Lake, and Eufaula Lake. For this study, simple Boolean viewsheds were generated, defining the location of everything in view and not in view from the mounds. Maximum viewing distance was set at 25 km. At this distance, not only the locations of the sites but the mounds themselves would have been visible to observers, as long as they contrasted sufficiently with their backgrounds. Considering the potential for brightly colored soils covering many of the mounds and evidence for large structures and even fires on the summits of platform mounds, it is likely that the mounds contrasted highly with their backgrounds, at least periodically if not regularly. Most of the mounds no longer exist due to agriculture, Works Progress Administration excavations, the construction of reservoirs, and intentional removal. Of the 72 mounds formally documented in the Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

149

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

1 5 6 8 9 12 16 18 19 20 21 22 27 28 30 31 32 33 34 35

Figure key

3BE245 3CN3 3FR1 3LO15 3LO208 3MA22 3PP23 3PP49 3SB11 3SB3 3SC1835 3WA1 3WA1241 3YE15 3YE211 3YE352 23MD46 23SN422 34AD11 34CK6

Site

Goforth-Saindon Point Remove Spinach Patch Page Logan Eddy Huntsville Scotia Cove Creek Guy Brittain Cavanaugh Unauthorized Elkins Goshen Bluffton Berry Carden Bottom Cem. Pineville Loftin Ewing Chapel Cem. Harlan

Table 6.2. All mound viewsheds

Copyright © 2012. Nebraska Paperback. All rights reserved.

Mound 1 Main Mound West Mound Feature 1 Mound A Mound A Feature 1 Mound Mound Mound Mound Mound A Mound South Mound Mound Mound Mound Mound Mound Unit 1

Mound 2 2 1 2 2 2 1 1 1 2 1 2 2 2 1 2 2 2 2 3

Site echelon Platform Platform Burial Unknown Unknown Platform Unknown Burial Burial Platform Unknown Unknown Unknown Platform Unknown Platform Platform Platform Platform Burial

Type of mound 3.5 9 0.5 2.4 1.6 3.1 4.5 1 — 9 1 3 1 5.5 — 1 1.2 1 1.5 2

Mound height (m) 699 10109 8729 4546 2678 749 4713 3597 15580 7819 189 2960 2824 7779 5572 5632 411 1013 3352 2432

From base

1020 19009 8989 4927 2846 860 5268 3881 — 15967 196 3384 2992 11604 — 7429 480 — 3709 2783

From summit

321 8900 260 381 168 111 555 284 — 8148 7 424 168 3825 — 1797 69 — 357 351

Base to summit difference

Viewshed size in hectares:

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

34CK6 34CK6 34CK43 34DL1 34DL4 34DL411,3 34LF61 34LF91 34LF161 34LF37 34LF37 34LF37 34LF37 34LF70 34LF1171 34MI451 34MS4 34MS175 34MS522 34SQ12 34WG2 34WG2

Harlan Harlan Brackett Reed Reed Lillie Creek Victor Area Holson Creek Sol Thompson Spiro Spiro Spiro Spiro area Skidgel Borrow Pit Eufala Hughes #1 Fort Davis Cat Smith Parris Norman Norman

Unit 4 Unit 7 Mound Mound 1 Mound 4 Mound Mound Mound Mound Craig Brown Copple (Round Mtn.) Mound Mound Mound Mound Mound Mound Mound Mound I-1 Mound II-1 3 3 2 2 2 2 3 3 3 3 3 3 — 2 2 1 2 2 3 2 3 3

1. Mean of samples from a systematic grid of “trial viewsheds” across the site. 2. 1.5 m above uncorrected reservoir surface. 3. Thoburn’s 1931 height of 14.5 m.

37 39 40 41 42 43 44 45 47 48 49 50 — 59 60 62 63 64 65 66 67 68

Copyright © 2012. Nebraska Paperback. All rights reserved.

Platform Platform Platform Platform Burial Platform Unknown Unknown Unknown Burial Platform Platform — Platform Unknown Burial Platform Unknown Burial Platform Platform Burial 1 4 2 2 1 14.5 2 — — 12 6 2.5 — 5.5 — 1.5 6 3 — 4 8 3

2583 2695 1474 774 755 618 6858 4410 3357 4600 9794 11329 38030 22082 2113 2727 7239 14157 2845 1831 1436 1738

2718 3183 1726 920 842 1299 8091 — — 14494 14753 13115 — 25446 — 3340 11056 16288 — 2127 2441 2075

135 485 252 146 87 681 1233 — — 9894 4959 1786 — 3364 — 613 3817 2131 — 296 1005 337

Copyright © 2012. Nebraska Paperback. All rights reserved.

Arkansas Basin, portions of only 23 are preserved on the landscape today. Most of the destroyed mounds, however, are visible in aerial photographs taken by various divisions of the U.S. Department of Agriculture beginning in the 1930s. These early aerial photographs were scanned and georeferenced into gis, and independent information concerning mound locations (early sketch maps, etc.) were used to aid in interpreting and digitizing mound locations from these photographs. In most cases, mound locations could be confidently determined to within a few decameters, well within the maximum precision of the topographic data employed. Viewsheds were generated from the base of each mound (current topographic elevation in most cases, where the mounds have been removed) to determine pre-mound views, and the summit of each mound (estimated from various sources, see Vogel 2005b for details) to determine viewsheds after the mounds had been constructed. Iterations using various viewer heights, between 1–2 m, demonstrated that a difference of a few decimeters of viewing height made little if any difference in the size or morphology of overall viewsheds. A viewer height of 1.5 m is used throughout the study. For mounds whose precise location could not be determined with confidence, a systematic grid of viewsheds was generated across the site area and averaged for an estimate (Vogel 2005b:65–69). Two mounds (at the Cavanaugh and Point Remove sites) are prominent enough to be expressed on the modern elevation model as topographic highs. In order to generate pre-mound viewsheds, these were digitally removed by averaging ground elevations surrounding the prominences. As most of the mounds were constructed in stages, viewsheds from early stages, although not estimated here, would be smaller than those shown. Also, many of the platform mounds served as the base of substantial structures. Accordingly these would have increased viewsheds to the mounds, while not affecting viewsheds from the mounds, unless an observer were standing on top of the structure. Table 6-2 lists the individual mounds used in this analysis and the absolute sizes of their generated viewsheds. Quantitative Aspects of Viewsheds I: Mound Comparisons An initial question in mound/viewshed relationships is the similarity or difference of viewshed size between different types of mounds. Figure 6-2 plots the size of the ground-based viewsheds for all mounds (from 152

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

6-2. Viewshed size from the ground surface (+1.5 m viewer height), plotted against mound heights. Vertical bars represent potential viewshed sizes where relevant variables were not known with confidence. Mounds of unknown height are plotted with a height of 0. Key to mounds is given in table 6-2.

the base of each mound with a 1.5 m viewing height), plotted against mound heights. The viewshed from the ground surface at the top of Round Mountain near the Spiro site—the most prominent peak of the surrounding landscape—is included for comparison. Burial mounds and mounds of unknown function cluster tightly in the lower left corner of the plot, as these are almost exclusively under 3 m in height and with viewsheds under 7,000 ha in size. The Guy Brittain (#19 on the plot) mound may have had a very large viewshed, but as the mound’s exact location on the landscape is unknown, the actual viewshed of the mound may have been anywhere within the plotted bar. Platform mounds clearly dominate both the taller mounds (over 3 m) and those with the largest viewsheds (over 7,000 ha). Of the mounds with a viewshed over 7,000 ha from the ground surface, seven out of a possible 10 are known platform mounds. Craig mound at the Spiro site (#48) is a clear exception, as is Fort Davis (#64). The Fort Davis mound has never been investigated in detail, however, and it may be a platform Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

153

6-3. Viewshed size from mound summits (+1.5 m viewer height), plotted against mound heights. Vertical bars represent potential viewshed sizes where relevant variables were not known with confidence. Key to mounds is given in table 6-2.

Copyright © 2012. Nebraska Paperback. All rights reserved.

mound. Brown (1996:75–83) has noted that the Craig mound holds evidence of construction and use as a platform mound at least twice in its history. Only one of these (Spinach Patch, #6) mounds with a larger confirmed viewshed (over 7,000 ha) is a known burial mound. A large number of platform mounds have relatively small viewsheds as well (10 of 18 have viewsheds of less than 7,000 ha), and statistically the viewsheds from the base of the platform mounds are undifferentiated from those of other mounds (two-sample t-test at 95 percent Confidence Interval [CI], T = 0.22, P = 0.82, df = 34; Mann-Whitney test at 95.1 percent CI, W = 395.0, P = 0.5228). The locations of the platform mounds, as a group, do not appear to have been chosen for their higher visibility any more than the burial mounds or mounds of unknown function. One possible exception is at the Spiro site, where the three platform mounds (Brown [#49]; Copple [#50], and Skidgel [#59]) comprise three of the four largest platform mound viewsheds and three of the six largest viewsheds. Figure 6-3 plots the size of the viewsheds from the summits of those 154

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

6-4. Gain in mound viewsheds (summit viewshed–ground surface viewshed), plotted against mound heights. Vertical bars represent potential viewshed sizes where relevant variables were not known with confidence. Key to mounds is given in table 6-2.

mounds for which there are reliable heights. Again platform mounds dominate the larger viewsheds, with the notable exceptions of Fort Davis (#64), which may be a platform mound, and the Craig mound at Spiro (#48). Platform mounds tend to be taller than other types of mounds, and because of this the viewsheds from the tops of these mounds are correspondingly larger. Statistically, however, viewsheds from the tops of platform mounds are still not significantly different from those of other mounds (two-sample t-test at 95 percent CI, T = 0.93, P = 0.36, df = 30; Mann-Whitney test at 95.2 percent CI, W = 332.0, P = 1.00). Figure 6-4 depicts the gain in viewshed size from the base to the top of each mound. Most are still clustered in the lower left-hand corner, indicating that these mounds are relatively low in height and that there is little difference between the base and summit viewsheds. The Victor Area mound (#44) may have gained significantly in viewshed with only 3 m of height, but this is equivocal because its exact position on the landscape is unknown. While the platform mounds in general are further differentiated by Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

155

Copyright © 2012. Nebraska Paperback. All rights reserved.

the gain in viewshed from their base to their top, the difference between them and non-platform mounds is still not statistically significant (twosample t-test at 95 percent CI, T = 1.12, P = 0.27, df = 31; Mann-Whitney test at 95.2 percent CI, W = 365.0, P = 0.267). Of the mounds not clustered in the lower left-hand corner, 9 of 12 form a clear pattern in the plot, closely following a trend from the Copple mound at Spiro (#50) to the Craig mound at Spiro (#48). These mounds gain about 800 ha of viewshed for each additional meter of height. The exceptions to this pattern are the Carden Bottom Cemetery (#31), Lillie Creek (#43), and Norman mound II-1 (#67). The Carden Bottom Cemetery mound is situated in a broad, flat bottomland area of the Arkansas River, with only slight variations in topography for several km on all sides of the mound. It would be expected that even a small rise in elevation in this area could bring much more of the bottomland into view. An uncorrected flaw in the elevation model may also account for this, however. About 200 m west of the site is a very sharp 2 m rise in elevation that is not actually present on the ground (based on personal observation). This raised area corresponds to the edge of a 7.5-minute U.S. Geological Survey map sheet, from which the digital elevations were initially derived. It appears that elevations from this map sheet were digitized incorrectly, demonstrating a limitation of the existing data. From 1.5 m above the ground surface at the location of the Carden Bottom Cemetery mound, this false rise blocks the view, but with the additional meter of mound height, the viewshed increases dramatically in this direction. Lillie Creek (#43) has a relatively small viewshed to begin with (from the ground surface), and even with the additional elevation of the mound, the viewshed is not increased a great deal. This is somewhat surprising because Lillie Creek may be the tallest mound (from Thoburn’s 1931 description, which is almost certainly unrealistically tall) in the study sample. Both the Norman and Lillie Creek sites are situated in relatively narrow stretches of bottomland along the main channel of the Neosho River, and even relatively large increases in elevation gain them little in the way of additional viewshed. Figure 6-5 depicts mound viewshed sizes arranged not by mound type but by site echelon. Viewsheds from mound locations (ground surface) are not significantly different between the three echelons. Viewsheds from mound summits are significantly different only between the third and first echelon sites (two-sample t-test at 95 percent CI, T = 2.61, P = .024, df = 11; Mann-Whitney test at 95.5 percent CI, W = 90.0, P = .0317). Con156

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

6-5. Mound viewsheds plotted by echelon. Boxes represent quartiles, vertical lines extend the full range of the data (asterisks indicate outliers), and horizontal lines within boxes are median values.

sidering that the presence of a platform or other large mound is necessary for a site to be considered a second or third echelon mound site, this is hardly surprising.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Quantitative Aspects of Viewsheds II: The Statistical Background Here mound viewsheds are tested against a statistical background: namely, are they larger or smaller than we should expect by chance alone? This question is addressed through the gis Monte Carlo randomization method, as described in Kvamme (1996). The Monte Carlo method is particularly useful where the parameters of a variable are difficult to discern or to quantify, which is certainly the case with viewsheds. In this process, the variable in question (viewshed size) is tested against a set of randomly generated samples (within realistic parameters drawn from observation) for statistical significance. If 99 random samples are generated, the significance of a mound’s viewshed is computed simply by ranking it within the generated sample set. If the mound viewshed is larger than 95 of the generated samples, or smaller than 5 of them, its significance is demonstrated at the .05 level. Holding viewer height and maximum search distance constant, Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

157

Table 6.3. Parameters used to generate the predictive model of mound locations within the Arkansas and Neosho River regions Elevation above bottomland

Region

Sites used to define model parameters

Arkansas River Region 1

Cavanaugh, Spiro, Skidgel

Arkansas River Region 2

Point Remove, Carden Bottom Cemetery

Neosho River Region 1

Harlan, Norman, Reed, Lillie Creek

Neosho River Region 2

Hughes, Fort Davis

3-25

Meters from bottomland

Maximum slope (degrees)

0-2600

4.5

0

(within bottomland)

4.5

0

(within bottomland)

4.5

0

(within bottomland)

4.5

Copyright © 2012. Nebraska Paperback. All rights reserved.

viewshed size is a function of viewer location and the surrounding topography. A region of broad, level plains, for example, would produce a very different set of randomly generated viewsheds than would a region with high topographic variability. Six regions of distinct topography were therefore defined for independent viewshed analysis, consisting of two along the Neosho River (Neosho River Region I along the northern reaches, centering around the Harlan, Norman, Reed, and Lillie Creek sites; Neosho River Region II farther south that is centered around the Hughes and Fort Davis sites), two along the Arkansas River (Arkansas River Region I centering around the Spiro site, including Skidgel and Cavanaugh; Arkansas River Region II, farther south, centering around Point Remove and Carden Bottom Cemetery), and one each for the Ozark Plateaus and Ouachita Mountains. Each region still contains a great deal of topographic variability, however, and the mounds are clearly not located randomly within the regions with respect to streams and topography. Caddo mound center relationships with water resources and other environmental variables are explored in detail in Brooks and Lockhart’s chapters. Simple predictive models of mound locations were therefore generated within each region to define more realistic locational parameters, and 99 locations were randomly generated from within these modeled areas for Monte Carlo sampling. For the Arkansas and Neosho River regions, elevation above, and distance from, local bottomlands as well as surrounding slopes served to generate simple but realistic predictive models (table 6-3). 158

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 6.4. Parameters used to generate the predictive model of mound locations within the Ozark Plateau and Ouachita Mountain regions

Region Ozark

Copyright © 2012. Nebraska Paperback. All rights reserved.

Ouachita

Sites used to define model parameters Goforth-Saindon, Huntsville, Elkins, Pineville, Loftin, Ewing Chapel Cemetery Page, Logan Eddy, Bluffton, Borrow Pit

Maximum Cumulative Elevation in Maximum distance from runoff in de-trended slope major stream adjacent DEM (degrees) (meters) stream

-187 to 17

4.5

800

1,000+

—

7.5

400

2,000+

Greater variability in elevation within the Ozark Plateaus and Ouachita Mountain Regions required a slightly different approach. Within these regions, mound sites are located near the larger streams, definable in this case through a minimum runoff value (derived from a gis model of stream runoff ), and in the case of the Ozark Plateaus Region, elevation parameters from an elevation model de-trended for stream gradients (Vogel 2005b:76–88) (table 6-4). Note that first echelon and provisional mound sites were not used to generate these models. Table 6-5 and figure 6-6 present the results of these randomization tests with the viewshed size of each mound ranked by its order within the 99 generated samples from its region. Viewshed ranks in parentheses indicate mounds whose locations were not used to generate the modeled areas and are included for comparison only; the randomization method may not represent an appropriate statistical background for these samples. Because the random samples were generated with a viewer height of 1.5 m, they are more appropriately used to rank the viewsheds from the base of each mound (generated from the same height above the ground surface); the ranking of views from mound summits is included here for comparison and as further quantification of the gain in viewshed by mound height. Within Arkansas River Region I, the Brown and Copple mounds at Spiro and Skidgel mound have significantly large base viewsheds, with Skidgel’s being larger than any generated sample (22,082 ha vs. the largest generated sample of 16,102 ha). Mapping these viewsheds onto the elevation model (see figs. 6-7 to 6-9) shows that much of the Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

159

Table 6.5. Results of Monte Carlo analysis: All mound viewsheds ranked against 99 samples generated for their region Site

Viewshed rank Viewshed rank from base from summit

Mound

Mound type

Mound Brown Craig Copple Mound West Mound Mound (Round Mt.)

Platform Platform Burial Platform Platform Burial Burial —

81 92 56 94 100 (84) (99) (100)

99 99 99 99 100 (88) (—) (—)

Main Mound Platform Mound Platform Feature 1 Unknown Mound Burial

81 17 (12) (4)

100 52 (15) (8)

Arkansas River Region 1: 3SB3 34LF37 34LF37 34LF37 34LF70 3FR1 3SB11 34LF37

Cavanaugh Spiro Spiro Spiro Skidgel Spinach Patch Guy Brittain Spiro

Arkansas River Region II: 3CN4 3YE352 3PP23 3PP49

Point Remove Car. Btm. Cem. Scotia Cove Creek

Neosho River Region I: 34CK6 34CK6 34CK6 34WG2 34WG2 34MS4 34MS175 34MS522

Harlan Harlan Harlan Norman Norman Hughes #1 Fort Davis Cat Smith

Unit 1 Unit 3 Unit 7 Mound I-1 Mound II-1 Mound Mound Mound

Burial Platform Platform Platform Burial Platform Unknown Burial

96 98 98 65 83 (100) (100) (99)

98 98 100 97 93 (100) (100) (—)

Mound 1 Mound 4 Mound

Platform Burial Platform

52 51 35

60 56 81

Mound 1 Mound A Mound A Mound Mound Mound Mound Mound Mound

Platform Platform Unknown Platform Platform Platform Unknown Platform Platform

6 7 49 4 57 (11) (47) (21) (28)

11 8 57 4 62 (—) (50) (27) (35)

Copyright © 2012. Nebraska Paperback. All rights reserved.

Neosho River Region II: 34DL1 Reed 34DL4 Reed 34DL411,3 Lillie Creek Ozark Plateaus Region: 3BE245 3MA22 3WA1 23MD46 34AD11 23SN422 3WA1241 34CK43 34SQ12

Goforth-Saindon Huntsville Elkins Pineville Ewing Chap. Cem. Loftin Goshen Brackett Parris

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Site

Mound

Mound type

Feature 1 Mound A S. Mound Mound Mound Mound Mound Mound Mound

Unknown Unknown Platform Unknown Unknown Unknown Unknown Unknown Unknown

Viewshed rank Viewshed rank from base from summit

Ouachita Mountain Region: 3LO15 3LO208 3YE15 34LF1171 3SC1835 3YE211 34LF61 34LF91 34LF161

Page Logan Eddy Bluffton Borrow Pit Unauthorized Berry Victor Area Holson Creek Sol Thompson

100 94 100 89 (25) (100) (100) (99) (96)

100 94 100 — (25) (—) (100) (—) (—)

Copyright © 2012. Nebraska Paperback. All rights reserved.

1. Mean of samples from a systematic grid of viewsheds across the site. 2. 1.5 m above uncorrected reservoir surface. 3. Thoburn’s (1931) height of 14.5 m, which is likely a great overestimation. Note: Values within parentheses represent viewsheds from mounds that were not used to generate parameters for the modeled areas.

view from these mounds consists of bottomlands, primarily the Poteau/ Arkansas River bottomlands between Spiro and Cavanaugh. The view from the base of Cavanaugh is fairly large within the sample (ranked 81). The views from the summits of all of the mounds within this region are ranked quite high. The size of the viewshed from the top of Round Mountain stands out as larger than any created for this study. Within Arkansas River Region II, neither mound used in constructing the modeled area has a highly significant viewshed from its base. The mound at Point Remove has the largest viewshed of any sample from its summit; at 9 m it is one of the tallest mounds within the entire study area. The size of the viewshed from Carden Bottom Cemetery is small within the sample (ranked 17), and it gains significantly with a modest height of 1 m, but it is still near the median of the generated viewsheds. Within Neosho River Region I all three mounds at Harlan have highly ranked viewsheds. The Harlan site is not within the floodplain of the Arkansas River itself but is on a slight topographic rise within the bottoms of Fourteen Mile Creek, about 4 km upstream from its confluence with the Arkansas River. The mounds at the Norman site have average to slightly larger base viewsheds and significantly large viewsheds from the summits. Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

161

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

6-6. Mound viewsheds ranked against the statistical background of 99 generated samples. Mound names followed by an asterisk were not used to generate parameters for predictive model areas for the statistics. Viewshed from the Loftin site was generated 1.5 m above an uncorrected reservoir surface, so no summit viewshed was generated. Circled points indicate mounds within the Ozark Plateaus region.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Copyright © 2012. Nebraska Paperback. All rights reserved.

6-7. Relief map showing viewshed from the summit of Brown mound at Spiro (shaded white). Sites are labeled in fig. 6-6.

6-8. Relief map showing viewshed from the Craig mound at Spiro. Areas shaded dark are visible from the base of the mound, while areas shaded white represent the gain in viewshed from the mound summit. Sites are labeled in fig. 6-6.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

6-9. Relief map showing viewshed from the Cavanaugh mound. Shaded areas are visible from the base of the mound, while white areas represent the gain in viewshed from the mound summit. A to A’ locates the topographic profile depicted in fig. 6-10. Sites are labeled in fig. 6-6.

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Hughes #1 and Fort Davis sites are situated on the unique landscape position of a high bluff overlooking the confluence of three major streams and have viewsheds larger than any other mounds within the sample. The Cat Smith mound also has a large viewshed, but the appropriateness of the sample to this site is questionable, because the Cat Smith site is within a reservoir that remains uncorrected in the elevation model and the estimated viewshed may simply be wrong. Within Neosho River Region II, none of the three mounds used to construct the predicted viewshed area have significantly large or small viewsheds. This is true at the Lillie Creek site, even using an estimated and certainly unrealistic height of 14.5 m (after Thoburn’s 1931 estimated height). Within the Ozark Plateaus Region none of the five mounds used to construct the predicted area model have significantly large viewsheds, nor do any of the other four mounds within the region which were not used to construct the model. In fact, the platform mounds at the Huntsville and Pineville sites have viewsheds that may be considered signifi164

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

cantly smaller than would be expected (Huntsville ranked at seventh from the base and eighth from the summit, while Pineville ranked fourth from both the base and summit). Even when mound heights are taken into account, viewsheds from three of the mounds within this region fall below 90 percent of the generated samples. If the predictive model for mounds within this region is accurate, the Goforth-Saindon, Huntsville, and Pineville sites appear to be located for preferentially small viewsheds. Within the Ouachita Mountain Region all four of the mounds used to construct the predicted area have large viewsheds compared with the statistical background, and two of them (Page and Bluffton) have viewsheds larger than any generated sample. Assuming that the predicted area is a realistic approximation of mound locations within this region, these sites are clearly situated preferentially in locations with large viewsheds. The Logan Eddy and Borrow Pit sites are also situated in locations with large viewsheds, but not quite so significantly (ranked ninetyfourth and eighty-ninth from the mound bases, respectively). Viewshed Morphologies and Potential Intervisibility

Copyright © 2012. Nebraska Paperback. All rights reserved.

Aside from the purely quantitative analysis of viewshed size, a qualitative look at what is visible on the landscape from mound locations is also instructive. The intervisibility of viewsheds in the Spiro area (Arkansas River Region II) is particularly interesting and contrasts sharply with the Norman and Harlan sites (Neosho River Region I), which are concealed from one another. Viewsheds between sites in other regions show little potential for intervisibility. Arkansas River Region I Arkansas River Region I is centered around the Poteau/Arkansas River bottoms between Spiro and Cavanaugh. Figure 6-7 depicts the viewshed from the Brown mound, the largest platform mound at Spiro, in relation to the landscape and to the Skidgel and Cavanaugh mounds. The viewshed from Copple mound, the other platform mound at Spiro, overlaps almost completely with that from the Brown mound. The view is primarily directed to the east, overlooking the Poteau/Arkansas River bottoms. Close to half of the area between the Spiro and Cavanaugh sites is within the viewshed. Slight undulations in topography within the floodplain Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

165

Copyright © 2012. Nebraska Paperback. All rights reserved.

and low terraces control which areas are in and out of view within the bottomlands, and these would have been quite different in prehistoric times: the most obvious difference is the straightening of the Arkansas River for navigation (visible in fig. 6-7 running northeast from Spiro). Still the amount of bottomland in view would likely not have been significantly different, and regardless of exactly what bottomland locations would have been visible, it is clear that the view from Spiro is directed largely to the east, overlooking the bottoms. Large trees now obstruct the view from Spiro north and east into the bottomlands, and if any large trees were present just east of the site in prehistoric times they would have obstructed the view then as well. The view from the site today is directed to the south, where Round Mountain, Sugar Loaf Mountain, and other distant hills are silhouetted rather dramatically against the skyline, as they likely would have been in prehistoric times as well. The Skidgel mound is located on a topographic prominence overlooking the bottomlands to the west, but the largest extent of its viewshed is still directed to the east, overlooking the Poteau/Arkansas River bottoms. Similar to the viewshed from the Brown Mound, several of the hills well beyond the bluff line of the Arkansas River are within view, both to the south and north. The viewshed also encompasses much of the bottomlands west of Skidgel (the Redland Bottom). The high, steep northern bluff of the Arkansas River southwest of Skidgel obstructs the view into the bottoms farther up river. Copple mound at the Spiro site is the tallest mound in the region, situated on a very low terrace just a few meters above the floodplain. The view from the ground at Craig (see fig. 6-8) is highly restricted in all directions; only high hills surrounding the Poteau/Arkansas River bottoms would have been visible. Even low vegetation near the site or on slight topographic rises to the west and south would have obscured much of this viewshed, so it is likely that it was even further constricted than is shown here. Craig is one of the tallest mounds in the region, however (12 m at its peak), and even a modest rise in elevation in this location increases the viewshed dramatically. Areas shaded white in figure 6-8 indicate locations within the viewshed of the top of the mound. Similar to the Brown mound at Spiro and Skidgel, the view from the Craig mound is primarily to the east, overlooking the broad Poteau/Arkansas 166

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

River bottoms. Cavanaugh mound is not in view from the base of the Craig mound but is in view from the summit. Cavanaugh is a solitary platform mound above the eastern edge of the Poteau/Arkansas River bottoms. From the base of the mound the view is relatively constricted (shaded dark in fig. 6-9), with a great deal of the view confined to an area less than 5 km from the mound. The rest of the viewshed is directed to the west and north, overlooking two stretches of the bottoms. The Arkansas River bottoms to the east are well out of site. A low topographic rise about 150 m west of Cavanaugh obstructs the view into the southern portion of the bottoms. A few of the hills above the bluff line to the north of the Arkansas River are visible as well. The viewshed from this location encompasses the northern half of the Spiro site and comes within about 90 m of the Brown mound. It would be unwarranted to assume accuracy of the model at this level of detail, however, as 90 m only spans three pixels in the elevation model. Even low vegetation would have obscured the view from Cavanaugh to any portion of the Spiro site. Skidgel Mound, even though it sits on a local topographic prominence, is also obscured from view from the base of the Cavanaugh mound. The view from the summit of the Cavanaugh mound is substantially larger (shaded light in fig. 6-9). The viewshed to the west overlooking the bottoms is much wider, and the Spiro and Skidgel sites are clearly within view. Even relatively tall vegetation in the bottoms (10 m or more high) would not obstruct this intervisibility. The narrow views to the north and east from the summit of the Cavanaugh mound are questionable; elevations along these paths are not much lower than the mound itself, and even moderately tall vegetation would have obstructed these views. The potential intervisibility between these mounds is illustrated in figure 6-10. The mounds at Skidgel, Spiro, and Cavanaugh would have all been visible to one another. Neither Spiro nor Skidgel would have been visible from the ground surface at the Cavanaugh Mound (and therefore ground-level objects at Cavanaugh would not have been visible from either the Spiro or Skidgel sites). The added heights of the mounds themselves account for the intervisibility. In terms of viewer experience, the implications are quite interesting: individuals on the ground at Skidgel, Spiro, and Cavanaugh would not have been able to see one another, but individuals standing on top of the mounds would have had an unobstructed view to and from all three locations. Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

167

6-10. Topographic cross-section (located from A to A’ in fig. 6-9) across the Poteau/Arkansas bottoms, with mound base and summit elevations.

Copyright © 2012. Nebraska Paperback. All rights reserved.

To travelers coming up the Arkansas River valley either on foot or by boat, the first mounds to come into sight would have been those at Spiro and Skidgel. Barring tall vegetation obstructing the view to either, they would have come into view at about the same time, just as the travelers came into the broad stretch of bottomlands immediately north of the Cavanaugh mound. Although Skidgel would have been a little farther away, it is situated on a local prominence and may have come into view first, silhouetted against either the sky or very distant hills. About halfway through the bottomlands (again barring any tall vegetation in the way), the Cavanaugh mound would have come into view, off to the left. To travelers coming down the Arkansas River valley, Skidgel mound would have come into sight upon entering the Redland Bottoms, clearly visible as a mound on an already existing peak, right at the top of the steeply rising bluff line. Cavanaugh mound and the mounds at Spiro would not have been visible until the travelers rounded the final bend into the Poteau/Arkansas River bottoms. Regardless of the vegetation, it is likely that from many (if not most) locations within the Poteau/Arkansas River bottoms, a person would have been able to look up and see a mound at one or more of the sites. Round Mountain is a prominent landscape feature along this stretch of the Arkansas River, and at 38,030 ha (see table 6-2), the viewshed from the top of the mountain is larger than any other of the more than 1,000 viewsheds generated for this study. Its view encompasses nearly everything that is in view from the other mounds (with the exception of a portion of the landscape east of the Cavanaugh mound). Almost all of 168

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

the landscape within the Poteau/Arkansas River bottoms is in view, as are many of the hills and ridges beyond the bottoms.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Neosho River Region I Neosho River Region I centers around the Norman, Harlan, and Hill/ Fort Davis sites. Figure 6-11 depicts the viewsheds from the summits of Norman mound I-1, Harlan mound unit 7, and Fort Davis. These are all platform mounds and have the largest viewsheds of each site (considering Hill and Fort Davis to be the same site). Views from the base of each mound are only slightly smaller. The view from the Norman site is directed up and down the Neosho River, encompassing a great deal of the bottomlands within 5 km either way. The view from Harlan is directed primarily to the north and east, overlooking the bottoms of Fourteen Mile Creek. A narrow view south from Harlan extends to the Neosho River valley. The view from Fort Davis overlooks a large expanse of bottoms where the Arkansas, Neosho, and Verdigris Rivers meet, extending more than 10 km to the west and south. Areas of overlap between the viewshed are circled and labeled a through c on figure 6-11. Area a represents a minor overlap between the Norman and Fort Davis mound sites. Even low vegetation would have obscured the view to one or more sites from this location, however, and Fort Davis is over 20 km away and may have been difficult to discern against the background, even on clear days. Areas b and c are the only locations of overlap. The overlap in both cases is very minor, however (only a few pixels at each location), and even minor vegetation would have completely obscured the views to one or all of the sites from these locations. Area e represents a minor viewshed overlap between the Fort Davis and Harlan mound sites. This location is on top of a sharply rising bluff line about 100 m above the local bottomlands, with a straight view southwest down the Neosho River valley to the Fort Davis site; even relatively tall vegetation in the bottoms would not have obstructed this view. The view to Harlan is more questionable, as the line of sight passes over upland ridges, and even relatively short vegetation would have obstructed the view. Area d on figure 6-11 represents the most likely candidate for a point mutually visible to and from the sites. This represents a relatively large overlap in viewsheds between the Norman and Harlan sites on the slopes and ridge of a 90 m high bluff overlooking the bottoms of the Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

169

6-11. Relief map showing viewsheds from Mound I–II at Norman (shaded black), Mound Unit 7 at Harlan (shaded gray), and Fort Davis (shaded white). Circles represent areas of overlap discussed in the text.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Neosho River and Fourteen Mile Creek. From within this location it would likely have been possible to view both the Harlan and Norman sites simultaneously. For travelers coming up the Arkansas River (barring tall vegetation), Fort Davis would have been visible from a distance of about 8 km, as the travelers rounded the last meander bend into the broad bottoms of the three rivers. Traveling down the Arkansas River, Fort Davis would also have been visible for 10 km or more. In both cases, the mounds would have been some of the tallest features on the immediate landscape, silhouetted against very distant hills. This is somewhat similar to the view of Skidgel mound that travelers would have had approaching from upstream. The Norman and Harlan sites would both have been visible to travelers from a distance of about 5 km from either upstream or downstream. 170

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

It is unlikely that both would have been visible at the same time to travelers in the bottomlands.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Concealed to Revealed Landscapes through Mound Construction The Cavanaugh Mound (a platform mound) and Craig mound at the Spiro site, as discussed above, are in relatively concealed areas of the landscape from the surrounding bottomlands, but the construction of the mounds makes these locations significantly more visible, revealing mound-top structures in a location that would otherwise have been hidden from the broad floodplains they border. Other mounds appear to express a similar dynamic: Mound A at Goforth-Saindon, Ewing Chapel Cemetery, and possibly the Hughes/Fort Davis mounds. The Goforth-Saindon site is located on a low terrace at the confluence of the Illinois River and Cincinnati Creek. Low rises surrounding Mound A (the large platform mound at the site) obstruct the view into the long stretch of bottoms to the northeast and the broad bottoms at the confluence of the streams to the southwest. Even low vegetation would further obstruct the view. The construction of the mound, however, would raise a viewer on the mound over the tops of moderately tall vegetation and allow for a much larger view in both directions. To travelers approaching the site from any of the bottomlands, the area of the site itself would be hidden, while the mound (and especially any structure on top of it) would have been readily apparent. The viewshed from the Ewing Chapel Cemetery mound at the ground surface is directed primarily up into the hills east and west of the site. With a sharp terrace scarp to the east, very little is visible of the bottoms of Evansville Creek to the east and north. Only with the addition of the mound height are the bottomlands visible from the site area; ground observations confirm this analysis. Figure 6-12 is a photograph taken from the bottomland about 2 km north of the Ewing Chapel Cemetery mound that confirms the accuracy of the gis viewsheds in this respect. The mound itself is visible in this photograph, but the ground immediately surrounding it is not. Individuals standing on the mound are visible from this location, as are even tall (1.5 m) gravestones. Individuals standing on the ground in front of the mound are not visible because of the rise of the terrace. Fort Davis mound apparently holds a similar relationship to the landscape, although this is not strongly expressed in the generated Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

171

6-12. View from the bottoms approximately 2 km north of Ewing Chapel Cemetery. Arrow points to a 1 × 1 m target held by a person standing on the mound. People standing on the ground in front of the mound are not visible from the bottomlands. The mound itself is visible as slightly lighter-colored ground beneath the copse of trees in the middle of the photograph.

Copyright © 2012. Nebraska Paperback. All rights reserved.

viewshed. Overlooking the broad bottomlands at the confluence of the Arkansas, Verdigris, and Neosho Rivers, the mound was used during the Civil War by the Confederacy as a lookout point (within Fort Davis), with a view down into Union Fort Gibson across the river to the east. Foreman (1939:148) wrote: Fort Davis was built around a prehistoric mound standing some 25 or 30 feet high above the surrounding terrain. It was on the crest of a gentle elevation that sloped both east and west. On the west slope is was possible to drill and assemble a large force of men, horses, and mules, screened from observation from Fort Gibson. The east slope fell away gently toward the Arkansas River and Fort Gibson, leaving an unobstructed view across the river to the old fort and to the surrounding prairies. With this advantage the confederates could hide most of their men from the field glasses in the hands of union observers in Fort Gibson; while at the same time they had an excellent view of the surroundings of the latter fort, especially of the lands where their live stock was pastured. 172

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

While this does not imply that prehistoric mound locations were selected for defensive purposes, this passage highlights a complex landscape/viewshed relationship that several of the mounds apparently held. The mounds were (and many still are) impressive features on the landscape, and as monumental architecture they were certainly meant to be viewed or, as Bradley (2000:158) writes, “were originally conceived with the aim of conveying ideas to a large audience.” They also appear quite suited for concealing ideas from large audiences. It is possible that other mounds have a similar relationship to the landscape which is not revealed here because the exact mound locations are uncertain or because of imprecision in the elevation model. While difficult to quantify, the spatial “windows” for the mounds to express this relationship are relatively small. Had these particular mounds been constructed slightly closer to the terrace scarps that conceal their ground-level positions from large portions of the bottomlands, each would have held an impressive and commanding position over these areas. Individuals walking in front of the mounds would have been visible from the bottomlands. Had these mounds been constructed slightly farther from the terrace scarps, each would have been completely hidden from the bottomlands. As it is with these sites, portions of the mounds are both concealed from and revealed to large portions of the immediately surrounding bottomland.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Conclusions Are the mounds located preferentially with respect to viewshed size? The brief answer appears to be yes for at least some of the mounds, particularly in the larger river valleys and at the larger (third echelon) sites. Viewsheds from the landscape position of at least 3 of the mounds (Skidgel, Page, and Bluffton) rank significantly higher within the Monte Carlo analysis than all 99 samples generated for their regions, and viewsheds from 10 of 27 of the mounds (37 percent) rank above 90 percent of their samples. When mound heights are taken into account, exactly half of the mounds have viewsheds in the top 5 percent of the samples. While the generated samples were not designed to serve as a statistical background for mound summit viewsheds, this does demonstrate the prominence of many of the mounds on the landscape. Of mound viewsheds in the large river valleys (the Arkansas and Neosho River Regions), 14 of 16 (88 percent) rank higher than 50 percent of the generated samples, and 7 of them (44 percent) rank higher than Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

173

Copyright © 2012. Nebraska Paperback. All rights reserved.

90 percent of the generated sample. This pattern becomes even more striking when viewsheds from only the largest mound centers are considered (Norman, Harlan, and Spiro). Of viewsheds generated from the locations of mounds at these sites (10 in total, including Skidgel and Cavanaugh as part of a larger Spiro complex), 6 rank in the top 10 percent of the samples, and the lowest ranks at fifty-sixth of the 99 generated samples. When mound heights are taken into account, all 10 rank in the top 10 percent of generated samples, and all but 1 (Mound I–II at Norman) rank at 98 percent or above. These mound locations, as a group, have significantly larger viewsheds than would be expected by chance alone. When the mound heights are taken into account, they become some of the most visibly prominent features on the entire landscape. Viewsheds from the locations of three mounds (Pineville, GoforthSaindon, and Huntsville) rank in the lowest 7 percent of generated samples, and even when the mound heights are taken into account, they all still rank in the lowest 11 percent. These mounds are all within the Ozark Plateaus Region, within which viewsheds from none of the mounds ranks particularly high. These rankings from the Goforth-Saindon and Huntsville mound sites are at first somewhat counterintuitive, because these are two of the largest and most elaborate mound sites in the Ozark Plateaus. Instead of selecting for locations with large viewsheds, it is possible that mound locations in this region were selected for smaller ones. In light of this, Kay et al.’s (1989:137) impression of Ozark Plateaus mound centers is put in fuller context: “The visual effect is that of being on stage of a large amphitheater, as one has a commanding view of the valley and flanking uplands from any of the mounds.” The view of the flanking uplands is commanding because the uplands are so close as to restrict the view. If the mounds “were originally conceived with the aim of conveying ideas to a large audience” (Bradley 2000:158), it does not necessarily imply that they must have been visible from large portions of the landscape compared with other natural features, just large enough to be visible to the intended audience. The landscape relationship of Ozark Plateaus mound centers perhaps implies a more intimate feel than centers in the major river valleys with much larger viewsheds. The viewshed from Round Mountain, near the Spiro site, is a full order of magnitude larger than the majority of mounds and may also deserve further attention. Bradley’s 2000 argument that constructed mounds may invest already significant natural places with additional layers of symbolism, in this case, may apply to the “place” of the Poteau/ 174

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Arkansas River bottoms and surrounding uplands as a whole. Morrow’s 2004 observation that mounds at the Spiro site seem to mimic natural features, including Round Mountain, points strongly to this potential. Viewshed dynamics may also serve as one line of reasoning in studies of mound site relationships. Different portions of the Spiro site, for example, are officially recorded under separate state site numbers in the Oklahoma Archeological Survey site files, although they are commonly considered a single site. The solitary platform mound sites of Skidgel to its west and Cavanaugh to its east, however, are often presented as separate sites. In light of the mutual intervisibility of these sites, due in large part to the construction of the mounds, it may be more appropriate to consider all three part of a larger site complex. The relationship of the Norman and Harlan sites has also been of great interest. Other than the Spiro mound site, these are the two largest and most elaborate mound centers in the region, yet they are scarcely 5 km apart. These sites were previously interpreted as sequentially occupied, but recent analysis of radiocarbon dates (Cranford 2007) demonstrates a considerable temporal overlap in their construction and use. Although the straight-line distance between the sites is small, a tall ridge prohibits intervisibility. There is one position on the landscape, however (a few kilometers south of both sites, area d in fig. 6-11), from which both sites would have been clearly visible. Such points of mutual visibility may warrant further attention. Other chapters in this volume (by Brooks, Lockhart, and Walker and McKinnon) explore mound and habitation site relationships with the surrounding landscapes and multiple environmental variables. While there is little geographic overlap between these chapters (see fig. 1-3), some interesting points of intersection can be noted, particularly with the chapters by Brooks and Lockhart. Brooks, for example, concludes that regions in southeastern Oklahoma with wide river valleys provided enough resources to allow Caddo communities more room and flexibility in creating a built environment, while regions with narrower river valleys may have limited this ability. This dynamic may be mirrored in the larger viewsheds of the Neosho and Arkansas River regions, with wider valleys and more potential choices for mound site locations. An opposing case may be made, however, in the viewsheds of the Ozark Plateau region. River valleys here are more constrained along low-order streams, where smaller viewsheds would be expected by chance alone, but mound site viewsheds are statistically Viewshed Characteristics of Caddo Mounds

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

175

Copyright © 2012. Nebraska Paperback. All rights reserved.

smaller even than chance would predict. If smaller viewsheds leading to an amphitheater effect were an important consideration, the rugged topography of more constrained valleys offers greater flexibility of choice. Brooks also brings up the notion of gateway or gatekeeper communities at key points along linear valley systems. One important component of such a community that would allow it to serve effectively may have been simple visibility: from a practical point of view, it is difficult to control or manage unseen commerce. From a broader perspective, if the constructed landscape of these sites helped to serve as the gateway to the community, it must have been seen in order to do so. As with mound sites in the Arkansas River basin, the important thing may not be a simple matter of the size of the viewshed but the nature of the viewshed in relation to the valley system. Lockhart’s study of mound sites in the West Gulf Coastal Plain of Arkansas presents a powerful model of site locations based on numerous environmental variables. The mound sites in this region are primarily situated along river valleys, but 13 anomalous mound sites are found at higher elevations near or directly on watershed divides. Located at some of the highest elevations within the entire region, the viewsheds of these sites are likely quite large, with the potential for long-distance intervisibility between the mounds or other important sites. gis-driven research concerning viewsheds and other “experiential” aspects of landscapes are still in their infancy, as is gis as a whole. Advances in computer technologies and spatial modeling software will almost certainly render many of the technical aspects of current research obsolete within a relatively short time. Such rapid obsolescence may be more encouraging than discouraging, however: revisiting old data sets with new technologies and theoretical orientations has never held greater promise than today. I only hope (rewording Brown 1992) that our theoretical assumptions do not promise more than increased computing speed can endure. Acknowledgments Marvin Kay, George Sabo III, Ken Kvamme, and Peggy Guccione were committee members of the PhD dissertation from which this work is drawn, and their help and encouragement are still greatly appreciated. Much of the work for this project was made possible by a grant from the National Science Foundation (ddi 0341068).

176

Vogel

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

7. Exploring Prehistoric Caddo Communities through Archaeogeophysics chester p. walker and duncan p. mckinnon

Copyright © 2012. Nebraska Paperback. All rights reserved.

Introduction Our understanding of how the prehistoric Caddo Indians organized their communities across natural and cultural landscapes has been heavily influenced by two primary source images. These are the 1691 map of the Upper Nasoni community drawn by an anonymous scribe from the Domingo Terán de los Ríos expedition to the Red River (in what is now northeast Texas, see Sabo, this volume), and a series of late nineteenthcentury historic photographs taken by William S. Soule of the Smithsonian Institution. The Terán map depicts a dispersed Caddo community of partitioned farmsteads spread out over several kilometers along the Red River. Each partitioned farmstead consists of one or more large circular thatchedcovered dwellings with open-air arbors, storage structures or granaries, and ramadas. Located on the western periphery of the community is a large temple mound. A vacant area surrounds the mound, with a templo or temple on the mound and an open-air structure at the base. The community is presumed to be the Hatchel-Mitchell-Moores archaeological site complex located in Bowie County, Texas (Wedel 1978:14; see Perttula 2005). The Soule photographs were taken between 1868 and 1872 of the farmstead of Chief Long Hat near Binger, Oklahoma, complete with beehive-shaped house clusters, outdoor arbors, and storage shelters. The combination of farmstead structures in the Soule photographs is quite similar to the combination of structures in the Terán map farmsteads, despite the two centuries that separate the two sources.

177

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

These two sources are often conceptualized as ethnographic examples of Caddo community spatial organization across the landscape (cf. Lockhart 2007; Schambach 1982a; Trubowitz 1984). However, they have been overused in investigations and characterizations of community organization at Caddo archaeological sites, in part as a result of the scarcity of primary archaeological data of Caddo community organization, as they only represent a synchronic view of Caddo community spatial organization. Currently the most completely excavated Caddo community in the Caddo archaeological area may be the Oak Hill Village site (41RK214) in east Texas (Perttula and Rogers, chap. 8, this volume; Perttula and Rogers 2007). Prior to the remote sensing or archaeogeophysical “revolution” in Caddo archaeology (Perttula et al. 2008), the excavations at the Oak Hill Village site provided the most complete archaeological view of the organization and evolution of a prehistoric Caddo community. Now, however, recent archaeogeophysical investigations at a number of Caddo sites have obtained primary data sets of geophysical features that allow a more detailed consideration of community organization at a landscape scale as well as a better appreciation of the character and intrasite distribution of Caddo architectural features. It is becoming increasingly economically feasible to use archaeogeophysics as a complementary investigative tool in the study of Caddo community spatial organization across the landscape. Kenneth Kvamme (2003) was the first North American archaeologist to urge a shift from using archaeogeophysics exclusively as a tool for locating cultural anomalies to an approach where it is holistically integrated as an investigative strategy with archaeological investigations and interpretations. This paradigm shift concerning how to approach the study of the overall spatial structure of prehistoric Native American communities is one that is unlikely to be achieved solely through modern excavation strategies, if for no other reason than cost. With expansive archaeogeophysical investigations of large coverage areas at sites, it is often possible to reveal the spatial arrangements of individual structures in addition to various attributes about their geometric shape, spatial orientation, and interrelationships. Furthermore, it is often possible to gather archaeogeophysical data suitable for interpreting the character of interior features within a structure, such as fire hearths, storage pits, or burials underneath house floors. The ability to 178

Walker and McKinnon

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

acquire information on the settlement structure of prehistoric American Indian sites over large coverage areas, compared with slow-paced, expensive, and spatially limited excavations, has affected the way archaeology is being conducted today in North America and will have an even larger effect in the future. In this chapter, we discuss the archaeogeophysical findings from the George C. Davis (41CE19), Hill Farm (41BW169), and Battle Mound (3LA1) sites (see figs. 1-2 and 1-3) to highlight the various architectural attributes of Caddo communities that can be defined using landscape archaeogeophysical methodology. At the most basic and perhaps analytically most powerful resolution, the use of archaeogeophysical data as a primary data set in landscape archaeology studies permits a comprehensive image of overall site spatial layout and organization of structures, mounds, and other cultural and geological features that are difficult to identify using traditional archaeological methods. Where excavations have the ability to confirm interpretations of individual archaeogeophysical anomalies, a broad-scale map of the overall site spatial layout from archaeogeophysical investigations complements these data sets with an analysis of spatial relationships and patterning of cultural features across an entire site.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Archaeogeophysical Analysis of Caddo Architecture The most basic characterization of architectural attributes is geometric shape. The identification of shape is often crucial during preliminary data analysis and the creation of an interpretive framework that lends archaeological meaning to geophysical data sets. The identification of geometric shapes is important in the recognition of patterns in the data that can be attributed to a cultural origin (Kvamme 2008:68). Analyzing the spatial orientation of individual geometric shapes is a process that also contributes to archaeological analysis and interpretation. The orientation and arrangements of architectural attributes provides archaeological information at many levels, particularly with respect to the larger patterns that are defined by a comprehensive view of subtle intrasite orientations. The locations and arrangements of interior features (hearths, storage pits, burial pits) provide high-resolution archaeogeophysical data about the spatial organization of specific structures, while also contributing a significant amount of landscape information. At the structure level, the Exploring Prehistoric Communities

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

179

Copyright © 2012. Nebraska Paperback. All rights reserved.

identification of interior features offers a starting point for the pursuit of traditional household scale studies (i.e., Early 2000; Kelley 1997). At the landscape scale, the patterns and arrangements of internal features in different structures provides insights useful for analyzing different architectural styles and their social function (i.e., residential vs. public structures). For example, the identification of certain anomalies as extended entranceways to structures will influence interpretations of their social function, given current views of the purpose of extended entranceway structures in prehistoric Caddo communities (e.g., Early 1988, 2000; Kay and Sabo 2006; Perttula 2009). There are several architectural styles of Caddo structures present at the George C. Davis, Hill Farm, and Battle Mound sites, as described in this chapter. At all three of the sites, the basic architectural style is a circular structure ranging from 5 to 20 m in diameter. These circular structures are defined at the George C. Davis site by a circular low magnetic signature, believed to have been created by negative relief features extending into a magnetically enhanced ferrous subsoil filled with less magnetic soil (Bruseth et al. 2007; Creel et al. 2005, 2008; Osburn et al. 2008). At the Hill Farm and Battle Mound sites the circular structures are defined by a circular pattern of magnetic high values that are interpreted as the effects of high heat burning events (Walker and Schultz 2008; Perttula et al. 2008; McKinnon 2008, 2009, 2010a). Additionally, at the Battle site, many of the circular structures are defined by a concentric circle pattern of low magnetic signature, suggesting an architectural style composed of sand berms or larger partitioned structures (McKinnon 2008, 2009, 2010a). At the George C. Davis site, a central hearth surrounded by four interior roof support features is representative of an additional architectural style. These structures have been termed “Button Houses” and are fairly unique to the site (Wilson and Schultz 2009; Schultz 2010). At the Hill Farm and Battle sites a few circular structures appear to have extended entranceways (McKinnon 2008, 2009, 2010a; Perttula et al. 2008). Such structures have been archaeologically identified throughout much of the southern Caddo area (Perttula 2009). George C. Davis Site (41CE19) The George C. Davis site, on a large alluvial terrace of the Neches River (see fig. 1-2), is one of the best known and most extensively excavated 180

Walker and McKinnon

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Caddo sites in east Texas. Since the publication of the late 1930s and early 1940s Works Progress Administration (wpa)–University of Texas (ut) excavations at Davis (Newell and Krieger 1949), the site has loomed large in our understanding of the Formative Caddo and Early Caddo periods (A.D. 800/850–1200) and the findings from the site have framed much of what is known about Caddo native history in this part of the Caddo area, as well as Caddo monumental architecture, ceramic technologies, and ritual interment of the dead (Newell and Krieger 1949; Spock 1977; Story 1972, 1981, 1998, 2000). The majority of the intensive investigations over the years have focused on the three mounds constructed by the Caddo there (Mounds A–C; Mounds A and B are platform mounds, while Mound C is an accretional burial mound) and their adjacent architectural features. Archaeological excavations have documented some 51 structures, both domestic and ritual in use, at George C. Davis (Spock 1977; Story 1998:26; Schultz 2010). More recently, the Davis site has proven to be ideal for evaluating the utility of archaeogeophysical methods on Caddo mound centers. This work has located additional Caddo structures and has added much new information on the intrasite arrangements and architectural detail of structures (Bruseth and Pierson 2004; Creel et al. 2005, 2008; Osburn et al. 2008).

Copyright © 2012. Nebraska Paperback. All rights reserved.

Archaeogeophysical Investigations at the George C. Davis Site The geophysical research at the George C. Davis site presented here represents a group effort by the Texas Archeological Research Laboratory at the University of Texas at Austin and the Texas Historical Commission. In 2002, a ground penetrating radar and magnetometer survey of portions of this site was initiated to search for features noted in earlier excavations (Creel et al. 2008:177). The magnetometer survey was successful in locating several Caddo structures (Bruseth and Pierson 2004; see also Bruseth et al. 2007). This initial work was also instrumental in helping to start what has been referred to as a “revolution in Caddo archaeology” by ushering in the widespread use of geophysical methods in East Texas Caddo archaeology (e.g., Perttula et al. 2008; Walker and Perttula 2008). The magnetometer survey was continued in 2003–2004 to cover the full extent of the property (then owned by the Texas Parks and Wildlife Department), as well as a small portion of private land to the south of the park (Creel et al. 2005, 2008; Schultz 2010; Walker 2009) (fig. 7-1). Exploring Prehistoric Communities

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

181

Copyright © 2012. Nebraska Paperback. All rights reserved.

7-1. The magnetometer survey areas at the George C. Davis site (41CE19).

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

In 2008, a 1-hectare portion of the site east of Highway 21 on Texas Forest Service land was surveyed, and this work located several structures directly adjacent to Mound A (Osburn et al. 2008). An additional survey was done here in 2011 by Walker.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Geophysical Architectural Features Currently there is a greater concentration of geophysical features of an architectural nature in the area directly east of Mound B at the Davis site, with an additional concentration north of Mound A; additional geophysical surveys south of Highway 21 may change that finding. These architectural features have three types of structures: button houses, circular houses, and sub-square houses (fig. 7-2). Button houses range from 10 to 15 m in diameter. The hallmarks of the button house are four low magnetic circular anomalies arranged in a square pattern with a single circular high or dipolar anomaly in the center (fig. 7-3). This grouping of anomalies no doubt represent four large interior post hole support features surrounding a central hearth. Several of the button houses recorded in the geophysical survey are located to the east of Mound B intermixed within the dense cluster of structures in that area. There is also a second concentration of button houses located in the southern portion of the site northwest of Mound A, which is an area that otherwise seems to have been much less intensely occupied (see fig. 7-2). Feature 42 beneath Mound A is an unequivocal archaeological and architectural match to the button houses identified in the magnetic data (see fig.7-3). The partially excavated structure was in the south central corner of Mound A, under the mound (Spock 1977:89; Newell and Krieger 1949:44). The structure had 51 exterior and 45 interior post hole features and four interior and oval pit features that were filled with a reddish soil matrix and contained a small amount of ceramic sherds. Feature 31, also from Newell’s Mound A excavations, appears to be a button house in the southwest corner of Mound A and also under the mound (Spock 1977:59; Newell and Krieger 1949:fig. 15). This feature, superimposed on two earlier structures (features 37 and 38), is made up of 78 exterior and 31 interior posts and measures approximately 14.9 m in diameter (Spock 1977:59). The four internal pits were arranged in a pattern similar to the interior pits from feature 42 as well as the button houses identified in the geophysical data (see fig. 7-2). In addition to the interior pits, feature 31 had a central hearth that was round and Exploring Prehistoric Communities

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

183

Copyright © 2012. Nebraska Paperback. All rights reserved.

7-2. Interpretive map showing the architectural features identified in the magnetometer surveys and past archeological excavations.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

7-3. Examples of button houses at the George C. Davis site: a–c. excavated examples; features 31 and 42 are underneath mound A; d–e. geophysical examples.

basin-shaped. The hearth contained a burned sandy clay fill that “when first exposed appeared as an area of bright burned red and black soil on the floor” (Spock 1977:62). A prepared lining of burned clay covered the hearth with the exception of a center post hole. A comparison of feature 125 north of Mound B, another button house (Story 1997:fig. 42), with features 237 and 242 defined in the geophysical work (see fig. 7-3; Schultz 2010), also indicates that there is great consistency in the shape and internal architectural characteristics of these features. There are several clearly definable low magnetic circular anomalies at the Davis site that measure 5–20 m in diameter. These circular anomalies, or in some cases partial arcs, can be readily attributed to the well-known circular Caddo architectural style of house construction (see Schultz 2010). A negative relief feature with a fill that has a lower magnetic signature than the surrounding clay subsoil creates these anomalies. If these anomaly types do actually represent Caddo houses, four would Exploring Prehistoric Communities

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

185

Copyright © 2012. Nebraska Paperback. All rights reserved.

represent some of the largest Caddo structures identified to date, being up to 20 m in diameter (features 238, 255, and 265 on fig. 7-2; see also Osburn et al. 2008:table 1). Circular houses are the most widely distributed architectural form at the Davis site. In the portion of the site north of Highway 21, circular structures, as well as partial arcs that are probably partial signatures of additional circular structures, are concentrated in the area east and southeast of Mound B as well as the area between Mounds B and C. Southeast of the highway, there is also a concentrated area of circular structures (see fig. 7-2). There are many excavated circular structures at the Davis site to use for comparison with the structures detected in the geophysical survey. There are 47 circular structures and 15 arcs or partial circular structures in the current magnetometer data set (Schultz 2010; Walker 2009); this is 81 percent (n = 77) of the total number of identified geophysical features. Sub-square houses are simply circular type anomalies with truncated or rounded corners. This structure type is well represented in the excavations from the Davis site (Spock 1977), and a few are present in the geophysical data set. Sub-square structures are largely clustered just to the east of Mound B. However, two sub-square geophysical features (features 2 and 7) are located in the cluster of structures north of Mound A (see fig. 7-2; see also Osburn et al. 2008:table 1). Sub-square structures are much more limited in their distribution in both the excavated as well as the geophysical investigations. Nine percent of the geophysical features are sub-square in shape (Schultz 2010). Feature 139, about 60 m north of Mound A, is a clear example of a subsquare structure. It is 10.3–10.4 m in diameter and formed by 40 posts. Story (1997:86) noted that the posts slanted outwards, were oval, and were filled with sediments much browner than other post holes documented in the excavations. Story (1997:86) proposed a connection between the unique architectural characteristics of feature 139 and the known historic architecture form termed by Griffith (1954:99–100) as a “beehive” house, which was constructed using bent poles. Community Organization Combining the geophysical results with those from the hand and machine-aided excavations at the George C. Davis site has resulted in the most comprehensive architectural database for an early Caddo commu186

Walker and McKinnon

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

7-4. The excavated and geophysical architectural features from the George C. Davis site, including the structures excavated under Mound A. The dots represent probable hearth features.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

7-5. Locations of possible plazas and community spaces at the George C. Davis site. These areas are delimited by dashed lines.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

nity (fig. 7-4). This may also represent the most complete geophysical and archaeological map of any one single Caddo site. Spatially plotting the Caddo architectural remains at the Davis site defines a complex arrangement of areas of continued structure use and rebuilding over several hundred years in the life of the community as well as areas that appear to have been left vacant. These vacant areas may represent a series of possible plazas and community spaces. Plazas and/or community spaces are labeled in figure 7-5 with dashed lines to suggest their possible locations based on the clusters of houses. Delineating plazas through the use of geophysics has been possible for a few other Caddo sites in east Texas (Perttula and Walker 2008; Walker 2009). At the Davis site the possible plaza areas or community spaces are defined due to their proximity to mounds and clusters of houses as well as the absence of other archaeological and geophysical features. Directly south of Mound C is a large area with no features. This area has been adversely affected by feral hogs, thus making it difficult to define archaeological features (Creel et al. 2008:188). A large plaza (90 m east/west and 70 m north/south) is defined here by the absence of any clear evidence of structures and the area’s proximity to Mound C. The area surrounding Mound B appears to have been much more spatially and culturally complex, and there appears to have been a series of small possible plazas or community spaces to the north, east, and south of the mound. One of these possible small plazas lies to the north of the mound, with another slighter larger area to the south, both of which would have been accessed by the north and south ramps flanking Mound B. Four additional plazas or community spaces are identified to the east of Mound B and between a densely clustered series of structures. Hill Farm Site (41BW169) The Hill Farm site is a portion of the larger Hatchel site (41BW3) village and site complex (Perttula 2005). The results of this geophysical survey have identified clusters of houses and other architectural features in two Late Caddo period residential areas (Walker and Schultz 2008; Perttula et al. 2008) about 500–900 m from the Hatchel mound itself (fig. 7-6). These two clusters of architectural features can be contextualized by direct comparison with the 1691 Terán map (Perttula et al. 2008:103; see also Sabo, chap. 15, this volume) and further our understanding of the Hatchel–Mitchell–Moores site complex, occupied by the Nasoni Caddo Exploring Prehistoric Communities

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

189

Channel Lake

41BW169

Hatchel

Area B Ancient Abandoned Channel Area A

Bayo

u

0 100 200

Copyright © 2012. Nebraska Paperback. All rights reserved.

LEGEND Estimated Area Limits Mound

N

400

Meters

0

400

800

1000

Feet

7-6. The location of Area A and Area B at the Hill Farm site (41BW169), channel lakes of the Red River, and an ancient abandoned stream channel. The northern mound is the Hatchel. Prepared by Sandra Hannum for the author.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

from as early as the thirteenth century through the seventeenth century (Wedel 1978; Perttula 2005). European explorers visited the site complex several times, including the 1691 Spanish expedition led by Don Domingo Terán de los Rios (Hatcher 1932, 1999). The Hatchel complex refers to the series of archaeological sites that extend along the banks of the Red River near present-day Texarkana, Texas. This complex included as many as five mounds and numerous village areas that were organized into individual compounds (Perttula 2005:180). Archaeologically the site is best known for the wpa excavations conducted in the late 1930s by the University of Texas (ut) under the direction of William C. Beatty; however, the earliest work at the site was conducted by A. T. Jackson of the same institution (Creel 1996). The ut/wpa excavations explored the site’s main temple mound as well as village areas and burial plots (Perttula 2005:182).

Copyright © 2012. Nebraska Paperback. All rights reserved.

Archaeogeophysical Investigations at the Hill Farm Site Archaeogeophysical investigations were conducted at the Hill Farm site as part of a cultural resource management project for a proposed Red River levee revetment to be constructed by the U.S. Army Corps of Engineers, Tulsa District. The geophysical work began with broad-scale electromagnetic induction survey (Grealy and Conyers 2008) that located several relict meander scars of the Red River. These findings, along with available archaeological information about the Hill Farm site, were then used to target areas for shovel testing. Focused magnetometer surveys were conducted in areas identified by positive shovel testing (Walker and Schultz 2008). There is a modest amount of architectural variation noted in the magnetometer data from the Hill Farm site. However, all the structures can be divided into two architectural forms: circular houses with extended entrances and those without extended entranceways. Nine features are noted in Area A, the southern section of the site (fig. 7-7). Of these, eight are interpreted as domestic Caddo structures (features 1, 2, 4–8, and 13). Features 1, 2, and 4 are all large structures with central hearths. In these cases the magnetic patterning shows a series of positive magnetic signatures comprising the outer ring of the structures. Feature 1 is also interpreted as a burned feature due to the large amount of high magnetic returns associated with it.

Exploring Prehistoric Communities

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

191

7-7. Plan of Area A collection blocks and interpretation of the geophysical data.

Copyright © 2012. Nebraska Paperback. All rights reserved.

There are two structures present with extended entranceways, one in Area A (feature 5) (see fig. 7-7) and one in Area B (feature 9, fig. 7-8). Extended entrance structures have a circular magnetic trend with a linear anomaly radiating from a single point. These two structures measure 10–12 m in diameter. The structure in Area A opens to the northwest toward the Hatchel platform mound and the structure in Area B opens to the south-southeast toward Area A (see figs. 7-6 and 7-8). Extended entranceways are a fairly rare architectural feature of Caddo structures and have been documented throughout much of the Caddo area with directionality linked to cosmological beliefs about life and death (see Kay and Sabo 2006; Perttula 2009). Circular houses have a circular magnetic trend, which at Hill Farm is a positive magnetic anomaly. There are 10 circular structures identified in the geophysical data from the Hill Farm site and 2 additional possible circular structures. Area A has 6 circular structures (features 1, 2, 4–7), 5 of which appear to have central hearths, and 2 possible structures (fea192

Walker and McKinnon

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

7-8. Plan of Area B collection blocks and interpretation of the geophysical data.

tures 8 and 13), one of which has a central hearth (see fig. 7-7). Two circular structures are apparent in Area B (features 9 and 11 in fig. 7-8). These structures range between 8 and 12 m in diameter, with the exception of feature 4 from Area A, which is 17.5 m in diameter. Features 6-8 and 13 in Area A are more geophysically subtle features that are also interpreted as Caddo house structures. These four features consist of a central positive signature with a circular to semicircular series of magnetic signatures encircling the central feature. These Area A features are all less than 10 m in diameter but well within the size range for typical Caddo houses at the Hatchel site and at various other Caddo sites (see Kelley 1997; Perttula 2005; Schultz 2010; Story 1997; Webb 1959; Williams 1993). The central hearth varies in its magnetic patterning from structure to structure in Area A. For example, feature 1.1 has a positive signature, feature 2 has a central area that is a diffuse negative signature surrounded by a semicircular positive signature, and feature 4 has a small diffuse positive magnetic signature just off center. Feature 3 in Area A is a large di-polar signature interpreted as either a large borrow pit, where the underlying strata are of significantly higher magnetic susceptibility than the overlying strata, or a burned structure. Its small size (7 m in diameter) and its proximity to other houses suggest that it may represent a special use structure, perhaps a granary. Four features are noted in Area B (see fig. 7-8), two of which are Exploring Prehistoric Communities

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

193

interpreted as Caddo structures (features 9 and 11). Feature 9 is a circular series of complex monopolar to di-polar signatures. The feature has a di-polar signature in its center, suggesting the placement of a central hearth. Like feature 5 in Area A, this structure feature appears to have an extended entranceway. Feature 11 is also a Caddo structure with a monopolar positive magnetic signature in its center that likely represents a central hearth. The monopolar positive magnetic signatures that comprise the outside ring of the structures vary from solid high to lower and more diffuse magnetic values. Both features 9 and 11 appear to have been burned due to the high magnetic signatures that comprise much of their remote sensing spatial patterning. Features 10 and 12 have low diffuse monopolar linear magnetic signatures. These two features flank both the northeast and southwest sides of feature 11. They may be compound dividers or fences like those depicted on the Terán map (see Perttula et al. 2008:fig. 1) of the Nasoni Caddo village and identified at other Caddo sites (Williams 1993:42–46; Perttula 2005:192). Community Organization

Copyright © 2012. Nebraska Paperback. All rights reserved.

Given the circumscribed archaeogeophysical investigations completed at the Hill Farm site, the consideration of Caddo community organization is perforce rather basic. First, both structures with extended entranceways seem to open toward other parts of the community: feature 5 in Area A opens to the northwest and in the general direction of the Hatchel mound; feature 9 in Area B opens to the southeast toward the Area A residential compound. Second, there appear to be compound dividers or fences associated with feature 11 in Area B. Using the Terán map as a model for community organization at Hill Farm (fig. 7-9), the long linear anomalies may mark part of an individual Caddo farmstead compound. There is additional information to be gained from comparing the Hill Farm archaeogeophysical data to the Terán map. It is possible to pinpoint the very location of the Hill Farm site on the Terán map (Perttula et al. 2008:103), relying on the analysis of the decorated pottery from the site; the geomorphology of the area (Grealy and Conyers 2008), which places an old stream channel to the west of the site; and the locations of household compounds southeast of the Hatchel mound. The Terán map shows a set of two farmstead compounds in the correct spatial relationship to the Hatchel mound—divided into two discrete areas—with one 194

Walker and McKinnon

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

7-9. Detail of the western portion of the 1691 Terán map showing the old abandoned channel of the Red River and the household compounds that may represent the likely area of the Hill Farm site.

to the north and one to the south with a relict stream channel to the west (fig. 7-9). Perttula et al. (2008:103) have suggested that the two small compounds located on the Terán map southeast of the templo can be correlated with the Hill Farm site. Battle Mound (3LA1) Battle mound is located on the Red River in Lafayette County, Arkansas (see fig. 1-2). The current river channel is located about 1.5 km west of the site. High river activity and sediment deposition define this region, which is composed of numerous channel scars, oxbow lakes, and back swamps. Exploring Prehistoric Communities

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

195

The most prominent feature at the Battle mound site is a multilevel platform mound that looms over the current landscape. Directly east of the mound, two very low rises (0.5–1 m in height) are subtly discernible. The large mound is composed of at least three platform levels and a large slope on the eastern side of the mound, a construction that is considered unique among Caddo mound sites. Although large and significant in several archaeological aspects, this premier Middle and Late Caddo (ca. A.D. 1200–1680) civic ceremonial mound center has seen few systematic archaeological investigations, especially those related to understanding the cultural landscape of the Caddo community. The earliest recorded work is by C. B. Moore during his five months on the Red River in 1911–1912 (Moore 1912:484; Weinstein et al. 2003:59). His crew dug numerous “trial holes” in the mound with few results and then moved to low rises east of the mound as well as to a cemetery area to the north of the mound in an area frequently referred to as the Handy Place. In the summer of 1948, a two-month excavation at the Battle mound site, headed by Dr. Alex D. Krieger and supervised by Lynn E. Howard, excavated mostly on the mound (Krieger 1949; Howard 1948; McKinnon 2010b). Beginning in 1979 and continuing intermittently through the mid-1990s, the Arkansas Archeological Survey (aas) station in Magnolia began a program of artifact surface collections in 10 delineated areas at and around the Battle mound site. Archaeogeophysical Investigations at the Battle Mound

Copyright © 2012. Nebraska Paperback. All rights reserved.

Recent geophysical surveys at the Battle mound site are another example of an examination of Caddo landscapes using archaeogeophysics (McKinnon 2008, 2009, 2010a). To date, the survey area represents 8.48 hectares (20.95 acres) of archaeogeophysical data around the mound (fig. 7-10). Anomalies identified within the 8.48 hectare area include those of (a) recent cultural origin (i.e., scattered metallic debris, incising in the landscape as a result of agricultural processes, and the remains of a long demolished tenant structure); (b) anomalies interpreted to be of Caddo cultural origin (i.e., numerous rectangular and circular structures and associated pits and hearths, a possible large community cemetery, evidence of a possible compound fence surrounding a Caddo farmstead, the remains of a borrow pit, and a possible linear causeway); and (c) anomalies originating from natural processes (i.e., evidence of light196

Walker and McKinnon

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

7-10. Total coverage of the magnetic gradiometry survey at the Battle site (3LA1).

ning strikes, a sequence of Red River meander scars, and evidence of a modern flood channel). These archaeogeophysical investigations have produced an informative data set of numerous anomalies interpreted as being the product of both archaeological and geological sources. The results demonstrate the efficacy of archaeogeophysical survey methods at the Battle mound site and their contribution toward a holistic understanding of the spatial orientation of the cultural and natural anomalies over time. Several of these geophysical anomalies contain patterns similar to those defined archaeologically at other Caddo sites. The similarity of geophysical patterns to archaeological features identified at other Caddo sites form the basis of the Battle mound archaeogeophysical interpretations. Exploring Prehistoric Communities

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

197

7-11. Clusters of possible structures, a community cemetery, a potential compound fence, a borrow pit, and a possible linear causeway identified in the Battle site archaeogeophysical data.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Community Organization When the survey area (8.48 hectares or 20.95 acres) at the Battle mound site is viewed across the landscape, numerous anomalies interpreted to be of prehistoric cultural origin provide insight into community organization. To date, the corpus of possible Caddo archaeogeophysical anomalies represent a total of 19 circular structures, 12 rectangular structures, a dense farmstead area, four compound fences, a community cemetery, a linear causeway, and a buried borrow pit (McKinnon 2008, 2009, 2010a)(fig. 7-11). About 100 m west of the large platform mound are three rectangular and two circular structures. The dimensions of the rectangular structures range between 5 m2 and 12 m2. The circular structures are approximately 198

Walker and McKinnon

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

7-12. House 6 at the Belcher site (after Webb 1959:41) compared with a small circular structure with a possible extended entranceway identified in the Battle site data. Note the change of scale between compared images.

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

7-13. A complex grouping of high magnetic values forms a possible farmstead with a compound fence. Several other high magnetic values form various geometric shapes that could be possible perimeters of rectangular and circular structures. The rectangular perimeters all are arranged in a similar orientation.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Copyright © 2012. Nebraska Paperback. All rights reserved.

10 m in diameter. Within one of the circular structures, situated close to the mound, are geophysical anomalies that resemble a circular pattern of “ash pits” similar to those identified at the Belcher site (16CD13) in northwestern Louisiana (see Webb 1959:43–44 and figs. 32 and 39; see also Kelley, chap. 14, this volume) (fig. 7-12). This one circular structure at Battle also contains a linear anomaly that is interpreted as an extended entranceway facing the west side of the mound. The magnetic patterns in this area likely represent the remains of burned structural features, such as wall debris or wall trenches. Fairly centralized within most of the patterns are semicircular patterns of concentrated high magnetic values. These may represent the remains of hearths or subsurface pits. North of the large platform mound is a concentrated area of high and complex patterned magnetic values (fig. 7-13). This area of numerous anomalies borders the old Handy Place property and is noted by Moore (1912:484; Weinstein et al. 2003:59) as a location of a Caddo cemetery area. The old Handy Place property has a history of amateur archaeological excavations resulting in numerous artifact collections, many from burial contexts (Moore 1912; Kitchens and Kitchens 1968; Perttula et al. 2009). Within the complex patterning, at least five patterns of anomalies are interpreted as possible rectangular structures, each approximately 4 m on a side. Also identified is a single possible circular structure that is approximately 10 m in diameter. This interpretation of a cluster of structures in close proximity to the Handy Place excavations suggests that this cluster of geophysical anomalies is evidence of a possible Caddo farmstead compound. Additionally, a wide and long linear band of high magnetic values runs south to north in this farmstead area. The magnetic linear band continues to the north beyond the survey area and into the Handy Place property. The visible extent of the linear anomaly is approximately 15 m long and 2 m wide. This large linear anomaly could be the remains of a compound fence that defined a boundary between two farmstead habitation areas, similar to those farmstead divisions recorded in the Terán map (see fig. 7-9; see also fig. 15-1 in Sabo, this volume). Compound fences have been identified archaeologically at two Caddo sites. At the Hardman site (3CL418), a large distinctive arc of post holes about 20 m in length is interpreted as the remains of a compound fence (Williams 1993:42–46). At the Hatchel site (41BW3), a linear group of post holes that run for 15 m in one habitation area is interpreted as a Exploring Prehistoric Communities

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

201

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

7-14. Numerous circular structures located on two rises east of the large mound.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Copyright © 2012. Nebraska Paperback. All rights reserved.

possible compound fence (Perttula 2005:192). The high magnetic values that comprise the possible compound fence at the Battle site could be the result of a burning event, aeolian processes that deposited magnetically enriched topsoil along a fence, or both. East of the large platform mound are two low topographic rises no more than 1 m in height. According to Moore (1912:484; Weinstein et al. 2003:59), as many as four low rises were visible in this area at the Battle mound site during his visit in the early twentieth century; the two rises visible today are likely part of those four low rises. The archaeogeophysical data indicates that each rise contains numerous circular structures and associated features (see fig. 7-11). The first rise, approximately 100 m from the mound, contains at least 10 circular structures, two midden areas, and a possible “causeway” extending from the eastern side of the mound. The easternmost rise, approximately 200 m from the mound, contains at least two large circular structures and two small rectangular structures with numerous intrastructure anomalies (fig. 7-14). Within two of the large circular anomalies on the easternmost low rise, isolated high magnetic values are arranged in a linear pattern that aligns with the center of the circles, similar to the patterning of a structure excavated at the McLelland site (16BO236) on the Red River in northwestern Louisiana (Kelley 1997:30, see also Kelley, chap. 14, this volume) (fig. 7-15). These are likely the remains of large interior post holes, possible burials, or even central hearths. Both low rises contain numerous anomalies of various sizes that are easily identified in the magnetic gradiometry data as circles or concentric circles of generally low magnetism (see fig. 7-14). These low magnetic concentric circles may be the result of the piling of sandy soil to create a sand berm around the outside of the structure. Berms of sandy soil have been identified at other Caddo sites (cf. Davis and Gipson 1960; Krieger 1949; Perino 1994; Walters and Haskins 1998, 2000; Parsons 2010). For example, at the Bryan Hardy site (41SM55) in Smith County, Texas, excavations of a structure revealed “a ledge of soil around part of the house, particularly by the entrance” (Walters and Haskins 2000:4). An alternative interpretation of the concentric circle pattern is that it is a structure composed of concentric post hole patterning. For example, at the Werner site (16BO8) in northwestern Louisiana (see Girard, chap. 9, this volume), two concentric circles of post holes are interpreted as the remains of a burned ceremonial lodge, with the outer circle representing Exploring Prehistoric Communities

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

203

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

7-15. Structure 1 at the McLelland site (after Kelley 1997:30) compared with two large circular structures identified in the Battle mound site data. Note the change of scale between compared images.

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

7-16. Structure at the Werner site (after Webb 1983:220) compared with the several concentric circular patterns identified in the Battle mound site data. Note the change of scale between compared images.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Copyright © 2012. Nebraska Paperback. All rights reserved.

the exterior wall and the inner circle representing a series of roof supports (Webb 1983:217–221) (fig. 7-16). This concentric pattern of outer and inner post holes may have defined an interior partitioned space. If such a space existed, especially in a large ceremonial lodge, it could have been a partitioned space that would have been ritually cleaned. Such an activity would have inadvertently removed magnetically enriched topsoil and left a circular band of low magnetic soil matrix similar to the concentric pattern of low magnetism that defines the several structures on the two low rises at Battle. Directly east of the large mound, in an area roughly 1,600 m2 in size, a large concentration of isolated magnetic anomalies can be discerned (fig. 7-17). Some of the anomalies are as large as approximately 4 m2. Additionally, many of the larger anomalies occur in two possible linear groupings. Given the number of anomalies and their possible linear patterning in this part of the Battle site, these isolated high magnetic anomalies are likely cultural in origin and may represent a cluster of different types of subsurface pits, such as storage pits, refuse pits, or burial pits. The many anomalies and their spatial proximity to the two large and presumably ceremonial circular structures on the easternmost low rise lend credence to the likelihood that the cluster of pits represents a possible Caddo community cemetery. Community cemeteries have been noted to contain numerous individuals. Cemetery sizes likely depend on settlement densities of the surrounding populations (Perttula 1992:83; Story 1990:338–339; Brown 1984b:54). A total of 91 isolated magnetic anomalies are identified in this area, and that number provides initial insights into the spatial arrangement of what is considered to be a large community cemetery. The two possible linear groupings may represent several “grave groups” of related Caddo individuals, similar to the multiple subgroups of burials at the Cedar Grove site (3LA97), located not far from the Battle Mound site (Trubowitz 1984:108). Conclusions Archaeogeophysical data from the Caddo communities at the George C. Davis, Hill Farm, and Battle Mound sites demonstrate their importance in the continuing analysis of prehistoric Caddo community organization. The analysis of the internal character of Caddo sites at a landscape scale allows for the holistic integration of archaeological and geophysical investigations and interpretations. 206

Walker and McKinnon

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

7-17. A large cluster of pits possibly representing a large community cemetery about 300 m east of the large mound at the Battle site. There are two linear groups of larger pits.

Data sets such as the magnetometer survey from the Davis site contribute details of Caddo community organization that are unmatched throughout the rest of the Caddo archaeological area. The geophysical data provide a sufficiently complete picture of the distribution and character of archaeological features to understand the larger community patterns that existed during Caddo times. When interfaced with archaeological information, geoarchaeological interpretations, and historical documents, the Hill Farm site geophysical data have led to a more nuanced understanding of one of the more complex Caddo landscapes, that of an Upper Nasoni Caddo village on the Red River. The small project conducted at the Hill Farm site should only be a prelude to a much larger archaeogeophysical effort that could examine the entire area covered by the 1691–1692 Terán de los Rios map of the Upper Nasoni Caddo village. Finally, the geophysical data from the Battle mound site demonstrate the capability of such work to obtain a primary archaeological data set. It also demonstrates that it is possible to develop an understanding of community organization and site structure on one of the largest Caddo mound sites of the Red River valley. Traditional archaeological field methods and techniques are constrained in the scope of their investigative abilities by the sheer size of the larger sites. Consequently, to date, much of what we know about these larger sites in the Caddo area rests on a few excavation projects that were only able to examine small portions of the archaeological Exploring Prehistoric Communities

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

207

.

Copyright © 2012. Nebraska Paperback. All rights reserved.

deposits preserved at them, opportunistic surface collections, and incomplete information from artifact collectors. Because of its spatial scope, archaeogeophysics is well suited as a well-developed investigative approach to achieving a better understanding of Caddo mound centers and habitation sites, whatever their size and internal complexity. Archaeogeophysical research can aid in the planning of future archaeological projects as well as provide primary archaeological data suitable for interpretation at a landscape scale.

208

Walker and McKinnon

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

8. The Evolution of a Caddo Community in Northeast Texas timo thy k . perttula and robert rogers

Copyright © 2012. Nebraska Paperback. All rights reserved.

Introduction Only rarely do prehistoric archaeologists in North America completely excavate and study an entire Native American community or village, particularly in the Caddoan area of northeast Texas, southeast Oklahoma, southwest Arkansas, and northwest Louisiana. It is simply too costly in this day and age, even on cultural resources management projects, to complete such large excavations, and consequently our views and understanding of past prehistoric settlements are based on sampling considerations, limited and partial exposures of settlement features, and comparisons with historically recorded examples (i.e., maps and descriptions) of the character of Native American communities and their related parts. By an archaeological community, we mean “the conjunction of ‘people, place, and premise.’” Specifically, this is a group of households that share a regular social interaction in a specific “set of places within a particular span of time” (Yaeger and Canuto 2000:5). To be able to expose a prehistoric Native American village in its entirety provides a unique view of past community and social arrangements that existed among Native American societies before contact with Europeans, if the different components (either temporal or spatial in nature) can be recognized and defined in the archaeological record. The Oak Hill Village site (41RK214) in the Pineywoods of northeast Texas is such a site. It is the only prehistoric or early historic Caddo settlement in the entire Caddo area of northeast Texas (fig. 8-1) that has been completely excavated in the ca. 100 years of avocational and professional archaeological investigations (other than the School Land I and II sites in Oklahoma [see Duffield 1969] that have been conducted in 209

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-1. Oak Hill Village (41RK214) in the Caddo area of northeast Texas and other important contemporaneous Caddo sites and phases. Key to numbered sites: 1. Harling (41FN1); 2. Sanders (41LR2); 3. Fasken (41RR14); 4. Roitsch/Sam Kaufman (41RR16); 5. Holdeman (41RR11); 6. Hatchel (41BW3); 7. Hurricane Hill (41HP106); 8. 41RR181 and Little Mustang Creek sites; 9. 41TT670; 10. 41CS150; 11. Coker (41CS1); 12. 41TT372; 13. 41FK70; 14. Benson’s Crossing (41TT110); 15. 41TT650; 16. Harold Williams (41CP10); 17. 41UR21; 18. Big Oaks; 19. Griffin Mound (41UR142); 20. Rookery Ridge (41UR133); 21. 41UR9; 22. McKenzie; 23. Spoonbill (41WD109); 24. 41RA65; 25. T. M. Moody; 26. 41WD518; 27. Yarbrough; 28. Charlie Crews; 29. Jamestown (41SM54); 30. Carlisle (41WD46); 31. Langford; 32. Bryan Hardy; 33. 41HS74; 34. Oak Brown Place; 35. Oak Hill Village (41RK214); 36. 41PN14; 37. Musgano; 38. Pace McDonald; 39. 41CE42; 40. 41CE289; 41. George C. Davis (41CE19); 42. Forest Mound (41CE290); 43. 41NA20; 44. Washington Square; 45. Tyson;

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

northeast Texas, southwest Arkansas, eastern Oklahoma, or northwest Louisiana. These extensive hand- and machine-aided investigations at the site (Rogers and Perttula 2004) provide an unequaled opportunity to study the evolution of a single prehistoric Caddo community that was occupied (although probably not continuously) between ca. A.D. 1150 and 1450—a time of significant cultural changes among Caddo peoples in subsistence, settlement, and mound-building (Rogers and Perttula 2004:35–42)—and relate that community’s development and eventual abandonment to both local and regional trends in Caddo native history. The Oak Hill Village site is a large, nucleated village in the Sabine River basin that had been occupied by prehistoric Caddo peoples between about ca. A.D. 1150 and 1450. It lies within the modern and prenineteenth century limits of the Pineywoods, a biotic habitat composed of both deciduous and coniferous forests on upland landforms, with alluvial landforms dominated by hardwoods, particularly forest species that tolerate frequent flooding and swampy conditions. The modern climate is humid subtropical, with more than 240 frost-free days and precipitation averaging about 117 cm per year, but with considerable variation from year to year, ranging from less than 70 cm to more than 170 cm per year. Most of the precipitation falls in the spring and fall months, with droughty summers common. Paleoclimatic data from dendrochronological records (Stahle and Cleaveland 1995; see also Tree-Ring Data bank, igbp Pages/World Data Center for Paleoclimatology Program, noaa/ngdc, Boulder, Colorado) indicate that very dry years were quite rare between A.D. 1000 and 1400, but after that time dry years were more common. The driest and coolest years in the 500-year period between A.D. 1000 and 1500 (bracketing the overall occupational period at Oak Hill Village) occurred in the major drought years of 1444 and 1447 and 1455 and 1460; the period from ca. 1430 and 1470 was apparently the coolest and driest in the last 1,000 years in the general region. The village was fully uncovered in the mid-1990s prior to proposed strip mining of lignite coal (Rogers and Perttula 2004:1–3). The village

(Fig. 8-1 caption continued) 46. 41SA123; 47. 41SA89; 48. Knight’s Bluff (41CS14); 49. 41FK7; 50. Hudnall-Pirtle (41RK4); 51. Gray’s Pasture (41HS524); and 52. Redwine (41SM193). Bowman, Crenshaw, Battle, Haley, Belcher, Werner, and Vanceville are important prehistoric Caddo mound centers that were occupied at the time of the Oak Hill Village site.

Evolution of a Caddo Community

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

211

evolved over a 300-year period, in three distinct village eras (termed the earliest, middle, and latest village) and contained more than 43 structures, many pit features (n = 134), an impressive assemblage of decorated ceramic sherds (n = 3,955) and lithic artifacts, and preserved plant remains (mostly maize, along with charred hickory nuts and a few starchy seeds; see Dering 2004). The structures and features were arranged in a roughly circular fashion around the edges of a ridge top (fig. 8-2) overlooking Mill Creek—a tributary to the Sabine River—leaving an open area or small plaza in the center. There was no evidence in the complete stripping of a palisade or fortification around the village. The entrances or openings to most of the structures were oriented so that they opened into and faced the plaza. We propose to address three issues with this detailed examination of a single prehistoric Caddo community: (1) age and intrasite chronological relationships of the Oak Hill Village structures; (2) the occupational history of the site as reconstructed from structure rebuilding and superposition evidence, radiocarbon dates from structures and associated features (including a mound at the northern end of the village), and ceramic stylistic analyses of decorated sherds from different contexts across the village; and (3) structure/feature relationships and community patterns through time. From these basic building blocks of the site’s archaeological record, we turn to a broader view of the evolution and abandonment of the Oak Hill Village by the Caddo in regard to regional changes in settlement, subsistence, and sociopolitical organization within the dynamic prehistoric Caddo societies of northeastern Texas. Architecture in the Village Copyright © 2012. Nebraska Paperback. All rights reserved.

We have recognized eight groups of structures (groups A–H) in the Oak Hill Village, along with three other isolated structures (2, 5, and 10) (fig. 8-3). The structures are marked by clusters of patterned post hole stains that were exposed in the clay B-horizon following hand and machineaided excavations, along with a variety of pit features (including many smudge pits) and associated artifacts. Many of the structures appear to have been rebuilt, and they occur in clusters across the site, and in some cases the structure groups consist of spatially distinct clusters of buildings (groups C, D, E, F, and H), while others (groups A, B, and G) are composed of many overlapping dwellings (between 3 and 10) from a number of successive building episodes (at least five episodes are apparent in structure group G). 212

Perttula and Rogers

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-2. Plan of structures and features at the Oak Hill Village.

Remains of over 43 wooden structures were uncovered at the Oak Hill Village (fig. 8-2). Four kinds of wood structures have been preserved in the archaeological and architectural record. They include a number of rectangular and circular household structures (fig. 8-4a–b), a few circular structures with extended entranceways (fig. 8-4c) that may be the households of the political elite, and several small (less than 3 m in Evolution of a Caddo Community

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

213

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-3. Structure groups A–H and structures 2, 5, and 10 at the Oak Hill Village.

diameter) structures (fig. 8-4d) that likely served as aboveground granaries, which we know from ethnographic accounts that the Caddo built, to hold the maize crop and seed stock. The four rectangular structures (37, 38, 39, and 43) are in the northern part of the Oak Hill Village site, in structure group A and G (see figs. 8-2 and 8-3). Three are oriented more or less north/south, while structure 39 is oriented northeast/southwest. They range from 6.5 × 9 m in 214

Perttula and Rogers

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-4. Plans of the different kinds of structures at Oak Hill Village: a. large rectangular household structure; b. circular household structure; c. circular structure with extended entranceway; and d. small structure, probably a granary.

size (structure 43) to 8 × 12 m in size (structure 37). The mean enclosed area of the four rectangular structures is 84.25 m2. The rectangular buildings lacked post holes at the corners, suggesting they had rounded rather than square corners, and only one (structure 38) had a single center post/roof support; we suspect, however, on architectural grounds, that they all had a center post to support corner Evolution of a Caddo Community

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

215

Copyright © 2012. Nebraska Paperback. All rights reserved.

rafters and the roof. Entrances are not obvious on them, although gaps in the north and west walls of structure 39 and the east wall of structure 38 may be the entranceways to these structures. The mean spacing of post holes ranged from 44 to 65 cm, and two of the rectangular structures have evidence of wall repair. The 32 circular structures without extended entranceways range from 5.25 to 11 m in diameter; the circular structures were intended for domestic use based on their archaeological context and general architectural, artifactual, and feature characteristics, and they are found widely across the site (see fig. 8-2). They have a single center post and regularly spaced wall posts, with gaps in the posts marking the structure entrance (table 8-1). There does appear to be a temporal trend in the size of the circular structures, as they became smaller over time. In the middle village (ca. A.D. 1250–1350/1375), the mean enclosed area of the circular structures is 54.64 m2. The mean enclosed area of the circular structures in the latest village (ca. A.D. 1350/1375–1450) is only 39.75 m2. Structure entrances for the circular structures tend to open toward the plaza. Two unique circular structures (structures 2 and 18, see figs. 8-2, 8-3) at the site had 3–4 m long extended entranceways that opened either to the north or northwest toward the plaza. We suspect these are important structures, probably the residence of an elite member of the village: (a) they are set apart (on a higher elevation) or isolated from the remainder of the structures; (b) construction details are different than the typical domestic residence; (c) the extended entranceway would have restricted ready access to the structure interior; and (d) Caddo structures in northeast Texas with extended entranceways tend to be found almost exclusively on earthen mound platforms (see Jackson 2004; Story 2000; Thurmond 1990) or are found buried beneath small earthen mounds (see Jelks and Tunnell 1959; Story 2000; Walters et al. 1998; Walters and Haskins 2000). Structure 2 was 6.5 m in diameter, with 71 wall support posts closely spaced every 10 to 25 cm (see table 8-1) and a single center post. The entranceway was first delineated by two parallel trenches that each then had 12–13 small and closely packed posts placed in the entranceway trench; two larger posts marked the outer edge of the entranceway (Rogers 2004:fig. 20). Inside the structure was a central hearth overlying the pulled center post (a typical Caddo construction technique) and two pit features that contained large quantities of corn, including many corn 216

Perttula and Rogers

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 8-1. Details of the circular structures at the Oak Hill Village site Structure group

Structure no.

Diameter (in m)

Posthole spacing (mean, in cm)

Center post

Entrance direction

Copyright © 2012. Nebraska Paperback. All rights reserved.

Middle Village A A A A A A A A Near A B B B C C D D D Near F G G G G G G G

3 14 28 29 30 31 40 41 5 4 25 24 6 8 7 12 13 2 1 9 32 33 34 35 36

6.8 7.5 9.5 7.8 9.0 9.5 6.7 8.2 6.5 7.5 7.8 7.7 5.25 6.4 8.1 8.8 8.5 6.5 8.0 9.0 10.0 11.0 9.0 10.0 10.3

40.5 54.9 73.1 67.1 64.6 56.2 50.0 83 44–61 54.2 53.9 51.6 41 38.2 57 63 79 10–25 70 52 80 61 60 60 88

Yes Yes Yes Yes ? Yes Yes Yes Yes* — Yes — Yes Yes Yes — — Yes — Yes Yes Yes Yes Yes ?

E NE NE ? NE ? SE? ? NE and SW ? ? ? S ? E, SE, and N? E and SE NW NW SE SE E NE SE SE SE

Late Village E E F H H H H H

17 26 10 18 19 20 21 42

6.5 7.1 6.0 8.7 6.7 8.0 6.8 7.2

50 67 38–77 59 49 61 53 51

Yes Yes —** Yes Yes Yes Yes Yes

SE SE W N ? ? N ?

*Also has four interior support posts. **Two interior compartments.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

cobs (Elson et al. 2004:fig. 132). Structure 2 is situated near structure group F and its several possible elevated granaries (see fig. 8-3). Structure 18 in structure group H was 8.7 m in diameter with a 4 m long entranceway and a single center post. The four possible aboveground granaries are in structure group F in the eastern part of the site, just north of structure 42 (see figs. 8-2 and 8-3). They range from 1.6 to 3.0 m in diameter and have no obvious entrance; instead, entrance to the structures was probably by a ladder to an elevated platform. The posts are spaced, on average, 53–95 cm apart. Structure 11 had a second line of post holes on the interior of the structure, probably serving as the supports for an interior compartment that would have had good air circulation and ready access; the other three possible granaries did not have this architectural feature. Radiocarbon Dates from the Village

Copyright © 2012. Nebraska Paperback. All rights reserved.

Calibrated radiocarbon dates have been obtained from seven structure groups and important features at the Oak Hill Village. There are 36 dates from the site, which is one of the largest samples of dates from any single prehistoric Caddo site in the Caddo area. In general, the calibrated radiocarbon dates provide only limited temporal resolution to the age of individual structures or the age of structure groups, because the calibrated dates provide at best an interval of 130–200 years (see table 8-2). They do at least provide broad points in time that can be corroborated by, and supplemented with, architectural data on structure superpositioning and rebuilding as well as estimated ages obtained from the ceramic stylistic evidence. There are six calibrated radiocarbon dates from Early Village contexts. At 2-sigma, these range from A.D. 1032 to 1301, and at 1-sigma, the age range is A.D. 1044 to 1285. Twenty-one dates are from Middle Village contexts, and most of these are from structure 2 (n = 10) and structure 5 (n = 6). With the exception of one anomalous older date from structure 12 (Beta-96913)1 and the wide-ranging dates on maize from the same feature (86 in structure 2), the 2-sigma age range of all the Middle Village dates is A.D. 1162–1442; most fall between A.D. 1211 and 1407. The mean age of the four charcoal dates from structure 2 is A.D. 1238–1404 (2-sigma),2 while the mean age of the six charcoal samples is A.D. 1237–1400 (2-sigma) for structure 5.3 The remaining nine radiocarbon dates are from Late Village structures and features. At 2-sigma, they have a broad range: A.D. 1288–1494. 218

Perttula and Rogers

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 8-2. Radiocarbon dates from the Oak Hill Village site Structure group

Structure/ feature

Conventional age (B.P.)

1-sigma cal. age (A.D.)

2-sigma cal. age (A.D.)

900 ± 40 830 ± 40 880 ± 60 800 ± 40 850 ± 40 890 ± 40

1051-1210 1206-1263 1044-1256 1207-1285 1186-1259 1058-1214

1027-1221 1153-1284 1028-1263 1150-1301 1050-1279 1032-1223

B-96909 B-96911 B-96913 B-110065 B-110061 B-96915 B-96920 B-96921 B-96910 B-73939 B-138870 B-138871 B-73940 B-138868 B-138869 B-96908 B-96912 B-96914 B-96916 B-96918 B-96919

700 ± 50 720 ± 50 850 ± 60 630 ± 40 640 ± 40 630 ± 50 740 ± 60 610 ± 50 720 ± 60 810 ± 100 830 ± 70 760 ± 90 610 ± 80 650 ± 40 650 ± 40 660 ± 50 670 ± 50 730 ± 60 660 ± 50 750 ± 50 730 ± 50

1264-1388 1259-1386 1071-1275 1301-1392 1299-1390 1300-1393 1219-1301 1304-1402 1257-1323 1153-1292

1290-1389 1286-1389 1222-1385 1290-1389 1222-1294 1223-1383

1223-1399 1217-1392 1035-1148 1292-1406 1289-1405 1288-1410 1162-1393 1291-1416 1211-1399 1022-1389 1030-1290 1040-1400 1264-1442 1280-1405 1280-1405 1280-1407 1264-1406 1194-1398 1280-1407 1193-1389 1213-1390

Late Village (ca. A.D. 1350/1375-1450) A S1 B-73941 E S26 B-110064 F S11 B-120290 G F261 B-110068 H S19 B-107399 H S19 B-110067 H S21 B-110066 Area A F65 B-73936 Area A F81 B-73938

550 ± 60 570 ± 40 580 ± 70 570 ± 40 500 ± 50 570 ± 40 570 ± 40 560 ± 60 570 ± 140

1326-1447 1326-1421 1305-1419 1326-1421 1398-1449 1326-1421 1326-1421 1325-1432 1281-1485

1294-1447 1302-1433 1288-1441 1302-1433 1319-1494 1302-1433 1302-1433 1296-1441 1191-1652

Lab no.

Early Village (ca. A.D. 1150-1250) A B G G Area A Area A

F219 S4/F146 S39 S39 F78 F166

B-110063 B-119889 B-81681 B-71486 B-119885 B-119886

Copyright © 2012. Nebraska Paperback. All rights reserved.

Middle Village (ca. A.D. 1250-1350/1375) D1 D D D G — — — — — — — — — — — — — — — —

S12 S12 S12 F105 S35 S22 S2 S2 S2 S2 (maize) S2 (maize) S2 (maize) S2 (maize) S2 (maize) S2 (maize) S53 S5 S5 S5 S5 S5

1302-1406

All radiocarbon samples were dated by Beta Analytic, Inc. Note: calibrations follow CALIB 3.03c (Stuiver and Reimer 1993) and INTCAL98 (Stuiver et al. 1998)

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Occupational History of the Oak Hill Village Site as Reconstructed from Architectural, Ceramic, and Radiocarbon Data

Copyright © 2012. Nebraska Paperback. All rights reserved.

In attempting to reconstruct the occupational history of the Oak Hill Village site, we have had to rely heavily on architectural information about the different structures, as well as the ceramic stylistic differences in the more than 3,950 decorated sherds from structure groups (decorated sherds by structure group range from 84–408, with 858 decorated sherds from structure 5 and 84 from structure 10) and area to another, and radiocarbon dates, to further tease out and bolster the chronological relationships between different groups of structures. Because at least half of the structures at the site overlapped (figs. 8-5 and 8-6) and were apparently rebuilt, with the newer structure placed adjacent to the first, the construction sequence of two or more houses could be determined by intersecting post holes, if such were identified. Similarly, several of the houses were rebuilt, with the newer structure placed adjacent to the first; it also closely paralleled it in the number, spacing, and size of the wall posts (i.e., structures 34 and 35 in structure group G, fig. 8-7), suggesting the reuse of existing structural materials. Often the two houses differed in size, with the later structure being slightly smaller. This suggests that the second and rebuilt house contained the recycled wall posts of the first structure, which may have rotted at the ground level and thus were shorter. As a consequence of using the shorter posts in building a circular structure, the second structure was smaller in diameter. In cases where there were few or no intersecting post holes (as exemplified in structure group G), reconstructing the architectural and construction sequences of many overlapping buildings (fig. 8-8) relied on the position of the houses relative to one another, as overlapping or very closely positioned houses could not have coexisted (see discussion in Rogers 2004:49–50). Additionally, the post hole patterns of earlier structures were often partially obliterated by later construction. Finally, apparent associations or relationships among houses could be seen in certain post hole characteristics, including wall post hole spacing and the ratio of the distances between wall post holes and center posts (i.e., structures 28 and 29 in group A, see fig. 8-6, and structures 32, 34, and 36 in group G, see figs. 8-7 and 8-8). To bolster the reconstructed sequence of house construction, as well as the extent of the village at different points in time, we also relied on 220

Perttula and Rogers

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-5. Three overlapping circular structures in the northeastern part of the site (structure group D).

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-6. Circular structures in the western part of the site (structure group A).

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-7. Overlapping structures in the northwestern part of the Oak Hill Village site (structure group G).

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-8. Unraveling the sequence of structures in the northwestern part of the Oak Hill site, from A (oldest) to E (youngest).

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

the distribution of decorated sherds (n = 3,955) and distinctive decorative elements on sherds (Perttula 2004a:table 69) in 16 gradall trenches (sherds recovered from the trench as a whole) excavated from south to north across the site at 8–10 m intervals, and distributional information on the density of brushed sherds from the southern end of the site. The frequency of brushed pottery is strongly time-dependent, as brushed pottery became more and more frequent in northeast Texas Caddo sites, beginning after around A.D. 1150 (Perttula 2004a:253). In the earliest contexts at the Oak Hill Village, brushed sherds comprise 16 percent of all the decorated sherds, to 27 percent in the latest parts of the village (i.e., structure group D). We defined three distinct ceramic assemblage groups using the Brainerd-Robinson (br) coefficient of similarity (Perttula 2004a:tables 66–67). The br similarity coefficients among the principal excavated areas/sherd assemblages at Oak Hill (Area A, structure 5, structure 10, structure group B, structure group C, structure group D, structure group G, and the Plaza area) range between 150 and 176 (out of a possible score of 200, see Robinson 1951). The br coefficient indicates that there is considerable intergroup stylistic coherence, with only a 13 percent difference (26 br points) between the highest and lowest similarity coefficient. This is not surprising given the stylistic relationships from one area and structure group to another and the character of this extensive and planned community. There are four groupings in the data where the br coefficient between two or more assemblages is consistently higher than among the other assemblages: Area A–structure group D (br coefficient of 171) Structure group G–Plaza Area (br coefficient of 172) Structure 5–structure 10–structure group B–structure group C (br coefficient of 169–175) Structure group D–structure group B (br coefficient of 176) As the br coefficient of similarity has been used to chronologically order groups of assemblages, we interpret the groupings as representing chronologically distinct assemblages, with structure groups B and D linking two of the groups. The relatively low percentage of brushed sherds from structure group G and the Plaza (13.6–16.8 percent of the Evolution of a Caddo Community

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

225

Copyright © 2012. Nebraska Paperback. All rights reserved.

decorated sherds) suggests that these ceramic assemblages are the oldest at the Oak Hill Village. Structure groups B and D have the highest percentage of brushed sherds (22.6–26.7 percent). In structure 5, structure 10, and structure group C, brushed sherds comprise 16.1–19.2 percent of the decorated sherds. The high similarity coefficient values between structure group B and structure group D suggest that if structure group D is the youngest ceramic assemblage of the excavated areas, then structure group B may be the most recent of the structure 5–structure 10–group B–group C cluster. The highest percentages of brushed pottery (27–35 percent) are found in trench collections of decorated sherds in the area of structure groups E, F, and H. These structure areas, as well as parts of Area A (see Perttula 2004a:253 and fig. 106), mark the latest use of Caddo settlement at the Oak Hill Village. From the age ranges of the calibrated radiocarbon assays (see table 8-2), a ca. A.D. 1150–1450 range for the length of the site occupation seems reasonable. As discussed below, this occupation was not continuous across the site over that +300-year period, and each part of the site was not apparently used for the same length of time or at the same level of intensity. The earliest occupation at the Oak Hill Village site had rectangular structures and predates ca. A.D. 1250, based on the distribution of certain early Caddo styles of decorated ceramic sherds (i.e., Holly Fine Engraved) and vessels (Perttula 2004a:257). This occupation is concentrated in two areas: (1) at the southern end of the site in Area A and vicinity, and (2) about 90 m to the north at the northern tip of the ridge (fig. 8-9). This latter area includes structure groups A, C, and G, the feature 245 pit, and burial 3. The habitation area at maximum covers ca. 60 × 60 m, a little bit less than an acre in size. Only structure groups A and G are known to have had structures built and occupied during the earliest “village,” and there was a scatter of associated ceramics (including Holly Fine Engraved and red-slipped sherds) with these structures as well as in an area to the east as far as structure group C, across what was later to become (but was not one then) the plaza for the larger village. The single burial associated with the earliest village is 10 m away from the main occupation area. The Area A archaeological deposit at the southern end of the site may not be directly associated with the small village at the north end of the 226

Perttula and Rogers

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-9. Inferred distribution of the earliest village at the Oak Hill Village site. The dashed line here (and on figs. 8-10 and 8-11) indicates the presumed extent of the settlement based on the structure distribution.

Oak Hill Village site, given the distance between the two areas. There is no evidence that the midden contained structures at this time. It may be the case that Area A is more closely related to 41RK215, a short distance to the west (Rogers et al. 1994:fig. 17). This particular site also had an extensive midden deposit (ca. 40 × 20 m in size) with a prehistoric Caddo Evolution of a Caddo Community

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

227

Copyright © 2012. Nebraska Paperback. All rights reserved.

ceramic assemblage that appears to be relatively early given the low percentage of brushed sherds at the site. The middle-era village (estimated to date from the mid-thirteenth century to the mid-fourteenth century) is the lengthiest and most intensive occupation at the Oak Hill Village site. The similarities in ceramic decorative methods and major decorative elements (brushed body sherds, appliqued body sherds, horizontal incised rims, and Weches Fingernail Impressed) and the distribution of these ceramics in excavation blocks, features, and trenches suggest that the middle-era village was concentrated on the northern tip of the ridge and covered an area ca. 100 × 60 m in size (fig. 8-10). During this period, circular structures were built in several different structure groups, including structure groups A, B, C, and D, along with structures 2 and 5. Insufficient ceramic evidence (i.e., number of decorated sherds) is available from structure group F to determine if any of the structures within that group were part of the middle-era village; radiocarbon dates on maize from structure 2 are somewhat ambiguous, suggesting it may have been built and used during either the middle or late villages (see table 8-2), although we favor its construction during the middle-era village. The structures were arranged around a small plaza that was ca. 65 × 15 m in size, and Burial 1 was interred at the plaza’s southwestern corner (see fig. 8-10). A small earthen mound (not used for burials) was built at the northern margins of the village, in structure group B. The mound was constructed over structures 4, 24, and 25 in structure group B (see figs. 8-2 and 8-3). It was composed of 2–3 mound fill zones that were ca. 40 cm thick in total. The combination of mound, plaza, and encircling circular residential structures indicates the mature development of the Oak Hill Village during the middle era. Certainly structures were built and rebuilt during this occupation (see figs. 8-5 to 8-8), mainly in proximity to the mound, and a small midden deposit accumulated by (and under) structure 5 in the southwestern part of the village. In structure group G, there may be four episodes of structure use and rebuilding during the middle era (see fig. 8-8, episodes B through E). The obvious concentration of structures near the earthen mound and the rebuilding of many structures here seem analogous to the “inner precincts” defined by Story (1998:26) for the Early to Middle Caddo period George C. Davis village and mound center (41CE19) (see fig. 8-1). There she interprets the inner precincts “as districts that are central functionally, if not also 228

Perttula and Rogers

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-10. Inferred distribution of structures in the middle-era village.

physically, to the community as a whole . . . the inner precinct areas are defined by structures and courtyards or plazas.” The much larger Area A deposits (see fig. 8-10) also continued to accumulate during the occupation of the middle-era village. Feature 166, a large trash pit, probably dates to this period based on its ceramic content, and the occurrence of a large pit (large pits probably used for Evolution of a Caddo Community

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

229

Copyright © 2012. Nebraska Paperback. All rights reserved.

storage are commonly found in proximity to Caddo structures) in the midden suggests that at least one structure was built on the earlier midden deposits during this era. At least 50 m separates Area A from the remainder of the middle-era village archaeological deposits. The final occupation at the Oak Hill Village site (beginning perhaps about the middle to third quarter of the fourteenth century) indicates a further and broad expansion in the size and complexity of the village during its final settlement. First, the ceramic evidence (i.e., increasing amounts of brushing, brushed-punctated sherds, Pease BrushedIncised sherds, rims with nodes and tabs, cross-hatched incised rims) suggests that there was significant building and use of circular structures at the northern end of the site, including a possible new group of structures in structure group E (fig. 8-11) and several possible rebuilding episodes during the latest village, possibly in structure groups D and G (see figs. 8-7 and 8-8). Second, new structures and structure groups were built along the midpoint of the ridge—including structure 18 and structure group H (and burial 2)—and structure 18 appears to have been part of the final episode in the planning and use of the village plaza. And third, the distribution of brushed sherds (with the highest frequencies of brushed sherds in any part of the site) and radiocarbon-dated features in and around Area A (see table 8-2) suggests that this area was fully incorporated into the latest village at the site, and there may have been feature and structure clusters here that were not fully recognized during the site excavations. While the plaza area is not well defined from the ceramic distributional evidence from the trenches or the structure group, relying on the available information and the range in radiocarbon dates from structure 5, structure 10, structure 19, and structure 21 (see table 8-2), the plaza may have extended from the southern end of structure groups C and G to the northern end of structure 10 (see fig. 8-11). If this is an accurate assessment, then the latest village plaza area was perhaps twice as long as the middle-era plaza. The entranceway for structure 18 clearly points into the plaza (as did the structure 2 entranceway). Structure/Feature Relationships and Community Patterns through Time The analysis of architecture, changes in ceramic styles and decorative elements, the placement of structures, the plaza, and the earthen mound, as well as the radiocarbon dates on key features are further considered 230

Perttula and Rogers

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-11. Inferred distribution of the latest village at the Oak Hill Village site. The latest structures in structure group G may also be part of this component.

to examine structure/feature assemblages and changes over time and space. There are six key elements in the relationships. First, the earliest structures built at the Oak Hill Village were rectangular. Second, circular structures with extended entranceways facing north and northwest toward the plaza, probably public or special Evolution of a Caddo Community

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

231

Copyright © 2012. Nebraska Paperback. All rights reserved.

purpose buildings, were apparently built during the middle and latest or final village. The first (structure 2) was probably contemporaneous with the construction of the earthen mound in structure group B at the northern end of the site. Third, burials are rare (n = 3) during each village episode. No cemetery is currently known, but may be nearby at a separate but unidentified site. The location of Burials 1 and 3 outside of domestic structures (and in the case of Burial 1, being apparently interred in the plaza established during the middle era) suggests they were important or highstatus individuals in the Oak Hill Village community during the early and middle-era communities, although they were not accompanied by many funerary objects. The fourth key relationship is that 13 small pits filled with charred corn (n = 162 cobs) are widely distributed across the Oak Hill Village only during the middle and final village occupations, being concentrated in structure 2, structure group H, structure group E, and particularly structure group D. While their function is debatable, their ubiquity in the final village certainly suggests a definite change in the use of corn between this occupation and the early and middle-era villages. In fact, Dering (2004:table 88) has determined that maize ubiquity in flotation samples increases from 50 percent in the middle village to 96.9 percent in the late village. This local change seems to match stable carbon isotopic evidence from Caddo burials in many areas in northeast Texas (see Perttula 2005:fig. 11-20), which suggests that the importance of corn in the diet of Caddo peoples began to intensify after 1300. Fifth, although it is unfortunate that so little archaeological evidence is currently available from structure group F (particularly radiocarbon dates and decorated ceramics) to accurately determine when these structures were built, four of the small structures (less than 3 m in diameter) in structure group F may be above-ground granaries built and used during the final village (cf. Story 1998:26; Spock 1977). These are the only four very small buildings at Oak Hill Village, and it is tempting to speculate—given the ubiquity of pits with charred corn discussed above—that these structures were also built and used during the final village Finally, while some midden deposits in Area A accumulated throughout the settlement of each Oak Hill Village community, a clear relationship or association between both the northern and southern parts of the site (separated by more than 50 m) is not fully established until the final 232

Perttula and Rogers

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

village community. Midden B (feature 106) near structure 5 is the only midden deposit associated with the early and middle-era communities at the northern end of the site. If the use of Area A for trash disposal cannot be clearly linked at these times with the remainder of the site (or at least not until the final village), it leaves open the question of where trash was deposited during the early and middle-era villages, or indeed if the occupations were of sufficient length and intensity for midden deposits to develop. Perhaps midden deposits in Area A represent a community trash dump shared by many households in the village.

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Regional Context for Establishment, Aggregation, and Abandonment of the Oak Hill Village At the time the Oak Hill Village was established around A.D. 1150, it was simply a small Caddo settlement that had no more than four households (see fig. 8-9). At that time, this part of northeast Texas was not heavily populated, and Story (2000:23) has suggested that most of the Caddo living on the Sabine River and areas to the south lived in the large civicceremonial centers at the George C. Davis and Hudnall-Pirtle (41RK4) sites (see fig. 8-1). Settlements were widely scattered (see Story 2000:fig. 5), but considerable similarities in the styles of the ceramics (especially the engraved fine wares) found on these early Caddo sites suggests that social distances were minimized by extensive contact and interaction. If all four structures at Oak Hill Village were occupied at the same time, we estimate (based on an enclosed area of 337 m2) that the population was between 54–71 people (Rogers 2004:table 34), with each house structure occupied by extended families.4 By A.D. 1250, the Caddo settlement at Oak Hill Village had grown into a larger community, with a well-planned arrangement of structures centered around a plaza, and between a small earthen mound at the northern edge of the community and a special purpose structure (structure 2) at its southern end (see fig. 8-10). Houses had entrances that faced the plaza. Since not all the houses were occupied contemporaneously during the Middle Village, with at least three episodes of structure construction and use apparent in some structure groups, the population at any one time was probably between 74 and 95 people (Rogers 2004:table 35); the community population was not significantly larger at this time than the earliest village, just better and differently organized. The Oak Hill Village was not an isolated settlement around A.D. 1250, Evolution of a Caddo Community

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

233

Copyright © 2012. Nebraska Paperback. All rights reserved.

as the cultural landscape of the Caddo peoples was changing dramatically. The land was filling up with large and more independent communities (covering more than 3–5 acres), hamlets, and farmsteads (fig. 8-12). The larger communities tended to occur along tributaries to the main rivers, such as the Sabine River, as did mound centers; these were usually single earthen mounds covering structures, and they were an integral part of the planning and arrangement of the larger communities. These mound centers reflect a “change in emphasis from individualoriented ceremonialism to public building–oriented ceremonialism . . . the major mounds of the Late Caddo era [dating after A.D. 1300] contain the remains of important buildings rather than important people” (Schambach 1996:40–41). In the area of the Oak Hill Village site (see figs. 8-1 and 8-12), the Caddo settlement of the middle Sabine River basin appears to have been more intensive between ca. A.D. 1250 and 1400 than at any other time (Rogers and Perttula 2004:23; Walters et al. 1998; Walters and Haskins 2000). At the mound center at the George C. Davis site, however, Story (1997:64) has suggested that a decline seen there in the elaborateness of the mortuary practices after A.D. 1200 may be a response to population declines after that time. Thus, one of the premier mound centers in northeast Texas was being eclipsed by the expansion and dispersal of many Caddo peoples who were living in a series of new, related, but socially independent farming communities across much of the Pineywoods; the Oak Hill Village was a nexus for one of those communities. It grew in size after A.D. 1250 and flourished, maintaining its importance as a social and political center (albeit for perhaps a small locale) for the next 100 years or more for specific Caddo groups living in the Sabine River basin. Ceramic stylistic changes around this time suggest that local communities were developing their own ethnic and stylistic expressions, while Caddoan area-wide subsistence and bioarchaeological information points to the contemporaneous intensification of maize agriculture. The Caddo societies living between ca. A.D. 1200 and 1400 set themselves apart from the societies that came before as well as those that came after, not just in terms of settlement and subsistence strategies but also in social organization and religious activities. The development at this time of many independent and socially dynamic communities across the Caddoan area is, we suspect, linked to the intensification of maize agri234

Perttula and Rogers

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-12. Middle Caddo period ceramic complexes in northeast Texas, including a complex on the Sabine River and tributaries where the Oak Hill Village site is located.

culture among these self-sufficient and dispersed Caddo communities. There was a loosening of long-standing links between the public and ritual in Caddo life, just as there was between the secular and residential, at least at the regional scale. Instead, the ritual and secular relationships in Caddo societies, like the one composed of Caddo that lived at Oak Hill Village, were embedded in the intertwined relationship of the people in the community with the construction and use of the small and single mounds and public structures that characterized many communities. The human and spiritual lived together at Caddo places like Oak Hill Village, as they consisted of “matrilineally organized communities of humans and spiritual beings” (Sabo 1998:172–173). Evolution of a Caddo Community

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

235

Copyright © 2012. Nebraska Paperback. All rights reserved.

There were two or three building episodes at the Oak Hill Village site during the latest village (see fig. 8-11). With an enclosed area of only 318 m2 in the eight structures (including structure 18, the possible residence of an elite Caddo), the maximum number of people who lived at the Oak Hill Village during the latest village was only 25–35. The Oak Hill Village was abandoned by the Caddo about A.D. 1450. The agricultural Caddo never came back to the site, and it also seems to be the case that the Caddo left this part of the Sabine River basin and never returned. What does this abandonment tell us about the end of the Oak Hill Village and, in a larger sense, the success or failure of Caddo groups in adapting and surviving in this particular part of the Pineywoods? In the case of Caddo communities living in the middle of the Sabine River basin, as well as parts of the upper Angelina basin (see fig. 8-1), the regional archeological data on the density and distribution of settlements after ca. A.D. 1450 (see Perttula and Nelson 2003a, 2004a; Walters 2001, 2003; Walters and Haskins 2000; Walters et al. 1998) points to a broad and regional (or larger-scale) cause for regional abandonment around A.D. 1450. We suggest that the root cause of the Caddo’s abandonment of the Oak Hill Village and many other Caddo sites around this time is paleoclimatic (cf. Anderson 2001:166). Climatic proxies such as tree-ring data (Stahle and Cleaveland 1995), reconstructed northern hemisphere mean annual temperatures (Bradley et al. 2003:fig. 6.4; see also Frink and Perttula 2002:figs. 9.1–9.3), and solar radiation inferred from atmospheric delta 14C variation (Bradley et al. 2003:fig. 6.13) all point to increasingly colder and drier conditions (and probably to a shorter growing season) in the region beginning around A.D. 1375 (fig. 8-13) and lasting well into the early sixteenth century. Peak periods of colder temperatures and drier conditions occurred between A.D. 1440 and 1470. For Caddo societies that were dependent upon agricultural crops for subsistence, such as the Caddo group living at the Oak Hill Village (cf. Dering 2004), especially those living in the latest village (ca. A.D. 1350/1375–1450), but were actually living in parts of the region that were more marginal (in soil fertility, rainfall, incidence of flooding, or incidence of droughty weather) for successful and predictable crop production, repeated crop failures brought on by cooler and droughty conditions extending over several years may have had disastrous and stressful results for these Caddo peoples. Having lived in a region subject to re236

Perttula and Rogers

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

8-13. Changes in atmospheric delta 14C and climatic minima (peaks in atmospheric delta 14C) during the three village components at the Oak Hill Village site. The climatic minima during the Middle Village is the Wolf minima, while the later and longer-lasting minima is the Sporer minima.

peated droughts, the Caddo probably developed cropping and storage strategies to successfully sustain themselves through a one- to two-year drought. However, the likelihood of continued poor harvests and the lack of storable food reserves over a decadal-scale drought, when linked with evidence for regional abandonment around A.D. 1450, suggests they were only successful by abandoning these more marginal areas; we do not know where they relocated. Acknowledgments This chapter was reprinted with the permission of the Society for American Archaeology from American Antiquity 72(1). We would also like to thank txu Business Services, txu Environmental Services, and txu Mining for their support during the course of the excavations at the Oak Hill Village site in 1994–1995 as well as during the preparation of the final technical report on the work. Sandra Hannum prepared several of the figures and maps in the article, while the rest were prepared by PBS&J staff. Thanks also to the review comments provided by Pete Thurmond and Mark Walters and two anonymous reviewers. Evolution of a Caddo Community

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

237

Notes

Copyright © 2012. Nebraska Paperback. All rights reserved.

1. The mean age of the three dates from structure 12 in structure group D is 757 ± 53 B.P. (conventional age), A.D. 1065–1350 (one sigma), and A.D. 1158– 1313 (two sigma). [Editors: future analyses of the large suite of radiocarbon dates from the Oak Hill Village may want to consider Bayesian methods in calibrating radiocarbon dates from various archaeological contexts instead of using mean methods, for established refined chronological estimates of the construction of specific houses as well as the probable duration of the different village eras (Bronk Ramsey 2009; Levy et al. 2008; Kidder et al. 2010:131–132, 142]. 2. The mean age of the four charcoal dates from structure 2 is 675 ± 55 B.P. (conventional age), A.D. 1270–1354 (one sigma), and A.D. 1238–1404 (two sigma). The six dates on maize from structure 2 fall readily into two different sets of dates, even though they are primarily from one feature (feature 86) that contained more than 100 charred corn cobs. The first (12- and 14-rowed cobs) group has a mean conventional age of 800 ± 87 B.P. and a two sigma age range of A.D. 1031–1360. The second group of maize dates (8- and 10-rowed cobs) has a mean conventional age of 637 ± 53 B.P. and a two sigma age range of A.D. 1275–1417. 3. The mean age of six charcoal dates from structure 5 is 700 ± 52 B.P. (conventional age), A.D. 1255–1371 (one sigma), and A.D. 1237–1400 (two sigma). 4. Population estimates are based on the works of Cook (1972), Casselberry (1974), and Smith (1978), who have proposed different measures of population using floor-space equations. For example, Smith (1978b) proposed that in small settlements, approximately 4.74 m2 of floor space was typical for each occupant.

238

Perttula and Rogers

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

9. Settlement Patterns and Variation in Caddo Pottery Decoration A Case Study of the Willow Chute Bayou Locality jeffrey s. girard

Copyright © 2012. Nebraska Paperback. All rights reserved.

Introduction In the Caddo area, variation in pottery decoration serves as the major criterion for dividing the archaeological record into the spatial and temporal units upon which almost all studies are predicated. Despite being of fundamental importance, our understanding of such variation is limited, in part, because of a lack of attention to problems of scale. Regional patterns are difficult to comprehend when little is known about variation at local levels. This study examines variation in decorations on Caddo pottery from a series of archaeological sites located along Willow Chute Bayou, a secondary stream in the Red River floodplain of northwest Louisiana (fig. 9-1). The sites share similar ceramic assemblages and appear to date between approximately A.D. 1100 and 1450. The Willow Chute locality extends for nearly 12 km on the eastern side of the Red River floodplain, a spatial scale comparable with Caddo floodplain villages described in historic records (Swanton 1942; Schambach 1982a; Trubowitz 1984) and with patterns found elsewhere in late prehistoric contexts in the southeastern United States. Floodplain villages on the larger rivers were one of the basic community patterns for the Caddo people from at least the twelfth century. Examination of ceramic decorative variability at the spatial level of the locality should be of fundamental importance for understanding patterns across broader expanses of the landscape. In this chapter I examine variation in ceramic decorations for contexts within a single site, the Vanceville mound site (16BO7), and for different sites within the Willow Chute locality. Almost all of the site collections contain pottery considered diagnostic of the traditional 239

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

9-1. Location of the Willow Chute Bayou locality.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Bossier focus (Webb 1948, 1983), the Middle Caddo period (ca. A.D. 1200–1500) cultural taxonomic unit recognized throughout northwest Louisiana. However, there is considerable variation in the percentages of types for differing contexts. Because of the likelihood of close social interactions between households, ideas concerning pottery decoration probably were widely shared within Caddo communities, and thus social factors should be minimal for generating decorative diversity within the locality. Analyses of the Willow Chute data presented here suggest that decorative variation reflects differing times of occupations and, to a lesser degree, distinct site activities. I argue that attempts to understand regional social configurations based on ceramic distinctions must take these variables into account. Dispersed Floodplain Villages A dispersed Caddo floodplain village was depicted on the Terán map of 1691–1692 (Swanton 1942:plate 1) (fig. 9-2). The Hatchel site (41BW3) in 240

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

9-2. Terán map of 1691–1692. (Courtesy of the J. P. Bryan Map Collection, CT0108, the Center for American History at the University of Texas at Austin.)

Bowie County, Texas, may be a remnant of a portion of this village (Perttula 2005). Schambach (1982a) and Trubowitz (1984) made detailed studies of the map and presented a set of expectations regarding the archaeological record of a dispersed Caddo floodplain community. The Terán map shows 25 clusters of buildings or compounds separated by hedges or low embankments. The compounds are located on both sides of the Red River, with some houses apparently situated along abandoned course segments. Schambach (1982a:7) suggested that the entire village extended for at least 4 km along the river. On the basis of new archaeological studies in the Hatchel locality and new geoarchaeological studies of the Red River floodplain (Guccione and Hays 2008), Perttula suggests that the village covered an area of about 9 km. One compound, located at the western end of the village, contained a mound with a structure on its summit. Schambach and others have identified a possible elite residence east of the mound (see Sabo, chap. 15, this volume). Early development of the dispersed village pattern in the Caddo area is not well understood. In the Red River floodplain of southwest Arkansas, numerous Late Woodland period (Fourche Maline) sites have been identified. Although little information is available for non-mound contexts, the sites likely represent small aggregated settlements (Kelley and Coxe 1998:213). In northwest Louisiana, only a small number of sites that might date to the late Woodland period (ca. A.D. 400–800) are known, and none have been subject to significant study. One village appears to Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

241

Copyright © 2012. Nebraska Paperback. All rights reserved.

have been located at the Mounds Plantation site (16CD12), across the Red River from the Willow Chute locality. The date of the earliest occupation of Mounds Plantation is not known, but substantial habitation apparently took place in the interval from approximately 900 to 1050 when ceramics with Coles Creek decorations were common in the region. Mounds Plantation developed into a regional ceremonial center by the twelfth century, eventually including seven mounds arranged around a plaza with two or three additional mounds nearby (Webb and McKinney 1975). Concentrations of habitation debris located along an old channel scar on the northern side of the site indicate that the site contained a relatively large residential population during the Early Caddo period (ca. A.D. 900–1200). Deceased village leaders were buried in Mound 5 at the eastern edge of the plaza. A central shaft grave contained the remains of 12 individuals, including males, females, and juveniles. This grave contained a wide range of artifacts, including highly polished and decorated ceramic vessels, a long stem Red River style ceramic pipe, split cane matting, tubular stone beads, finely made Gahagan bifaces, copper-plated stone ear spools, caches of arrow points, and a wooden bow. The Mound 5 burials clearly indicate that a social hierarchy existed within the village by the twelfth century. However, the degree to which Mounds Plantation, in its early years, can be considered a “regional” center is not clear. Populations apparently were aggregated in the site area, as few contemporary outlying sites have been identified. Small habitation sites are likely to be present in upland settings (e.g., Lintz et al. 2007), as well as buried beneath floodplain deposits. However, relative to later periods, sites are sparse, suggesting that the incipient elites at Mounds Plantation had influence over a relatively limited population that was concentrated around the mound/plaza complex. After A.D. 1200, Mounds Plantation appears to have lost most of its residential population, and south of the Haley site in southwest Arkansas there were no ceremonial centers maintained by multiple communities, nor is there evidence that some communities were significantly larger or more powerful than others. In the Red River floodplain, the Willow Chute village was one of several that consisted of scattered habitation loci with a few widely separated solitary mounds. The earlier “mound-plaza” site plan disappeared from the southeastern periphery of the Caddo Area. Clarence H. Webb, the first to synthesize the archaeological record 242

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

for the northwest Louisiana portion of the Caddo area, defined the Bossier focus of the Fulton aspect (now the Late Caddo period) and contrasted it with components of the earlier Gibson Aspect (now pre-1400 sites) in being characterized by relatively small sites located primarily in upland settings. Bossier focus sites lacked evidence of mound construction and the elaborate burial practices noted for earlier sites such as Mounds Plantation and Gahagan. Webb (1948:140) initially argued that the Bossier focus represented a dispersal of population into small farming hamlets with a consequent reduction in regional social integration as well as a decline in emphasis on ceremonialism: The people responsible for [the] Bossier Focus apparently were hunters, fishermen and agriculturists living in small, closely placed communities away from the major streams. Locally self-contained, they traded very little, were satisfied with limited contacts and avoided highly developed ceremonials. Webb (1948:139) identified ceramic continuity from the Alto and Gahagan foci to the Bossier focus, but initially hesitated to draw a direct cultural link between the two:

Copyright © 2012. Nebraska Paperback. All rights reserved.

On these same dispersed sites, we see development of the Bossier pottery complex alongside or out of the Alto—chiefly the utility wares. . . . It is still not clear whether Alto Focus is directly ancestral to Bossier (physically or culturally) or whether Bossier peoples, deriving cultural influences from both Coles Creek and Alto, coexisted with Alto people in numerous villages during the latter phases of Alto Focus. A decade later he argued that small upland villages became widespread late in the Alto focus (ca. A.D. 900–1200), with many of these sites showing continuity between the foci. A corresponding spread of occupation within the floodplain may have occurred, with sites having ceramic influences from the Haley site in Arkansas. The early occupation of the Belcher site (the Belcher I occupation) is an example (Webb 1959:201–202). In a still later summary, he noted that many Bossier focus sites were present in the Red River floodplain along the lateral streams both east and west of the present channel (Webb 1983:217– 219). Scattered sites have been identified along Cowhide Bayou and Red/ Stumpy Bayou in north Caddo Parish; and along Red Chute Bayou in southern Bossier Parish (Louisiana Division of Archaeology, site files). Perhaps farthest south are sites in southern Red River Parish such as Hanna, Charles Webb, and Gahagan, all of which exhibit a continuity Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

243

of occupation from the Early Caddo period (Thomas et al. 1980; Hunter et al. 2002; Webb and Dodd 1939). Although substantial populations probably remained in the Red River floodplain during the Middle Caddo period, they were not integrated by large ceremonial centers as was the case prior to the thirteenth century. Most of the identified Caddo sites are located along abandoned courses in older meander belts, away from the active channel of the Red River. These areas consist of extensive contiguous natural levees that would have been relatively free from the frequent flooding and bank caving that occurs closer to the river. The secondary streams likely contained sufficient flowing water for human consumption, acquisition of riverine food resources, and transportation. It may be inaccurate to convey too systematic a picture of Caddo settlements, however. Each community likely had unique histories and timing of expansion, mound construction, and abandonment. Defining distinct communities remains a problem, and understanding how they interacted socially and economically will require significantly more archaeological data than are currently available. The Willow Chute Bayou Locality Natural Setting

Copyright © 2012. Nebraska Paperback. All rights reserved.

Willow Chute Bayou is an abandoned course of the Red River located on the eastern side of the active channel. Geomorphological studies indicate that the Red River has occupied its current valley for the past 12,000 to 14,000 years. It was originally entrenched at least 4–5 m below the present elevation of the floodplain, and it has been aggrading throughout the Holocene (Pearson and Hunter 1993:31). The development of the floodplain has been complex, and it has proven difficult to isolate and date specific depositional surfaces. Saucier and Snead (1989) identified six meander belts within the floodplain and mapped surface deposits associated with each. Willow Chute flows through Hrm2, the meander belt occupied immediately prior to that of the modern river (Hrm1), with portions of an older meander belt, Hmr3, remaining in the southern portion of the Willow Chute area. Extensive back swamp deposits (B) are present between the active and abandoned meander belts. Pleistocene terrace deposits bind the floodplain to the north and east of the locality (fig. 9-3). The time of the shift from Hrm2 to the modern belt is not known. 244

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

9-3. Land surfaces in the Willow Chute Bayou locality (after Autin 1997).

Albertson et al. (1996:appendix C) report an uncorrected radiocarbon date of 4610 ± 60 B.P. from an Hrm2 overbank deposit along Finn Bayou in southwest Arkansas. Archaeological sites along Finn Bayou date back to the Late Archaic period, suggested to date from 2500 to 5000 B.P. in that region (Kelley and Coxe 1998). Radiocarbon dates from surface deposits along Willow Chute suggest that, in the Willow Chute Bayou locality, Hmr2 was active by 3300 B.P. (perhaps considerably earlier). Landforms became stable by the 1700–2100 B.P. interval, indicating that the Red River had shifted to the modern belt by that time (Autin 1997; Girard 2007a). Radiocarbon dates and ceramic data suggest that the area was depopulated during the late fifteenth century, and by the early nineteenth century Willow Chute was an uninhabited back swamp. A 30–60 cm layer of clay or silt loam now covers lower portions of the natural levees. Caddo artifacts have been exposed along the slopes to the bayou by deep plowing. Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

245

Archaeological Investigations

Copyright © 2012. Nebraska Paperback. All rights reserved.

In 1928, Gerard Fowke of the Smithsonian Institution described a "flattopped mound, quadrilateral but not rectangular, 16 feet high and about 100 feet across at the base," located four miles northeast of Shreveport (Fowke 1928:407). Fowke probably was noting one of the four mounds later recorded by Clarence Webb in the Willow Chute Bayou locality. Webb identified single mounds at the Vanceville (16BO7), Werner (16BO8), Swan Lake (16BO11), and Ed Johnson (16BO12) sites. The Vanceville site was mentioned in his 1948 summary of the Bossier focus (Webb 1948), but no descriptions of investigations were provided. However, in his unpublished notes (Webb n.d.), he stated that he dug into the top of the mound and encountered hard soil (mixed sand) at 3 ft below the surface. He also noted that a fired area containing sherds, shell, and animal bone was encountered on the east side of the mound. In a later paper on the Bossier focus, he provided type percentages for 232 recovered sherds from the site (Webb 1983:227–228). Boy Scouts under the supervision of Webb, Mike Beckman, and Robert Fulton excavated the former Werner mound area in 1958 and 1959. The excavations revealed a floor of "packed, level red clay" containing post holes and two ash pits. Most of the post molds were arranged in two concentric circles: the outer circle was approximately 24 m in diameter, and the inner circle was 14 m in diameter. Webb interpreted the pattern as representing a large structure with an inner chamber. The nature of the artifact and faunal assemblages suggested an exclusively ceremonial function for the site. A sample of charred wood from the site yielded a radiocarbon age of 510 ± 70 B.P. (TX-628; wood charcoal, cal A.D. 1318–1450, 1 sigma) (Webb 1983:219–226). Webb described the Swan Lake mound as circular with a diameter of 60 ft at the base and 20 ft at the summit. It was 8 to 10 ft tall and contained several shallow holes at the top. In several places around the mound, 3 to 4 inches of sandy soil were noted, but no sherds were found on or around the mound. However, he did find a "knife or lance point of banded chert" on the surface in a field near the mound (Webb n.d.). The landowners subsequently recovered another large stemmed biface from an eroded portion of the mound. They also had a garden from which numerous sherds were recovered. Subsequent investigations have demonstrated that the site was occupied during the Middle Woodland period, 246

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

abandoned for many years, and reoccupied by the Middle Caddo period inhabitants who constructed the mound (Girard 2007a). The mound at the Ed Johnson site was plowed down before Webb visited the area. Webb (n.d.) stated that the mound apparently was 50 to 60 ft in diameter. Two large points made of "rose-red quartzite" were found by the landowner in a nearby field. Both specimens are now in the collections of the Louisiana State Exhibit Museum in Shreveport. During the 1980s, Louis Baker, an amateur archaeologist associated with the Louisiana Archaeological Society, recorded extensive artifact scatters along Willow Chute when deep plowing and erosion exposed materials along breaks in slopes to the stream channel (fig. 9-4). Most sites consist of artifact scatters that range from 50 to 200 m in extent along the natural levees bordering the stream. Baker attempted to delimit each scatter, plotted them on 7.5 ft usgs topographic sheets, and assigned field numbers to the scatters. Claude McCrocklin, another member of the Louisiana Archaeological Society, recorded some of Baker's sites with the Louisiana Division of Archaeology, where they were assigned state trinomials. McCrocklin also recorded a few additional scatters not investigated by Baker. Over a period of slightly more than 10 years, Baker repeatedly collected the sites after plowing during the winter months, recovering everything he saw, including small undecorated sherds and chipping debris. Beginning in 1992, Baker and I revisited all of the scatters as an ongoing project of the Regional Archaeology Program at Northwestern State University of Louisiana. We surveyed additional areas along Willow Chute as well. When surface scatters were found, collections were made in the same manner. Baker and I also excavated several test units at the Vanceville, Festervan (16BO327), and Swan Lake sites. The surface collections and excavated materials constitute the data for the analyses presented in this study. The sites are concentrations of artifacts and likely represent distinct households or activity areas similar to those depicted on the Terán map. Since little excavation has been carried out along Willow Chute, we have no information about possible house patterns or other features. However, lumps of fired clay, many with impressions of grass or other vegetation (daub), are ubiquitous, suggesting the presence of structures. All sites contained moderate amounts of chipping debris and low numbers of stone tools, particularly arrow points, retouched flakes, perforators, and celt fragments. Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

247

Copyright © 2012. Nebraska Paperback. All rights reserved.

9-4. Distribution of recorded archaeological sites along Willow Chute Bayou.

The most intensively occupied portion of the locality appears to have been near the Vanceville mound site (16BO7) and along a stretch of the bayou to the north that was initially recorded as a single site designated 16BO168 (called the Vanceville Village) (see fig. 9-4). The natural levee in this area consists of a series of low rises separated by swales. Baker collected each rise as a distinct unit, designated Areas 1 and 2 in the mound area (now assigned to separate sites 16BO7 and 16BO264) and Areas 3 through 11 to the north. We are not sure whether each of the collection units represent distinct activity areas or households, or whether occupa248

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

tion was essentially contiguous with the swales occurring from lateral drainage of the fields in historic times. Although survey coverage of the area has been extensive, there are gaps due to the presence of modern roads, businesses, and residential areas. The locale is in one of the most rapidly expanding metropolitan areas in the state, and many of the sites collected by Baker have now been lost due to recent residential and commercial developments. Despite these limitations, the Willow Chute locality is excellent for addressing questions regarding ceramic variation at a village level for several reasons. First, the data base is on a spatial scale comparable with Caddo floodplain villages described in historic records (Swanton 1942; Schambach 1982a; Trubowitz 1984) and with patterns found elsewhere in late prehistoric contexts in the southeastern United States. Second, the locality contains numerous archaeological sites with similar ceramic assemblages and is bounded by natural areas that have significant limitations for human habitation. Communication within the area would have been easy as all sites are within a half-day walk along contiguous well-drained natural levees. Finally, most of the areas suitable for prehistoric settlement recently were in cultivated fields with excellent surface visibility. Baker collected everything that was exposed and carefully recorded site locations on topographic maps.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Synopsis of Human Occupation in the Willow Chute Locality Pre-Caddo Occupations Dart points and pottery sherds along the edge of the Pleistocene terrace north of the locality indicate sporadic occupation beginning in the Middle Archaic period (ca. 6000–2000 B.C.) and continuing into the Early to Middle Woodland period (ca. 500 B.C. to A.D. 400). Several Archaic period dart points recovered from the Allison site (16BO167) have rectangular stems and resemble those from the Middle Archaic period Conly site (16BI19), which has been radiocarbon dated to the 7500 to 8000 B.P. interval (Girard 2000). Also likely pertaining to the occupation of the locality in the Archaic period are a flat celt, a magnetite plummet, and a fragment of an engraved slate pendant. Pottery from the Allison site has distinctive poorly fired paste that contains abundant grog and sand as temper, with some fragments of crushed sandstone and bone. Only two specimens are decorated, both with wide incised lines resembling Marksville period pottery from the lower Mississippi valley. Deposits Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

249

Copyright © 2012. Nebraska Paperback. All rights reserved.

at the Allison site are shallow, and it appears that materials pertaining to multiple occupations are present that spanned the Middle Archaic through Middle Woodland periods. A small mound dating to the Middle Woodland period was present at the Jim Burt site (16BO23) (Webb 1984; Girard 2002). The mound is situated on Pleistocene terrace deposits west of the Allison site. Webb (1984:264–265) and colleagues began an excavation into the mound but stopped when historic burials were encountered. They did recover a few thick, clay-grit-bone-tempered sherds, some with irregularly incised lines, and classified the mound with the pre-Caddo Bellevue focus. In 1995, Baker and I excavated a test unit 8 m south of the mound and recovered similar sherds along with flakes and fire-cracked rock. A lump of charcoal from the unit was radiocarbon dated to 2200 ± 70 B.P. (Beta8280; wood charcoal; δ13C = -25.5%) (Girard 2002:14). The Jim Burt mound is one of several isolated burial mounds that relate to the Early to Middle Woodland periods in the Trans-Mississippi South (Webb 1984; Schambach 1997). Most are located on upland terrace settings overlooking the floodplains of the Red River and its major tributaries. Little or no habitation debris is associated with the mounds, suggesting their construction and use by relatively mobile populations. Habitation sites relating to these periods have been difficult to isolate, possibly because of the minimal use of ceramics at that time in this area. It also is possible that many components are deeply buried in floodplain deposits. Use of the floodplain during the Woodland period is evident at the Swan Lake site, located along an oxbow of that name within the Willow Chute locality. The oxbow might be a channel segment that relates to an old course of the Red River in the Hmr3 meander belt. Test excavations at the site identified a deep pit feature that contained burned sediments at its base along with thick, grog-tempered sherds and a Gary dart point. Radiocarbon dates on charcoal (1690 ± 80 B.P., wood charcoal, δ13C = -28.4%; 1830 ± 70 B.P., wood charcoal, δ13C = -26.4%; 2020 ± 60 B.P., wood charcoal, δ13C = -28.2%) from the base of the pit indicate use sometime in the interval from cal 184 B.C. to A.D. 540 (2 sigma) (Girard 1995). Sherds from other test units and surface collections at Swan Lake also relate to this Woodland period occupation (Girard 2007a). Other evidence of occupation during the Woodland period consists of numerous small Gary points (Gary, var. Camden) found at several sites. The presence of these points poses an interpretive problem as the sites 250

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

lack ceramic evidence of Late Woodland occupation. It is possible that the floodplain was utilized on a sporadic basis and the points are remnants of small camps that were reoccupied as more permanent residential areas during the Early to Middle Caddo periods. It also is possible that some undecorated pottery in the collections relates to pre-Caddo occupations, but we are unable to distinguish it at this time. Early Occupation at the Festervan Site

Copyright © 2012. Nebraska Paperback. All rights reserved.

The permanent village along Willow Chute appears to have started at the north end of the stream across the Red River from the Mounds Plantation site. The Festervan site (16BO327) covers approximately 12 ha and contains a large quantity of Late Woodland (ca. A.D. 400–800) to Early Caddo period ceramics. A deep test pit near the western edge of the site sampled a buried A horizon approximately 1.2 m below the present surface. The horizon was 35 to 40 cm thick and contained artifacts throughout. A few brushed sherds were found at the top of the horizon, but these likely were from a Late Caddo period occupation more abundantly represented in the eastern portion of the site. Ceramics included Coles Creek Incised vars. Hunt or Phillips, Coles Creek, Greenhouse or Blakely, Keo, and Hardy (Phillips 1970). One dentate stamped sherd probably relates to the earlier Middle Woodland type Marksville Stamped. A few specimens of Kiam Incised, and perhaps Dunkin Incised, also were recovered. Three small sherds have engraved lines. Scattered chunks of charcoal recovered near the base of the A horizon were submitted for radiocarbon analysis. The sample dated to 1250 ± 60 B.P. (Beta-8281; δ13C = -25.5%; wood charcoal; cal A.D. 657–934, 2 sigma). The Caddo Dispersed Village The Festervan ceramic assemblage is not replicated elsewhere along Willow Chute Bayou. The site area is relatively extensive, suggesting that population was aggregated there from the Late Woodland period into the interval between A.D. 900 and 1100 when Coles Creek ceramic types (or locally made copies) were common in northwest Louisiana. Given the general absence elsewhere along Willow Chute Bayou of Coles Creek ceramic types, expansion downstream along Willow Chute appears to have begun no earlier than the late twelfth century. By the Middle Caddo period in this part of northwestern Louisiana (ca. 1200–1500), virtually all of the higher natural levees along the stream were utilized by Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

251

the Caddo people. Occupation also is evident in the uplands north of the floodplain. Several sites have been identified on Black Lake Bayou around the margins of the lake. Collections are small and not well studied, but it is possible that the sites represent seasonal, or short-term, occupations by people who resided in the Willow Chute community. Another possibility is that the sites represent the relatively permanent residences of a separate community, or sub-community, tied economically and socially to the floodplain inhabitants. Post-Village Occupations

Copyright © 2012. Nebraska Paperback. All rights reserved.

A significant change in the depositional processes affecting the Willow Chute locality took place following the Caddo abandonment in the mid-fifteenth century. By 1806, Willow Chute had changed from a flowing distributary channel surrounded by well-drained natural levees to a cypress and willow swamp. The Freeman-Custis expedition probably passed through the Willow Chute area in late June 1806. Freeman described the area as “a swamp covered with Willow bushes growing very close, and through which it was almost impossible to propel the boats” (Flores 2002:143). What happened to change the conditions? Two possibilities are suggested: the effects of the “Great Raft” and the effects of the buildup of an alluvial ridge in the modern meander belt. During the early nineteenth century, the complex of log jams known as the Great Raft had diverted the Red River into numerous channels and back swamp lakes from Campti to a point near the head of Willow Chute Bayou (Bagur 2001; Triska 2008). Because the expansion and destruction of the Great Raft has historic documentation, rafting has been considered a specific event rather than a process that also occurred throughout late prehistoric times with varying local effects. It is possible that the “Great Raft” was the latest phase of rafting that happened to be observed and described by European writers. Albertson et al. (1996:24) and Albertson and Dunbar (1993:36) suggest that initiation of the rafting process may have been linked to a hypothesized climatic change toward increasingly dry conditions that took place across the southern Great Plains around 1000 B.P. (Hall 1990). However, no direct evidence of fluvial change at 1000 B.P. has been recognized in this portion of the Red River floodplain, and possible effects of the hypothesized climatic change in the Eastern woodlands east of the plains are not known. Pearson (1986) estimated that the 252

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

onset of rafting occurred at a much earlier time. At approximately 1800 B.P. the Red River may have shifted from directly draining into the Gulf of Mexico to its present course through Moncla Gap and into the Mississippi River. Such a shift would have shortened the river’s course and also could have triggered channel incision, bank erosion, and raft formation. In summary, although it is possible that rafting was a factor in the changes evident in the Willow Chute area, we currently understand little about the timing of this process or its effects on local environments. An alternative explanation is that the Willow Chute area was increasingly prone to flooding as a result of the building of an alluvial ridge in the modern meander belt (Hmr1) of the Red River. As long as the natural levees along the Red River were lower in elevation than those along Willow Chute, overbank flooding in the old meander belt would be an insignificant problem. However, on current usgs topographic maps, the natural levees along the river are 1.5–3 m higher than surface elevations along Willow Chute. At some point the Willow Chute area began to receive vertical accretion deposits during flood stages. Eventually, 40 to 60 cm of silt loam and silty clay loam were deposited over much of the natural levees and point bars along Willow Chute, turning the entire area into a back swamp by the early nineteenth century when described in the Freeman-Custis narrative. It is possible that the Caddo abandonment of the Willow Chute locality in the fifteenth century was due to this increased propensity for flooding because of the buildup of the alluvial ridge to the west. During the sixteenth and early seventeenth centuries, other areas such as Cowhide Bayou to the north and Red Chute Bayou to the south may have been more attractive for habitation. However, at least intermittent use of portions of the Willow Chute locality and nearby areas to the south is evident after the mid-seventeenth century. The Festervan site, described earlier as containing the earliest occupation of the Willow Chute community, is situated at the northwest side of the locality at the highest elevations. Ceramics from the site suggest that little, if any, habitation took place during the Early and Middle Caddo periods, the span when the dominant occupation of the Willow Chute locality occurred. A later reoccupation is evidenced by pottery recovered on the surface and in a test unit on the eastern side of the site. In the test unit, the edge of a pit feature was exposed with an intact ceramic vessel at the base, approximately 1.3 m below the present surface. Although no human bone was found, I suspected that the pit represented Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

253

a human burial, and we left the vessel in place and backfilled the excavation. The vessel was a small engraved bowl that appeared to relate to the Lester Bend variety of the type Natchitoches Engraved, and perhaps it dates to the late seventeenth century (Schambach and Miller 1984:124). Sherds from the buried A horizon from which the pit originated include Late Caddo period types Keno Trailed and Belcher Ridged. Another late seventeenth- to early eighteenth-century Caddo occupation is located along an abandoned channel course 12 km to the south at the Beene Plantation site. Both the late Festervan and Beene Plantation sites might be part of the widespread Yatasi village visited by Henri de Tonti in 1690 and Bienville in 1699 (Girard 2006). These sites are situated on remnant natural levees in the modern Red River meander belt (Hrm1) along abandoned channel segments or oxbow lakes. The settlement configuration appears considerably more fractionated relative to the earlier Willow Chute locality where the Caddos formed a community along essentially contiguous natural levees in an inactive meander belt. Perttula (1992:227) has suggested that Caddo population decreases in the Red River valley after ca. 1650 resulted in a settlement pattern consisting of isolated farmsteads and small family communities.1 Caddo Ceramic Diversity in the Willow Chute Locality

Copyright © 2012. Nebraska Paperback. All rights reserved.

The archaeological record for northwest Louisiana originally was organized through use of the Midwestern Taxonomic Method (e.g., Webb 1948, 1959). The former “foci” from that method now are often referred to as “phases” (sensu Willey and Phillips 1958), but the lack of precise spatial parameters and reliance on ceramic variation make them essentially ceramic “periods” employed throughout the upper Red River drainage in northwest Louisiana. Three periods are generally recognized in this region: Early Caddo period (ca. A.D. 900–1200), Middle Caddo period (ca. A.D. 1200–1500; elsewhere the Middle Caddo period is estimated to end at A.D. 1400), and the Late Caddo period (ca. A.D. 1500–1700; in other regions and localities, the Late Caddo period is estimated to begin at ca. A.D. 1400). These correspond to the Alto or Gahagan, Bossier, and Belcher foci or phases. Almost all of the Willow Chute sites include types relating to both the Early and Middle Caddo periods and would necessarily be classified as “mixed” components if quantitative data were not considered. The absence or paucity of diagnostic 254

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Middle Caddo period attributes (particularly brushed sherds) at a few sites suggest that some areas were occupied and abandoned before approximately 1200. The rarity of Late Caddo period diagnostic attributes further indicates movement out of the locality prior to the mid-fifteenth century. The few available radiocarbon dates corroborate these chronological parameters. Based on current knowledge of the regional ceramic chronology, there is no reason to believe that significant temporal gaps are represented in the Caddo occupation of the Willow Chute locality. The main Caddo occupation along Willow Chute appears to have existed for approximately 300 years, a period within which considerable changes in ceramic decoration likely occurred. For this study, I use a sherd classification that approximates the traditional types that have been developed for Caddo archaeological studies in northwest Louisiana and surrounding areas (e.g., Webb 1948, 1959; Newell and Krieger 1949; Suhm et al. 1954; Suhm and Jelks 1962). Rigorous use of these types usually is not possible when working with sherd collections because the types were defined from complete vessels that often had different decorative patterns on vessel rims and bodies. The Willow Chute sherds are generally small, and breakage at the rim-body juncture is common. The groups employed here are based on attributes that can be monitored in a relatively unambiguous manner on small sherds. The groups are defined on the basis of primary decorative techniques that characterize the major types for the region. Eleven groups are recognized: Appliqué Ridged, Brushed, Curvilinear Incised, Diagonal Incised, Engraved, Linear Punctated, Miscellaneous Punctated, Ridged, Undecorated, and Zoned Punctated. Brushed and Appliqué Ridged In northwest Louisiana, two major groups of brushed pottery generally are recognized: sherds with vertical appliqué ridges that separate brushed or incised zones into panels (coded ap), and sherds that lack these ridges (coded br). Both Pease Brushed-Incised and Bossier Brushed jars often had brushed rims. When working with sherds, it cannot be determined whether the brushed specimens are from vessels that lack the ridges or are simply portions of appliqué ridged vessels derived from areas between widely spaced appliqué ridges. Despite the high numbers of simple brushed sherds, larger specimens tend to have appliqué ridges, and I suspect that simple brushed vessels (Bossier Brushed) Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

255

were far less common than appliqué ridged vessels (Pease Brushed-Incised) in the Willow Chute locality. Brushed pottery has been used as a diagnostic trait of the Middle Caddo period and thus is most common in contexts dating after about A.D. 1200. However, it likely began to appear earlier (Webb 1983:192) and did not decline in use until well into the Late Caddo period (Girard 2006). Ridged Ridged pottery, generally classified as Belcher Ridged, was made by forming clay into narrow ridges with the use of a tool or finger (Webb 1959:136). Sorting difficulties arise in specimens that have the distinctive low ridges as noted for the type, but the ridges are widely spaced and intervening surfaces are lightly brushed (Webb 1983:185; Girard 2007b). Most of the ridged Willow Chute specimens exhibit these characteristics. The paucity of ridged specimens suggests that the Willow Chute locality was abandoned prior to the Late Caddo period when Belcher Ridged became the major type for utilitarian jars in this area. However, as discussed later, there may be other reasons as well. Incised

Copyright © 2012. Nebraska Paperback. All rights reserved.

Incised sherds are separated into three groups: diagonal incised (DI), miscellaneous incised (MI), and curvilinear incised (CI). Decorations consisting of horizontal bands of diagonal lines usually are classified in the type Dunkin Incised. Most incised decorations on the Willow Chute specimens appear to be relatively simple, consisting of a series of parallel lines, either in single or multiple bands. Nested triangle patterns also are common. A small number of specimens have cross-hatched bands, a trait usually considered diagnostic of the type Harrison Bayou Incised. Diagonal incised elements tend to be on rim or neck sherds, indicating that this trait was most commonly applied to the upper portions of jars. Webb (1948, 1983) noted that cylindrical jars tend to be earlier than those with everted rims, and he referred to the latter as a late variant of Dunkin Incised. A large number of specimens from Willow Chute can be identified as neck sherds (everted rims lacking the vessel lip), indicating that this late variant is common. Webb (1983:193) also suggested that notched lips, another common trait in the area, is a later trait. Most sherds classified as “miscellaneous incised” appear to have multiple straight incised lines oriented horizontally. Body sherds that 256

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

have multiple parallel lines in a single direction are classified as “miscellaneous incised” unless a diagonal orientation is evident from the form of the sherd. Horizontal bands formed by multiple parallel incised lines are associated with several types. Vessels decorated in this manner exclusively around the rim generally are classified as a variety of Coles Creek Incised. Most of the Vanceville (and Willow Chute) specimens have multiple lines applied in a relatively sloppy manner and resemble Coles Creek Incised, var. Hardy. However, paste characteristics suggest that they are locally produced and do not represent vessels imported from the lower Mississippi valley. A few specimens have a single row of punctations beneath the band of horizontal lines. Punctation shapes range from short and linear to round. There are no specimens with the arc forms characteristic of the type Weches Fingernail Impressed, nor are there any with the distinctive triangular punctations of the type Coles Creek Incised. Bands of horizontal incised lines on rims are present on some vessels of other types including Kiam Incised, Dunkin Incised, Pease Brushed-Incised, and Sinner Linear Punctated. These all are utilitarian jar forms that were made primarily in the Early and Middle Caddo periods. In the Middle Caddo period, the Caddos combined incised, brushed, and punctated elements in a wide range of patterns on utilitarian jars. Curvilinear incised lines appear to have been used to form scrolls, circles, and filler elements. But in few cases is it possible to determine design motifs or even elements. Use of curvilinear incising was a minor technique in the Willow Chute locality. The specimens may pertain to vessels of the types Crockett Curvilinear Incised or French Fork Incised, both of which were more prevalent in the Early Caddo period. It also is possible that a few specimens of the Late Caddo period, Foster Trailed Incised type, are represented, although no rims that could be related to this type were identified. Punctated Punctated sherds are separated into three categories: miscellaneous punctated (mp), linear punctated (lp), and zoned punctated (zp). Miscellaneous punctated sherds have multiple punctations with no apparent incised framing lines. Vessels of the types Kiam Incised (with incised rims) and Wilkinson Punctated (punctated to the rim), both of which relate primarily to the Early Caddo period, are likely to be represented. Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

257

Linear punctated sherds have small punctations aligned in closespaced rows. The type Sinner Linear Punctated, likely to be primarily represented, has panels consisting of multiple parallel rows in contrasting directions. It is possible that the type Haley Complicated Incised also is represented, although it does not appear to have been abundant in Louisiana. Both types are associated with the Middle Caddo period. Punctations are framed by incised lines on zoned punctated specimens. Most sherds are likely to be from vessels of the type Pennington Punctated Incised, which appears to have been most abundant in the Early Caddo period. It is possible that some examples of Haley Complicated Incised are included in the assemblages. Stamped Although no specimens were recovered in the Vanceville excavations, eight stamped sherds are present in the Willow Chute surface collections. All are likely to relate to the Late Caddo period type Cowhide Stamped. The rarity of these specimens is another indication that the locality was abandoned when Late Caddo period ceramic types were just beginning to be made in the region. Engraved

Copyright © 2012. Nebraska Paperback. All rights reserved.

Engraving was a major decorative technique employed throughout the Caddo area. Complexities of design patterns make it difficult to relate engraved sherds to specific types. For this study I simply lumped all engraved specimens together into a single category with the realization that this category would not likely be useful for chronological purposes. The purpose of including the engraved specimens in the analyses is to identify another source of variation between the collections other than time. Engraved vessels appear to have been most often fine wares that differed in function from the more common jars and simple bowls that account for the vast majority of sherds in household contexts. In northwest Louisiana, a relatively high percentage of bottles and carinated bowls are engraved, and these specimens often have highly polished surfaces, thin vessel walls, and finely textured pastes. Undecorated These sherds likely are from both undecorated vessels and lower portions of vessels decorated primarily around the rims. Most undecorated 258

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

rims probably are from undecorated vessels, although in some cases decorations may have begun a significant distance below the lip. Because of the tendency to brush utilitarian jar bodies in the Middle Caddo period, it was anticipated that undecorated sherds would constitute smaller proportions of collections through time.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Temporal Variation at the Vanceville Site In order to determine whether or not variation in the selected sherd categories is likely to measure temporal changes, I first examine excavation contexts that can be placed in stratigraphic order and for which a few radiocarbon dates are available. The Vanceville mound site (16BO7) is located along the outer bend of a tight meander of Willow Chute Bayou (fig. 9-5). The site consists of a single conical mound and an extensive surrounding midden deposit. To the northwest, a small drainage, possibly related to an old crevasse, appears to mark the edge of the site. Immediately north of the drainage, the land is in pasture and has only been subjected to a few bucket auger tests that have not yielded artifacts. The eastern and southern boundaries of the mound site are not well defined. To the south, artifacts appear to gradually diminish in number with increasingly fine-grained deposits down the back slope of the natural levee. Baker initially collected surface artifacts in a cultivated field on the west side of the mound (Area 1). Additional artifacts later were collected farther east and recorded as a separate site, 16BO264 (later designated Area 2) (see fig. 9-4). In 1992 a wire fence surrounded the mound, and trees and dense underbrush covered the fenced area. The mound was approximately 25 m in diameter and 2 m tall. A relatively flat area, 8 to 10 m in diameter, was at the mound summit. Historic disturbances may have been responsible for the flat summit. Webb reported the mound to be 10 ft high with a basal diameter of 110 by 90 ft (slightly larger than our 1992 estimates). Several pits were visible in the top of the mound. All but one of these was relatively small and shallow. The largest pit was approximately 3 × 3 m in size and 1.5 m deep. The sides had slumped, and it is likely that the excavation was much deeper. Local informants told Baker that the remains of a Civil War veteran were removed from this pit. A child’s skeleton also reportedly had been excavated from the mound, and a headstone with a burial date of 1890 lay at the base on the southwest side. In 1992 Baker and I excavated several auger tests and a 1 × 1 m test pit Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

259

9-5. Contour map of the Vanceville mound site (16BO7).

Copyright © 2012. Nebraska Paperback. All rights reserved.

(TP92) west of the mound (see fig. 9-5). We excavated another test pit in the spring of 1996 east of the mound in an area that was slated to be disturbed by house construction (TP96A). Two more test units (TP96B and TP96C) were excavated the following fall in the immediate vicinity of the mound. The general depositional context was evident from the auger tests. A 30–60 cm thick layer of clay loam overlay a buried natural levee deposit consisting of very fine sandy loam sediments. The buried A soil horizon was 30–40 cm thick and contained numerous artifacts. In TP92, one pit feature contained a stash of 66 sherds that were sorted into eight vessel groups. No complete vessels could be reconstructed, but the vessel groups represent four Pease Brushed-Incised jars, one engraved carinated bowl, one untyped trailed vessel of unknown form, one base to a black polished vessel of unknown form, and one portion of an undecorated vessel, also of unknown form. Also recovered from the fill were 11 smaller sherds not assignable to any of the vessel groups, seven chert flakes, one flake from a polished celt, one fractured chert pebble, and two pieces of stream gravel. Charcoal fragments present at the base 260

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

of the pit yielded a radiocarbon age of 475 ± 40 B.P. (uga-00882, wood charcoal, δ13C = -22.88%, cal A.D. 1329–1484). Another pit feature from TP92 was filled with charred sweetgum fruits. A sample submitted for radiocarbon analysis yielded an age of 730 ± 70 B.P. (Beta-61743; charred sweetgum; δ13C = -27.7%, cal A.D. 1163–1398). A small pit feature (feature 18), possibly a post hole, encountered in TP96A, contained charcoal that yielded an age of 630 ± 60 B.P. (Beta-94395; charcoal; δ13C = -20.1%; cal A.D. 1276–1415). Distinct depositional strata containing artifacts were present in the two units (TP96B and TP96C) placed near the mound. These strata enable us to view the regional ceramic chronology in the context of changes within the Vanceville site. The immediate area surrounding the mound was used as a mule pen from at least the 1950s and had not been plowed since that time. The slope of the mound on the south side was relatively gradual and the edges were not distinct, either a result of slope wash or the presence of a constructed low platform adjacent to the mound. During the fall of 1996, Baker and I excavated a 2 × 2 m unit (TP96B) on the lower portion of the slope, approximately 16 m southeast of the mound summit. A thin (2–5 cm) layer of fine sandy loam was present directly beneath the surface of the unit (fig. 9-6). Recovered within this layer, likely slope wash from the mound, were a small number of prehistoric and historic artifacts, including a wire-wound glass trade bead. The upper deposit was underlain by the clay loam overburden (1A horizon) that covers the entire site. At approximately 35 to 40 cm below the surface (bs), the 2A horizon was encountered. The deposit became gradually lighter in color with depth until 55 to 65 cm bs, at which point a second buried A horizon (3A horizon) was encountered and artifact counts increased markedly. The 3A horizon continued to approximately 1 m bs. Unfortunately, the change from the 2A to the 3A horizons was too subtle to detect with a bucket auger, and we were not able to trace its extent. The burial of the 3A horizon by the 2A sediments, probably due to overbank flooding, must have been a fairly quick event. It seems unlikely that the 2A horizon represents slope wash because sediments washing off of the mound would have gradually thickened the A horizon on the top of the natural levee, rather than resulting in the distinct depositional break present in the profiles. Also, as discussed below, both horizons were present beneath the mound in TP96C. Scattered charcoal recovered from the 2A and 3A horizons in TP96B Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

261

Copyright © 2012. Nebraska Paperback. All rights reserved.

was submitted for radiocarbon analysis. The sample from the 2A horizon had an age of 616 ± 44 B.P. (uga-01290, wood charcoal, δ13C = -24.7%, cal A.D. 1294–1408, 2 sigma). The age of the sample from the 3A horizon was 665 ± 39 B.P. (uga-00884, wood charcoal, δ13C = -23.4%, cal A.D. 1272–1395, 2 sigma). A third sample was recovered from a probable post hole (feature 24) that likely originated from the top of the 3A horizon. An age of 720 ± 42 B.P. (uga-01287, wood charcoal, δ13C = -27.85%, cal A.D. 1221–1387, 2 sigma) was obtained from this feature. A 1 × 2 m test pit (TP96C) was later placed on the eastern slope of the mound with the hope of recovering charcoal in the sub-mound natural levee deposit to help date the mound construction. From the present surface to a depth of 60 to 80 cm, deposits in TP96C consisted of a reddish-brown fine sandy loam (fig. 9-7). The upper portion was relatively homogeneous in appearance, but the underlying deposits were mottled with small (generally 1–3 cm) patches of silt loam. These patches were scattered irregularly throughout the deposit. The upper deposit probably represents mound fill redeposited by slope wash; the lower deposit probably is in situ mound fill. Although the mound fill differed in appearance from the surrounding natural levee sediments, no distinct basket loads were visible. The boundary between the two strata was irregular, possibly due to natural disturbances. Artifacts were present throughout these deposits. Two charcoal samples recovered from the mound fill deposit were submitted for radiocarbon analysis. One result was 538 ± 43 B.P. (uga-01289; charcoal; δ13C = -26.78%, cal A.D. 1305– 1441, 2 sigma). The other, 370 ± 60 B.P. (Beta-106202; charcoal; δ13C = -24.8%, cal A.D. 1440–1643, 2 sigma), was considerably later. The boundary between the mound fill and underlying natural levee was abrupt and well defined. As in TP96B, both 2A and 3A soil horizons were identified in the buried natural levee. The 2A horizon was almost 30 cm thick on the northwest side and less than 15 cm thick on the southeast (down slope) side, suggesting that it might be thicker beneath the mound relative to the surrounding areas. A charcoal sample yielded an age of 592 ± 43 B.P. (uga-01288; charcoal; δ13C = -25.09%, cal A.D. 1299–1415, 2 sigma). The upper portion of the 3A horizon, although somewhat wavy, occurred at the same elevation as detected in TP96B. A small scatter of charcoal was recovered in the upper portion of the 3A horizon in TP96C-2. The result conforms to other samples from the buried natural levee deposits: 680 ± 60 B.P. (Beta–105513; charcoal; 262

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

δ13C = -24.7%, cal A.D. 1228–1405, 2 sigma). One feature, probably a post hole, was detected in the sub-mound deposits. Given the relatively abundant artifact content of the mound fill, I initially thought that the mound was constructed of old midden deposits. However, the date and ceramic evidence (see below) suggest that the mound fill contains materials that postdate the sub-mound midden. It thus is more likely that material in the mound fill resulted from habitation or other cultural activities that took place on the mound. The contexts at the Vanceville site enable us to address the problem of temporal variability in sherd collections, while minimizing possible effects of spatial variation. Materials from the 3A horizon in TP96B and TP96C are likely to have been deposited earlier than those from the 2A horizon. The mound fill context represents deposits transported by human agents from alluvial deposits, probably in the immediate vicinity of the mound. If artifacts contained therein were included in the deposits when the borrowing took place, I would expect the sherd counts to mirror those of the upper or lower midden (or a combination of the two). As noted previously, it appears more likely that the mound was constructed from deposits not containing artifacts (either prior to deposition in the borrow area or from somewhere off the site). Sherds were incorporated into the mound fill after the mound was in place, indicating that some activity, perhaps domestic habitation, took place on the mound summit in a manner similar to that at the Belcher site (Webb 1959). Thus there should be three temporally discrete and superimposed contexts for viewing changes in ceramic assemblages. Radiocarbon analyses (table 9-1) generally follow the stratigraphic order and suggest that the lower midden formed in the late thirteenth or early fourteenth century; the upper midden (2A horizon) formed in the fourteenth century, and use of the mound dates to the fifteenth century. The midden west of the mound sampled by TP92 may be contemporary with the Caddo occupation on the mound. Table 9-2 compares contexts that can be stratigraphically interrelated in TP96B and TP96C at the Vanceville site. The Lower Midden directly underlies the Upper Midden in both units. In TP96C, the lower portion of the mound overlays the Upper Midden with a few levels sampling both (the “Mound fill–Midden” context). TP92 and TP96A, and the surface collection, are included in table 9-2 for comparative purposes. Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

263

Copyright © 2012. Nebraska Paperback. All rights reserved. The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

9-6. Profile of test pit 96B.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Unit TP96C TP92 TP96C TP96C

TP96B

TP96A TP96B TP96C

TP96B TP92

Lab. No.

Beta-106202

UGA-00882

UGA-01289

UGA-01288

UGA-01290

Beta-9439

UGA-00884

Beta-105513

UGA-01287

Beta-61743

F4

F24

Lower midden

Lower midden

F18

Upper midden

Upper midden

Mound fill

F3

Mound fill

Feature

730 ± 70

720 ± 42

680 ± 60

665 ± 39

630 ± 60

616 ± 44

592 ± 43

538 ± 43

475 ± 40

370 ± 60

Conventional Range (B.P.) 1-sigma Calibration

A.D. 1260-1300 (0.94) A.D. 1373-1378 (0.06) A.D. 1219-1303 (0.88) A.D. 1366-1383 (0.12)

A.D. 1329-1343 (0.245) A.D. 1395-1432 (0.755) A.D. 1305-1355 (0.708) A.D. 1361-1365 (0.061) A.D. 1386-1403 (0.23) A.D. 1302-1329 (0.406) A.D. 1343-1370 (0.385) A.D. 1381-1395 (0.209) A.D. 1291-1325 (0.409) A.D. 1344-1394 (0.591) A.D. 1281-1310 (0.523) A.D. 1360-1386 (0.477) A.D. 1271-1316 (0.582) A.D. 1354-1389 (0.418)

A.D. 1451-1523 (0.989) A.D. 1572-1629 (0.011) A.D. 1417-1447 (1.00)

Table 9-1. Summary of radiocarbon assays from the Vanceville mound site

Copyright © 2012. Nebraska Paperback. All rights reserved.

A.D. 1272-1328 (0.517) A.D. 1341-1395 (0.483) A.D. 1228-1232 (0.008) A.D. 1239-1248 (0.015) A.D. 1251-1405 (0.977) A.D. 1221-1314 (0.854) A.D. 1353-1387 (0.146) A.D. 1163-1332 (0.809) A.D. 1337-1398 (0.191)

A.D. 1276-1415 (1.00)

A.D. 1294-1408 (1.00)

A.D. 1329-1340 (0.014) A.D. 1396-1484 (0.986) A.D. 1305-1366 (0.367) A.D. 1386-1441 (0.633) A.D. 1299-1415 (1.00)

A.D. 1440-1643 (1.00)

2-sigma Calibration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

51.31

Total 13.64

25.93 23.33 16.48 9.83 17.45 9.40 7.82

BR

8.04

20.99 15.56 10.53 2.49 12.77 1.96 8.53

AP

9.65

4.12 6.67 10.98 12.60 7.17 11.03 10.30

MI

4.24

5.35 5.56 5.72 3.88 4.36 3.86 2.84

EN

5.21

1.51 2.22 5.72 6.79 3.43 6.63 5.33

MP

5.00

3.16 4.44 3.66 8.45 2.49 4.87 5.86

DI

0.67

0.27 1.11 0.92 0.83 0.62 0.81 0.36

LP

0.25

0.82 1.11 0.69 0.00 0.00 0.07 0.00

RI

1.64

1.37 0.00 1.60 1.25 0.93 2.37 1.24

ZP

Note: UN = undecorated; BR = brushed; AP = applique ridged; MI = miscellaneous incised; EN = engraved; MP = miscellaneous punctated; DI = diagonal incised; LP = linear punctated; RI = ridged; ZP = zoned punctated.

36.35 40.00 43.48 53.60 50.78 58.46 57.19

Moundfill Moundfill-midden Upper Midden Lower Midden Test Pit 92 Test Pit 96A Surface

UN

Table 9-2. Sherd percentages for contexts at the Vanceville mound site (16BO7)

Copyright © 2012. Nebraska Paperback. All rights reserved.

4340

729 90 437 722 321 1478 563

Total

Copyright © 2012. Nebraska Paperback. All rights reserved. The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

9-7. Profile of test pit 96C.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

9-8. Percentage stratigraphy for contexts at the Vanceville site. ap = appliqué, br = brushed, ri = ridged, en = engraved, lp = linear punctated, ci = curvilinear incised, di = diagonal incised, mi = miscellaneous incised, mp = miscellaneous punctated, zp = zoned punctated, and un = undecorated.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Figure 9-8 presents a percentage stratigraphy (Lyman et al. 1998:243– 244) for TP96B and TP96C at the Vanceville site. Percentage stratigraphy

Copyright © 2012. Nebraska Paperback. All rights reserved.

involves placing the proportional abundances of artifact types per vertically defined assemblage against each assemblage’s vertical provenience within a single site. Unlike frequency seriation, which orders assemblages only on the basis of the popularity principle, percentage stratigraphy uses the vertical provenience of collections as the basis of ordering, with the expectation that the ordered frequencies will display a unimodal distribution (the popularity principle). (Lyman et al. 1998:243–244) With only minor exceptions, the graphs show that the unimodal expectation is met in this case. Percentages suggest that the lower midden dominates the “combined midden” samples. The results conform to expectations regarding regional changes in decorated sherd frequencies based on the traditional cultural historical studies of Webb and Krieger. Brushed sherds, both with and without appliqué ridges, became increasingly abundant. In contrast, undecorated sherds decreased as did other categories generally associated with the Early Caddo period: diagonal incised, miscellaneous incised, and miscellaneous punctated. Engraved sherds maintained relatively constant levels through time. In TP96A, the relatively high percentages of undecorated sherds and low percentages of brushed sherds suggest that the debris in this area dates relatively early in the site occupation (see table 9-2). The sherd data are comparable to the lower midden contexts in TP96B and TP96C, indicating that late midden accumulation did not occur in the eastern portion of the site. In contrast, TP92, on the west side, appears to have sampled a relatively late context, similar to the upper midden. Based on the radiocarbon dates, the lower midden apparently accumulated during the late thirteenth and early fourteenth centuries (early in the Middle Caddo period). Incised and punctated utilitarian jars were relatively numerous at that time, but brushed jars, some with appliqué ridges, were already in use. The upper midden relates primarily to the late fourteenth century with mound construction and habitation perhaps dating as late as the early fifteenth century. The dates and associated ceramics fit well with current understandings of the regional ceramic chronology.

Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

271

Ceramic and Settlement Variation across the Willow Chute Locality

Copyright © 2012. Nebraska Paperback. All rights reserved.

The percentage stratigraphy demonstrates that sherd categories based on major decorative techniques show unimodal distributions through time adjacent to the Vanceville mound. The spatial area sampled in that case was very limited, minimizing possible social and functional variation in ceramic decoration. This section explores the degree to which frequency variation in the ceramic decorative categories between sites can be attributed to shifts in settlement location through time or is likely to have resulted from other factors. Assuming that site duration tended to be less than the entire 300-year span of occupation along Willow Chute and does not vary significantly between sites, a significant portion of the variation between sites should make sense in terms of existing knowledge of the regional ceramic chronology. In other words, the site collections should seriate reasonably well. However, other sources of decorative variation between sites within the locality might be significant. The interpretation of sites as functionally equivalent households might be erroneous in some cases. This is particularly likely with sites that include mounds, such as Vanceville and Werner. It also is possible that intra-community social variation is reflected by differences in pottery decorations between sites. Rather than assume that the collections would order along a single (temporal) gradient, I here use correspondence analysis to determine to what degree the frequencies result from multiple factors (Smith and Nieman 2007:57–58). Correspondence analysis reduces multiple variables to a smaller number of components or dimensions with measures of how much information is lost in doing so. The purpose is to obtain a graphical view of the data that enables the analyst to make useful interpretations. Correspondence analysis works directly on count data in a manner similar to the familiar χ2 tests archaeologists use routinely. A graphic display is produced where both cases (sites) and variables (decorative categories) are plotted in the same spatial dimensions (Shennan 1988; Kintigh et al. 2004; Smith and Neiman 2007). The surface collections from the Willow Chute locality were sorted into the same categories used for the Vanceville site. However, I eliminated the ridged, stamped, and curvilinear incised categories because so few examples were present in the collections. For comparative purposes I include the published data from the excavations at the Werner Mound 272

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

(16BO8) (Webb 1983) in the analysis, as well as the excavated contexts at the Vanceville site (table 9-3). I limited the analysis to collections with 50 or more sherds. The large eigenvalue (a measure of the total inertia or amount of variation in the data set) for the first dimension (dim 1) indicates that most of the variation between collections falls along a single gradient (table 9-4). Locations of the scores for the decorative categories strongly suggest that this gradient can be interpreted in temporal terms. The categories brushed (br), appliqué ridged (ap), and linear punctated (lp), all of which are thought to relate primarily to the Middle Caddo period, have high dim 1 scores (fig. 9-9); so does the engraved (en) category. The categories diagonal incised (DI), miscellaneous punctated (mp), miscellaneous incised (mi), and zoned punctated (zp), all of which have been related to the kinds of decorated ceramics found in Early Caddo period contexts in northwest Louisiana, have low scores on the first dimension (fig. 9-9). The contexts near the Vanceville mound are ordered by the correspondence analysis in accordance with the percentage stratigraphy presented earlier. The outlying test pits also are arranged as expected. It is noteworthy that the surface collection obtained from the site falls approximately in the middle of the other site contexts along dim 1 (fig. 9-9, “7-s”) suggesting that it represents a sort of mean temporal position at the spatial scale of “site.” Figure 9-10 is a frequency seriation with the site order following the dim 1 scores. Generally unimodal distributions are apparent with the obvious exception of the undecorated category. Discrepancies probably relate to poor sample sizes as well as the influence of differing site activities. Dunnell (1982) and others have argued that the reason frequency seriation works is because of the tendency for ceramic styles not under selective pressure to wax and wane in a stochastic manner, thus producing the familiar lenticular or “battleship-shaped” curves of seriation charts (Dunnell 1982; Nieman 1995; Lipo et al. 1997). Most of the categories employed in this analysis relate to jars and simple bowls probably routinely used for storage and food preparation in household contexts. There are no obvious functional differences between the categories that would suggest that their differential use was under significant selective pressure. However, the inclusion of the undecorated and engraved categories is probably inappropriate when the goal is to produce a frequency seriation. Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

273

Table 9-3. Counts and percentages of sherd categories for collections in the Willow Chute locality (*collections used in the correspondence analysis)

Copyright © 2012. Nebraska Paperback. All rights reserved.

Site

Appliqué

BO7* 49 BO8* 1399 BO18 3 BO22* 2 BO97 3 BO168-3* 24 BO168-4* 78 BO168-5* 70 BO168-6* 118 BO168-7* 60 BO168-8* 16 BO168-9* 77 BO168-10* 54 BO168-11* 15 BO169* 0 BO219* 5 BO249* 19 BO250* 61 BO251 0 BO252 0 BO253 4 BO255 1 BO256 0 BO257 5 BO258* 0 BO262 0 BO264* 48 BO265 0 BO266* 62 BO267* 49 BO268 2 BO269 1 BO271 0 BO272* 0 BO273 1 BO276 4 BO304* 30 BO305 6 BO306* 8 BO307* 18

11.58 23.48 8.33 1.75 16.67 7.55 13.22 19.83 23.27 21.66 6.96 5.70 9.20 4.18 0.00 0.82 28.79 33.89 0.00 0.00 26.67 25.00 0.00 13.16 0.00 0.00 8.50 0.00 24.70 24.75 6.06 2.50 0.00 0.00 9.09 12.50 25.00 28.57 10.39 21.43

Brushed 57 1376 1 9 3 51 124 65 99 63 16 53 53 22 0 43 2 21 1 0 2 2 0 7 5 0 44 0 39 38 2 5 0 0 0 4 18 5 20 15

13.48 23.09 2.78 7.89 16.67 16.04 21.02 18.41 19.53 22.74 6.96 3.92 9.03 6.13 0.00 7.05 3.03 11.67 14.29 0.00 13.33 50.00 0.00 18.42 5.81 0.00 7.79 0.00 15.54 19.19 6.06 12.50 0.00 0.00 0.00 12.50 15.00 23.81 25.97 17.86

Curvilinear Incised 1 0 0 0 0 0 0 1 2 1 0 9 2 0 0 6 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 1 0 0 0

0.24 0.00 0.00 0.00 0.00 0.00 0.00 0.28 0.39 0.36 0.00 0.67 0.34 0.00 0.00 0.98 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.53 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.83 0.00 0.00 0.00

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Diagonal Incised 11 2.60 76 1.28 0 0.00 1 0.88 0 0.00 14 4.40 29 4.92 8 2.27 2 0.39 2 0.72 13 5.65 56 4.14 34 5.79 10 2.79 9 4.43 48 7.87 1 1.52 2 1.11 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 5 13.16 4 4.65 0 0.00 33 5.84 0 0.00 16 6.37 2 1.01 1 3.03 1 2.50 0 0.00 6 3.73 3 27.27 0 0.00 3 2.50 0 0.00 1 1.30 1 1.19

Engraved 7 647 1 3 1 15 34 23 24 21 5 40 16 9 8 15 3 7 0 0 0 1 0 1 3 0 16 1 9 8 1 3 1 3 0 5 2 0 2 2

1.65 10.86 2.78 2.63 5.56 4.72 5.76 6.52 4.73 7.58 2.17 2.96 2.73 2.51 3.94 2.46 4.55 3.89 0.00 0.00 0.00 25.00 0.00 2.63 3.49 0.00 2.83 33.33 3.59 4.04 3.03 7.50 33.33 1.86 0.00 15.63 1.67 0.00 2.60 2.38

Linear Punctated 0 88 0 1 0 1 0 1 4 0 0 0 2 2 0 14 0 0 0 0 0 0 0 0 1 0 2 0 0 2 0 1 0 0 1 0 0 0 0 1

0.00 1.48 0.00 0.88 0.00 0.31 0.00 0.28 0.79 0.00 0.00 0.00 0.34 0.56 0.00 2.30 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.16 0.00 0.35 0.00 0.00 1.01 0.00 2.50 0.00 0.00 9.09 0.00 0.00 0.00 0.00 1.19

Copyright © 2012. Nebraska Paperback. All rights reserved.

Misc. Incised 18 179 0 8 0 16 42 13 6 3 16 138 84 47 27 82 1 0 0 0 0 0 1 1 4 2 58 0 2 2 0 7 1 22 0 2 1 0 0 2

4.26 3.00 0.00 7.02 0.00 5.03 7.12 3.68 1.18 1.08 6.96 10.21 14.31 13.09 13.30 13.44 1.52 0.00 0.00 0.00 0.00 0.00 50.00 2.63 4.65 13.33 10.27 0.00 0.80 1.01 0.00 17.50 33.33 13.66 0.00 6.25 0.83 0.00 0.00 2.38

Misc. Punctated 9 2.13 61 1.02 0 0.00 9 7.89 0 0.00 17 5.35 28 4.75 5 1.42 3 0.59 0 0.00 10 4.35 66 4.88 32 5.45 15 4.18 8 3.94 35 5.74 0 0.00 0 0.00 0 0.00 1 50.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 30 5.31 0 0.00 1 0.40 1 0.51 0 0.00 2 5.00 0 0.00 5 3.11 0 0.00 1 3.13 0 0.00 0 0.00 1 1.30 0 0.00

Ridged 0 34 0 0 0 1 0 3 2 0 0 2 1 18 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 4 0 0 0 0 0 1 0 0 0 0

0.00 0.57 0.00 0.00 0.00 0.31 0.00 0.85 0.39 0.00 0.00 0.15 0.17 5.01 0.00 0.33 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 2.02 0.00 0.00 0.00 0.00 0.00 3.13 0.00 0.00 0.00 0.00

Stamped 0 0 0 0 0 0 2 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 2 0 0 1 0 0 0 0 0 0 0 0 0 0

0.00 0.00 0.00 0.00 0.00 0.00 0.34 0.00 0.20 0.00 0.00 0.00 0.00 0.00 0.00 0.16 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.35 0.00 0.00 0.51 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Undecorated

Zoned Punctated

Total

264 2099 30 79 11 178 248 164 246 127 151 902 297 219 151 343 40 89 6 1 9 0 0 19 69 13 322 2 122 91 27 20 1 124 6 15 65 10 45 45

7 1.65 0 0.00 1 2.78 2 1.75 0 0.00 1 0.31 5 0.85 0 0.00 0 0.00 0 0.00 3 1.30 9 0.67 12 2.04 2 0.56 0 0.00 16 2.62 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 1 50.00 0 0.00 0 0.00 0 0.00 7 1.24 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 1 0.62 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00

423 5959 36 114 18 318 590 353 507 277 230 1352 587 359 203 610 66 180 7 2 15 4 2 38 86 15 565 3 251 198 33 40 3 161 11 32 120 21 77 84

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

62.41 35.22 83.33 69.30 61.11 55.97 42.03 46.46 48.52 45.85 65.65 66.72 50.60 61.00 74.38 56.23 60.61 49.44 85.71 50.00 60.00 0.00 0.00 50.00 80.23 86.67 56.99 66.67 48.61 45.96 81.82 50.00 33.33 77.02 54.55 46.88 54.17 47.62 58.44 53.57

(continued)

Table 9-3. Counts and percentages of sherd categories for collections in the Willow Chute locality (*collections used in the correspondence analysis) (continued) Site

61 3 0 163 18 18 1 4 1 4 3 7 0 7 0 0 2 1 2585

9.82 17.65 0.00 11.79 7.83 15.00 3.70 6.25 0.64 5.19 13.04 30.43 0.00 33.33 0.00 0.00 11.11 4.55

Brushed 45 5 21 218 23 5 2 3 12 0 5 3 2 3 0 2 8 3 2625

7.25 29.41 3.00 15.77 10.00 4.17 7.41 4.69 7.64 0.00 21.74 13.04 18.18 14.29 0.00 11.11 44.44 13.64

Curvilinear Incised 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 29

0.16 0.00 0.00 0.07 0.00 0.00 0.00 0.00 0.00 0.00 4.35 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Copyright © 2012. Nebraska Paperback. All rights reserved.

BO308* BO309 BO310* BO322* BO323* BO324* BO326 BO328* BO329* BO361* BO384 BO386 BO387 BO388 BO468* BO521 BO522 BO523 Total

Appliqué

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

PB

Girard

Diagonal Incised

Engraved

27 4.35 19 1 5.88 0 26 3.72 17 20 1.45 35 3 1.30 5 0 0.00 3 3 11.11 2 1 1 1.56 14 8.92 5 1 1.30 0 0 0.00 0 0 0.00 2 0 0.00 0 0 0.00 0 0 0.00 0 0 0.00 1 0 0.00 0 1 4.55 2 489 1029

3.06 0.00 2.43 2.53 2.17 2.50 7.41 1.56 3.18 0.00 0.00 8.70 0.00 0.00 0.00 5.56 0.00 9.09

Linear Punctated 1 0 5 11 5 1 0 0 4 1 0 0 0 0 0 0 0 0 149

0.16 0.00 0.72 0.80 2.17 0.83 0.00 0.00 2.55 1.30 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Misc. Incised 71 0 125 46 15 1 0 7 32 6 0 0 0 0 8 0 0 1 1097

11.43 0.00 17.88 3.33 6.52 0.83 0.00 10.94 20.38 7.79 0.00 0.00 0.00 0.00 11.59 0.00 0.00 4.55

Misc. Punctated 21 0 32 7 6 2 0 3 12 6 0 0 0 0 1 1 0 0 431

3.38 0.00 4.58 0.51 2.61 1.67 0.00 4.69 7.64 7.79 0.00 0.00 0.00 0.00 1.45 5.56 0.00 0.00

Ridged 0 0 0 6 2 0 0 0 0 0 0 0 0 0 0 0 0 0 76

0.00 0.00 0.00 0.43 0.87 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Stamped

Undecorated

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 8

372 8 462 866 151 89 19 45 71 59 14 11 9 10 60 14 8 13 8931

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 4.55

59.90 47.06 66.09 62.66 65.65 74.17 70.37 70.31 45.22 76.62 60.87 47.83 81.82 47.62 86.96 77.78 44.44 59.09

Zoned Punctated 3 0 11 9 2 1 0 0 6 0 0 0 0 1 0 0 0 0 100

0.48 0.00 1.57 0.65 0.87 0.83 0.00 0.00 3.82 0.00 0.00 0.00 0.00 4.76 0.00 0.00 0.00 0.00

Total 621 17 699 1382 230 120 27 64 157 77 23 23 11 21 69 18 18 22 17549

Copyright © 2012. Nebraska Paperback. All rights reserved.

Table 9-4. Summary of correspondence analysis Dimension

Eigenvalue

Percent

1 2 3 4 5 6 7 8 Sum

0.180 0.027 0.008 0.007 0.008 0.003 0.003 0.001 0.233

77.28 11.72 3.49 2.85 1.54 1.41 1.28 0.42 total inertia

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Settlement Patterns and Pottery Decoration

PB

Copyright © 2012. Nebraska Paperback. All rights reserved.

9-9. Scatter plot of first two dimensions for the correspondence analysis. Sherd categories (squares) are br = brush, ap = appliqué, lp = linear punctated, di = diagonal incised, en = engraved, mi = miscellaneous incised, mp = miscellaneous punctated, un = undecorated, and zp = zoned punctated.

The analysis provides some insights into the manner in which the Willow Chute community developed. In figure 9-11, the sites used in the analysis are plotted and coded according to their dim 1 score. Rather than a gradual movement or expansion downstream from the Festervan site, it appears that several new households were established at various points along the stream to form core areas that later either expanded or 278

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

were abandoned. This process apparently began during the Early Caddo period with occupation of sites such as 16BO329 north of Festervan, the bend to the east at 16BO310, and the cluster of sites farther east around 16BO169 and 16BO219. These sites contain little brushed pottery and thus appear to have been vacated by the twelfth century, prior to the Middle Caddo period. Sample sizes are too small to determine much about the southeastern end of the Willow Chute locality, but the relatively low dim 1 scores for 16BO361 and the Modica site (16BO22) indicate that habitation had expanded to the southeastern side of the locality early in the Middle Caddo period. The second dimension of the CA (dim 2) accounts for almost 12 percent of the variation in the collection. The second dimension appears to separate contexts with high percentages of undecorated sherds, or in some cases contexts with many appliqué ridged sherds, from those with more incised, engraved, and punctated specimens (see fig. 9-9 and table 9-3). Sites with high dim 2 scores, such as 16BO468, 16BO258, and 16BO324, have high percentages of undecorated sherds, and the late sites with high scores such as 16BO249, 16BO250, and 16BO304 have relatively high percentages of appliqué ridged and undecorated specimens and little else. It is possible that sites with high dim 2 scores represent areas where utilitarian storage and cooking vessels were especially numerous, as opposed to areas with more diverse vessel assemblages. It also is possible that some contexts represent materials lost or discarded directly in everyday household activities, whereas others represent secondary trash deposition. However, in the absence of extensive excavation data, it is difficult to assess activity variation between the collection contexts. In any event, the analyses indicate that, at the spatial scale of the Willow Chute locality, percentages of undecorated sherds do not vary in a unimodal manner through time. Other researchers (e.g., Lipo et al. 1997; Smith and Nieman 2007) reject the use of undecorated specimens in seriation studies because relative frequencies may be measuring other factors rather than stylistic variation and are more subject to sampling bias than other sherds. Contexts where atypical sherd collections might be expected are those on or near mounds. An unusually high percentage of engraved sherds was recovered at the Werner site (16BO8). Webb (1983:221) suggested that ceremonial activities were associated with the very large structure beneath the Werner mound. Engraved sherds from Werner Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

279

Copyright © 2012. Nebraska Paperback. All rights reserved. The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

9-10. Frequency seriation of collections used in the correspondence analysis with order following dim1 scores. br = brushed, ap = appliqué, lp = linear punctated, di = diagonal incised, en = engraved, mi = miscellaneous incised, mp = miscellaneous punctated, un = undecorated, zp = zoned punctated.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Copyright © 2012. Nebraska Paperback. All rights reserved.

9-11. Schematic view of the Willow Chute locality with sites coded according to dim1 scores in the correspondence analysis.

appear to represent primarily serving vessels, especially bottles and carinated bowls. Although difficult to quantify, the engraved pottery from the Werner site appears to be of high aesthetic quality, with polished surfaces and use of pigment in the engraved lines more frequent than at other sites along Willow Chute. It is possible that these fine wares were brought to ceremonies where they underwent a relatively high rate of breakage. It also is possible that the vessels accumulated and were stored at the site under jurisdiction of the community leaders. The structure at Werner preceded construction of the mound; in fact, the mound 282

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

served to cap the remains of the collapsed and abandoned structure. The kind of ritual activity represented at Werner may have been unique in the Willow Chute community and only occurred late in its occupation. Although the Vanceville site also has a mound, the ceramic pattern shows less of a tendency toward high percentages of engraved sherds. Unlike Werner, Vanceville probably had a structure and associated trash deposition on the mound summit, and perhaps it served as an elite residence rather than as a communal ritual activity or meeting area. Limited exploration of the area around and beneath the mound at the Swan Lake site shows no evidence of an unusually high deposition of engraved ceramics (Girard 2007a).2 The overall similarities in the sherd collections throughout the Willow Chute locality suggest that, at the general level of decorative techniques, there was little variation between individual households. However, use of specific design elements and the manner in which these elements were assembled on individual vessels may have varied significantly. Brushed and appliqué ridged jars appear to have been relatively standardized, not only throughout the Willow Chute locality but across the region that Webb (1948, 1983) also defined as the Bossier focus. Greater variation is apparent with incised and engraved specimens, and more detailed classifications and analyses may enable identification of contemporary intra-community decorative variation.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Willow Chute in Regional Perspective The analysis presented in this chapter demonstrates considerable variation in percentages of sherd decorative categories from contexts across the Willow Chute locality and indicates that most of this variation can be attributed to spatial shifts in habitation within the total span of occupation of the village. The results have implications for interpreting patterns on a broader regional scale. In terms of the traditional cultural taxonomy, most of the Willow Chute sites could be classified as components in the Bossier focus. Webb (1948, 1983) defined the focus on the basis of several key sites such as Pease, Sinner, High Island, Marston, Montgomery I, and Mill Creek. With the exception of Montgomery I, these sites have not been systematically excavated, nor do they have associated radiocarbon dates. All, except Marston, are located on non-aggrading upland landforms, and several have archaeological evidence of mixture with earlier occupations. Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

283

Copyright © 2012. Nebraska Paperback. All rights reserved.

Webb (1983:table 10) recognized changes in percentages of ceramic types within the Bossier focus. He sorted 12 components into three groups interpreted as representing different periods. The earliest group is composed of Hanna, Smithport Landing, Colbert, and Greer. All are sites with substantial amounts of Early Caddo and Coles Creek pottery. Sites in the middle group (Mill Creek, Sinner, Pease, and High Island) are located in the lower Red Chute/Loggy Bayou area. The most striking differences between the groups are the lower percentages of Kiam Incised/Hardy Incised and higher percentages of Bossier Brushed, Pease Brushed-Incised, and Belcher Ridged in the Red Chute sites. The final group is composed of the two Willow Chute sites (Vanceville and Werner Mounds), the Marston site in the Red River floodplain near the mouth of Loggy Bayou, and the upland Montgomery I site on a small tributary of Bayou Dorcheat. Incised and punctated sherds in this third group are present only in low percentages, whereas percentages of Bossier Brushed and Pease Brushed-Incised are very high. Although the chronological interpretation of Webb’s groups appears to be valid, he did not address possible spatial variation within the Bossier focus. Gibson (2005) takes the opposite approach, arguing that Bossier sites can be considered a single chronological unit, with variation between sites indicative of differing social units or “tribes.” The Willow Chute study provides additional information regarding this issue by indicating that the spatial scale of “site” likely samples a limited range of temporal ceramic variation for past Caddo communities. For example, Webb (1983:227) noted that his collection from Vanceville came from the surface of the top of the mound and likely related to a final habitation on that surface. From our subsequent work at the site, it is clear that Webb was correct and other contexts at the site represent considerably earlier activities. Webb’s collection is not representative of the midden debris surrounding the Vanceville mound or of the Willow Chute community as a whole. As noted above, the other sample used by Webb from Willow Chute, that from the Werner site, also appears to pertain to a late context within the history of the Willow Chute community. The Vanceville and Werner collections fall together in Webb’s third group, not because they pertain to the same community but because both sherd assemblages were deposited within a similarly late temporal interval within the Middle Caddo period. Without more knowledge of the range of ceramic variation in different localities, identification of 284

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

social units within Webb’s Bossier focus, such as Gibson’s “tribes,” has to be considered tenuous. Despite this caution, there is likely to be some variation on a regional scale that can be attributed to differing ceramic traditions at the community level. For instance, the Vanceville and Werner collections contain some of the lowest percentages of ridged pottery (Belcher Ridged) among the Bossier focus collections (see Webb 1983:table 10). The data presented here demonstrate that this pattern is not confined to those two sites but is a characteristic of the Willow Chute locality in general. Because of the early fifteenth-century dates from the latest contexts at Vanceville and Werner, I think it unlikely that the locality was abandoned prior to the time when Belcher Ridged was made in significant amounts in northwest Louisiana. The relative scarcity of the type might be a distinctive social trait of the community. Belcher Ridged jars probably became popular in some communities during the Middle Caddo period, but remained rare in others such as Willow Chute.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Conclusions A widespread shared ceramic tradition across northwest Louisiana for the Middle Caddo period (ca. A.D. 1200–1500) produced the archaeological manifestation that we recognize as the Bossier focus. Both geographical and temporal parameters are too broad to consider the Bossier focus a “phase” in the traditional sense (e.g., Willey and Phillips 1958; Phillips 1970). Similarities in ceramic decorations suggest considerable interaction between communities, but a diverse range of settlement configurations are evident. These include multiple dispersed floodplain villages, communities surrounding marginal floodplain lakes, and poorly defined upland settlements along tributary streams of varying size. Little is understood about social, political, or economic links between these communities. Were past social groups within the region decorating their pottery in distinctive ways that can be detected by archaeological studies? Perhaps, but currently defined types appear throughout the Bossier focus region. More detailed studies of ceramic decorative elements and motifs are needed to detect those that may be unique to specific communities. What about the notion that different groups can be identified, not on the basis of manufacture and use of distinctive types, but on relative percentages of types represented in sherd debris? The Willow Chute Settlement Patterns in Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

285

data suggest that variation in percentages of currently defined types is more likely to be due to differing times of site occupations. Differential transmission of information between communities probably has resulted in variation on large spatial scales. However, I am skeptical that pottery vessels were produced, broken, and discarded in particular relative quantities over sufficiently long intervals to result in sherd collections with percentages distinctive of specific social groups. Caddo social and spatial boundaries are likely to have been highly dynamic: groups coalesced and fissioned at irregular intervals, villages probably moved as social and economic conditions warranted, and peripheral households may have shifted allegiances. Such factors would limit the degree to which ceramics in the archaeological record cluster into neat spatial and temporal units identifiable by distinctive percentages of sherd types. This study demonstrates the importance of spatial scale when comparing percentages of sherd categories to elucidate past social configurations across the Caddo landscape. Because dispersed floodplain villages were a fundamental form of settlement in the Red River drainage, analysis on a spatial scale similar to that described for the Willow Chute locality provides a view of ceramic variability likely to pertain to a past community. Differing contexts within sites such as the Vanceville mound site exhibit varying percentages of categories due to small-scale shifts in use of space. Changes in household locations within the total span of occupation of the village produce considerable variation in percentages of decorative categories. Such variability must be considered in attempts to understand Caddo social organization on the basis of ceramic decorations across larger geographic expanses. Copyright © 2012. Nebraska Paperback. All rights reserved.

Acknowledgments This research is based largely on the diligent field efforts of Louis Baker of Benton, Louisiana. Test excavations and analyses were carried out through the state of Louisiana’s Regional Archaeology Program based at Northwestern State University of Louisiana. The program has been financed with state funds and with federal funds from the National Park Service, U.S. Department of the Interior. Thanks to Joe Saunders and Robert C. Dunnell for providing comments on the chapter. Notes 1. Kelley (chap. 14, this volume) discusses Late Caddo period settlement in the region in greater detail in his chapter on the Belcher phase. 286

Girard

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

2. On the Terán map, George Sabo (chap. 15, this volume) identifies two central places, the caddi’s compound and the xinesi’s temple complex, and argues that these are of considerable importance for understanding the spatial structure of Caddo communities. For the Willow Chute locality, it is tempting to interpret the Vanceville site as a caddi’s residence and the Werner site as a xinesi’s temple complex. However, contrary to the Terán map situation, the former was associated with a mound and the latter was not (until the large structure was deliberately buried following its destruction).Ceramic analysis and radiocarbon dates indicate that the mound occupation at Vanceville, and activities associated with the large structure beneath the Werner mound, are among the latest events at Willow Chute, likely dating to the early fifteenth century. Thus it appears that neither was present throughout most of the community’s history. It is possible that other uninvestigated areas served these functions but were not associated with mounds and have not been identified in the archaeological record (the chronological positions of the Swan Lake and Johnson mounds are not known). It also is possible that Mounds Plantation, located across the Red River to the west, continued to serve the Willow Chute residents as a sacred place. However, it appears that residential occupation at Mounds Plantation, or other activities that resulted in significant artifact deposition, diminished greatly or ceased as the Willow Chute community expanded.It is important to note that the spatial configuration of the Willow Chute community developed over the course of two to three centuries, perhaps with complex spatial rearrangements of residences and communal spaces during that span. However, Sabo’s study points out the need to consider social issues and cosmological principles, in addition to simple landform configuration, for understanding how a Caddo community developed spatially.

Settlement Patterns and Pottery Decoration

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

287

10. Caddo in the Saline River Valley of Arkansas The Borderlands Project and the Hughes Site mary beth trubitt

Introduction

Copyright © 2012. Nebraska Paperback. All rights reserved.

Recent Caddo archaeology emphasizes the diversity in Caddo lifeways from one river valley to another across a broad region as well as transformations that took place through historical time. For example, Ann Early (2000:127) points to regional “variations on the common Caddo theme” that probably have roots before contact rather than being solely repercussions of historic era interactions with other Native and non-Native groups during the Historic Caddo period. Timothy K. Perttula (1996, 2008) highlights temporal trends with more intensive use of maize and more successful agriculture after A.D. 1400 (see also Wilson, chap. 4, this volume), resulting in sociopolitical changes, shifts in ritual behavior, a lessened reliance on elites, and a reduced social hierarchy. Regional variation and the nuances of differences or connections between communities may be identified in more recent centuries by information on dialects spoken or kin ties connecting Caddo peoples. Reconstructing these cultural patterns is more difficult for the distant past using only the archaeological record and material remains. While ceramic forms and decorations do show variation from river valley to river valley across the Caddo area, pottery vessels—or goods in pots— were traded between communities and between different linguistic and cultural groups. Artifact types may be problematic as ethnic indicators. If the presence of certain pottery types are taken as an indication of Caddo cultural affiliation, what do we do with “the recurring presence of apparently pan-regional (if not pan-cultural or multicultural) very late shell-tempered ceramic types such as Cowhide Stamped, Foster TrailedIncised, and Keno Trailed” (Jeter and Early 1999:60)? Kidder (1992, 288

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

1993, 1998) has emphasized that pottery shows evidence of interaction and exchange between the Caddo and lower Mississippi valley areas during the sixteenth and seventeenth centuries, and argues that Caddo ceramics were traded into non-Caddo communities such as the Keno, Glendora, and Jordan sites in northern Louisiana. In several articles, Perttula (2002; Perttula et al. 2001) has taken the issue of trade in Caddo ceramics head-on by applying instrumental neutron activation analysis to Caddo sherds from northeast Texas and elsewhere to find out where pots originated. This chapter will address this issue of regional variation using material culture and archaeological remains from a Late Caddo period mound site in the Saline River Valley borderlands, with comparisons to sites in the middle Ouachita River valley as well as the Arkansas River valley in central Arkansas. First of all, I need to clarify that there are actually two Saline Rivers in Arkansas: a smaller western one that forms the boundary between modern Howard and Sevier Counties, joins the Little River, and drains south into the Red River; and a larger eastern one that heads in modern Garland and Saline Counties, flows toward the southeast through Grant and Cleveland Counties, forms a boundary between Drew, Bradley, and Ashley Counties, and drains into the Ouachita River in the Felsenthal area near the southern border of the state. It is this eastern Saline River that is the setting for this discussion. This eastern Saline River in Arkansas is generally depicted as the eastern edge of the Caddo area or outside it altogether (for example, Jeter and Williams 1989:figs. 17–21; Perttula 1996:fig. 1; Rogers and Sabo 2004:fig. 1). The Saline River valley has been described as a “borderlands” or boundary zone between the trans-Mississippi South and lower Mississippi valley archaeological areas and between Caddo and Mississippian cultures in these respective areas. It also falls at the borders of Arkansas Archeological Survey research station territories and has not seen much focused research (Jeter and Early 1999). Recent archaeological fieldwork at the Hughes site (3SA11) is one by-product of renewed attention on the Saline River valley that my colleague Marvin Jeter has termed the “Borderlands Project.” Recent Archaeology at the Hughes Site Hughes (3SA11) is a Caddo mound center located in the Saline River valley near the boundary of the West Gulf Coastal Plain and Ouachita Mountains physiographic regions (fig. 10-1). Placed on the National Caddo in the Saline River Valley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

289

10-1. Portion of Arkansas and surrounding states with physiographic regions and locations of sites discussed in text: Hughes (3SA11), Peeler Bend canoe (3SA296), Eastwood Bayou (3JE285), Kuykendall Brake (3PU111), 3CL593; Hardman (3CL418); Winding Stair (3MN496); Standridge (3MN53); Cedar Grove (3LA97), Shallow Lake (3UN9/52), Gordon (3AS152), and Jordan (16MO1).

Copyright © 2012. Nebraska Paperback. All rights reserved.

Register of Historic Places in 1985, the site was deemed significant for its potential to contribute to an understanding of Late Caddo and protohistoric Caddo social and ceremonial systems, settlement patterning and economic organization, and the relationships between Caddo and Mississippian cultures in the centuries before European settlement of the region. Based on its geographic location, this site is important in understanding relationships between regions. Hughes is one of a small group of known mound sites clustered in the middle Saline River drainage, separated from other mound sites located near the confluence of the Saline and Ouachita Rivers (Lockhart 2007, and chapter 11, this volume; Prentice 2000). Artifacts from the site point to occupation in the Late Caddo period (ca. A.D. 1400–1680; Jeter and Early 1999:55–61; Schambach and Early 1982:111–121), and this timing raises the possibil290

Trubitt

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

ity that the site’s inhabitants may have encountered Spaniards passing through the area as part of the de Soto expedition (although evidence of a Spanish presence at Hughes has not been found to date). Investigating community changes between the “prehistoric” and “protohistoric” periods is a potential line of research at this site. Archaeologists and artifact collectors have known about the Hughes site for over a hundred years. It was visited by Edward Palmer in 1883 as part of the Smithsonian Institution’s Mound Survey (Jeter 1990). Palmer’s (1917:406–407) description of the site mentions several smaller mounds that had been leveled previously, but primarily focuses on the large two-stage main mound. In his limited exploration, Palmer found daub and charcoal/ash deposits that indicate there must have been at least one burned structure buried within the mound. A drawing made from a photograph Palmer had taken of the mound was included in Thomas’s (1894:fig. 151) summary volume of mound site investigations. The twentieth century saw a number of archaeological visits to the site by archaeologists. In 1972, several small test units were excavated at the site by amateur archaeologist Wanda Hoffman; the information retrieved from these units was recorded by J. C. Weber, then the Arkansas Archeological Survey’s Henderson State University (aas/hsu) station archaeologist in Arkadelphia. As the Survey’s next hsu station archaeologist, Ann Early conducted limited mapping, soil augering, and shovel testing of the site in 1982 as part of a National Register of Historic Places nomination for the Hughes Mound. Site survey along a gas pipeline right-of-way included limited visits to the Hughes site (Hinkle 1988). Trees and brush cover make photography difficult at the site, but figure 10-2 shows the mound in 1972. The 1972 test units, located near the east flank of the mound, uncovered midden deposits with numerous pieces of well-preserved animal bone and large sherds, including fragments of a vessel typed as Mound Place Incised and Brushed (fig. 10-3), a protohistoric ceramic type at the Jordan site in northern Louisiana (Kidder 1986, 1992). The sherds from the 1972 work, curated at aas/hsu, show a predominance of shell tempering (86 percent, with 8 percent grogtempered and 6 percent with a mixture or other tempers), with a variety of decorative surface treatments (incising and trailing, brushing, punctating, stamping, engraving, and slipping) that resemble those seen on Caddo ceramic types but also show affinities to some Plaquemine and Mississippian ceramic types. Caddo in the Saline River Valley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

291

10-2. Photograph of main mound at the Hughes site (3SA11), taken in 1972. (aas/hsu slide 2943.)

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Hughes site has also attracted the attention of artifact collectors. Anecdotal evidence indicates a cemetery area lay adjacent to the mound (aas/hsu research station photo files record three Bailey Engraved effigy bottles in the Gilcrease Museum’s Lemley Collection that reportedly came from graves dug at the Hughes site in the 1930s). Despite the preservation and stewardship efforts of the landowner, William Davenport, and Arkansas Archeological Society member Dr. Quin Baber, looting at the site continued. My research at Hughes was motivated in part by a need to document damage from recent digging in the summit of the main mound and to discourage further vandalism by refilling looter holes. In 2002 we undertook one field season of mapping, shovel testing, and unit excavation at Hughes as part of a Henderson State University/ University of Arkansas Archeological Field School. The resulting topographic map (fig. 10-4) shows a two-stage mound rising 5.5 m (about 18 ft.) above the surrounding terrace. The mound’s upper stage is conical in shape, and its lower stage extends to the northeast as an oblong 292

Trubitt

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

10-3. Partial vessel, Mound Place Incised and Brushed type, from 1972 testing. (aas/hsu D4313.)

to rectangular platform. There is bulldozing damage to the west end of the mound, as was noted by Early in 1982. Cleaning and profiling three recent mound summit potholes indicated that the deposits contain small fragments of burned clay and charcoal, with a possible burned clay feature exposed near the bottom of one. Despite the disturbances to the main mound, there are probably intact structural features still here. Other Caddo mound sites in southwest Arkansas (e.g., Antoine [3CL60], Carroll Bell [3CL51], Ferguson [3HE63], and probably Hays [3CL6] and Moore [3CL56]) have a similar two-stage configuration for the main mound. At Ferguson and Hays, excavations uncovered contemporaneous yet functionally distinct structures on the upper and lower platforms (Schambach 1996; Weber 1973). No clear indication of the former locations of other mounds was seen from the topography or shovel testing. Systematic shovel testing shows that the Hughes site extends across an area of about 270 × 140 m (roughly 3.8 hectares or 9 acres). In the 35 shovel tests dug in 2002, artifact densities were greatest in the vicinity of the mound and decreased north of the Caddo in the Saline River Valley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

293

10-4. Topographic plan map of the Hughes site.

Copyright © 2012. Nebraska Paperback. All rights reserved.

N280 grid line and south of N125 on the site grid. Artifacts extended to as much as 80 cm below surface (bs) in shovel tests in the site’s central area. Burned clay features were identified in two locales, and larger excavation units were placed to investigate these features further. In three 2 × 2 m units placed northeast of the mound, we uncovered a series of three burned surfaces that are probably portions of burned structures covered with midden deposits (see also Trubitt 2009). A stratified sequence of features (2, 3, 7, and 16) was identified in one unit (N199E244, fig. 10-5a–b). The upper features (2, 3, and 7) appeared to be cut through, and there was some mixing of deposits (several large ceramic sherds were reconstructed from proveniences at different depths in both the N199E244 and N199E248 units). The limited area excavated 294

Trubitt

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

in the summer of 2002 makes it difficult to correlate features and deposits across units, but feature 16 and feature 8 in units N199E244 and N199E248, respectively, are areas of burned clay/daub with charred timbers found at about the same depth (50 cm bs) and interpreted as part of the same burned structure. The site was part of a European American farm/plantation during the nineteenth century, and the area around the mound was used for residential and agricultural activities. The area north of the mound was apparently the focus of digging by collectors at the site during the mid to late twentieth century. Occasional metal and glass artifacts found in upper levels of the excavation units relate to these later activities. The burned clay, daub, charcoal, and charred wood features defined in Unit N199E244 were separated by middens with quantities of animal bone as well as artifacts. In a new analysis of N199E244 fauna, Lucretia Kelly has found a reliance on terrestrial animals rather than those from aquatic habitats. Deer and tree squirrels were most common, some turtles and mussels occurred, and birds and fish were not prevalent. Unusual occurrences in the assemblage were bones of several carnivores (cougar, mink, and gray fox). During excavation in the field and later during the cataloguing process in the lab, several teeth and tooth fragments were identified as possibly human. In 2006, these were examined by Jerome Rose and James Tyler at the University of Arkansas. They identified one tooth as human (2002-414-120-15 from unit N199E244, in the fill above feature 7), but could not make a determination of human or animal for five examined teeth or tooth fragments because of their irregular morphology and fragmentary condition. The human tooth has been curated at the Arkansas Archeological Survey’s Fayetteville facility. Three charred wood samples from the Hughes site were sent to Beta Analytic, Inc., for radiocarbon analysis (table 10-1 and fig. 10-6). Two samples from burned timbers came from features 8 and 16; their similar radiocarbon results support an interpretation that they are part of the same structure and indicate that structure had wood cut within several decades of A.D. 1300. The third date was run on charred wood from feature 7. This burned surface, stratigraphically higher, may be part of a structure that had wood cut within several decades of A.D. 1470. The dates provide temporal brackets for activities that took place in this part of the Hughes site and can be compared with other published dates from the region. But they are earlier than expected based Caddo in the Saline River Valley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

295

a.

Copyright © 2012. Nebraska Paperback. All rights reserved.

b.

10-5. Excavation unit N199E244, feature 16: a. plan view, and b. unit east profile.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 10-1. Results of radiocarbon dating of Hughes site (3SA11) samples

Sample ID and Provenience Beta-214276 2002-414-191, Feature 8 Beta-214277 2002-414-234, Feature 16 Beta-214275 2002-414-188, Feature 7

Measured and conventional radiocarbon ages before present

Calibrated age intercept

730 ± 40 B.P.; 710 ± 40 B.P.

A.D. 1290

680 ± 40 B.P.; 640 ± 40 B.P.

A.D. 1300

380 ± 40 B.P.; 390 ± 40 B.P.

A.D. 1470

Calibrated age range (1 sigma; 2 sigma) A.D. 1270–1300; A.D. 1260–1310 and A.D. 1370–1380 A.D. 1290–1320 and A.D. 1340–1390; A.D. 1280–1410 A.D. 1450–1510 and A.D. 1600–1620; A.D. 1430–1530 and A.D. 1550–1630

Copyright © 2012. Nebraska Paperback. All rights reserved.

Results reported from Beta Analytic, Inc., and calibrated using Intcal98 database (Stuiver et al. 1998).

on the analysis of the artifacts, which were predominantly Late Caddo period in style. Ceramic sherds and chipped stone arrow points found in the Hughes excavations in 2002 show similarities with site assemblages dating to the late fifteenth to sixteenth centuries in neighboring areas. The feature 7 date seems to reflect the artifact dating better than the earlier date from the features 8 and 16 timbers. This may be because of the reuse of old timbers in that structure, but more likely it reflects the deposition of midden after structure abandonment and the lack of a floor deposit contemporaneous with the lower structure in our limited excavations. Ceramics Small fragments of ceramic vessels were the most common artifact recovered during our excavations at the Hughes site. Based on the detailed analysis of sherds larger than a half inch in size, the majority of the sherds are tempered with crushed mussel shell. Shell was the predominant temper in 88 percent of the 4,217 analyzed sherds with identifiable temper, with grog in 10 percent, grit in 2 percent, and bone in less than 1 percent of the sherds from the 2002 excavations. In unit N199E244, shell predominates as the temper used for vessel sherds found both at and above feature 7 and between features 7 and 16, although grog is Caddo in the Saline River Valley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

297

Copyright © 2012. Nebraska Paperback. All rights reserved.

10-6. Probability curves of 2-sigma calibrated age ranges for radiocarbon assays from the Hughes site (3SA11) compared with other area sites. (Sources: AAS radiocarbon database; Early 1988; Early, ed., 1993, 2000; House 1997; House and Farmer 2001; Reynolds 2007; Trubitt 2002.) (Calibrations calculated with IntCal04 curve using calib 5.0.2 [Reimer et al. 2004; Stuiver and Reimer 1993; Stuiver et al. 2005].)

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

more common in the stratigraphically lower assemblage (grog-tempered sherds make up 3 percent of the sherds at or above feature 7 in depth and 26 percent of the sherds between feature 7 and feature 16). Current work with several colleagues is examining temporal trends in the adoption of shell tempering across the Caddo region. In the Middle Ouachita River drainage, shell tempering appears during the Mid-Ouachita phase (ca. A.D. 1350–1500) and dominates ceramic assemblages during the later Social Hill and Deceiper phases (ca. A.D. 1500–1700; Early, ed., 1993). An excavated Mid-Ouachita phase ceramic assemblage from 3CL593, associated with dates clustering in the early 1400s, had higher percentages of grog tempering: 52 percent of the sherds associated with structure feature 9 there had grog temper, with 81 percent grog-tempered sherds in the lower structure feature 12 assemblage (Reynolds 2007; Trubitt 2009). If temper trends are similar in the Saline River drainage, then the high percentage of shell-tempered sherds suggests this assemblage at the Hughes site was deposited at least a century after the ca. A.D. 1300 dates obtained for the feature 8/ feature 16 wood. Nearly half (43 percent) of all analyzed sherds had decorated exterior surfaces (n = 1,834), and 57 percent had plain surfaces (n = 2,395). Decorated sherds were analyzed using the “collegiate system” developed by Frank Schambach to grapple with combinations of decorations on rims and bodies of vessels in both the Trans-Mississippi South and lower Mississippi valley regions (Rolingson and Schambach 1981; Schambach and Miller 1984). The collegiate system is similar to decorative categories used in analyses of Caddo ceramics elsewhere (see Girard, this volume). At the broadest level, classes are identified: class A includes incised/trailed lines in vertical or diagonal rectilinear patterns; B has incised/trailed lines (sometimes with a minor amount of punctation) in horizontal and typically curvilinear patterns; C is punctated decoration (sometimes with a minor amount of incising/trailing); class D has brushing as the major decoration; E is engraved patterns; F has stamped decorations; G is reserved for slipped and painted pottery; and class H is made up of appliquéd, ridged, and noded decorations. Within each class there are identified rim and body patterns, with distinct variations numbered within each pattern as designs. Whole vessels can be identified to class, pattern, and design, but sherds from the Hughes site excavations were not sorted beyond class and pattern (table 10-2 and fig. 10-7) Caddo in the Saline River Valley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

299

Table 10-2. Ceramic sherd surface decoration by temper Grog

A A rims (Anthony) A bodies (Altus, Antioch)

Shell

A A rims (Agnes, Albertus, Alfred) A bodies (Alma/Antioch)

5 2 2 Total Grog

9 21 10 1

Total Shell Total A Grog

B B rims (Bates) B bodies (Bard)

Shell

B B rims (Barnard, Bates, Butler) B bodies (Babson, Baker, Bard, Bellarmine/Belmont?, Buffalo?)

32 41

1 2 1 Total Grog

4 189 16 15

Copyright © 2012. Nebraska Paperback. All rights reserved.

Total Shell Total B

220 224

Grit Grog Shell

C bodies (Castleton) 1 C rims (Caldwell, Chatham, Cornell) (or bodies Catawba?) 5 C rims (Caldwell, Cambridge, Cameron, Cedar?, Chatham, Chattanooga, Cisco?, Concord, Cornell?, Cranbrook?, Creighton) 387 C bodies (Caney/Coker?, Carver?, Cascade/Claflin?, Casper, Castleton, Clinton?, Curry) 18 Total Shell 405 Total C 411

Grog

D D rims (Dawson?) Total Grog D D rims (Dana, Dawson, Defiance?) D bodies (Dakota/Delaware/Denver, Danbury, Dartmouth, Drake, Drury, Dublin/Dwight?)

Shell

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

1 1 2 491 34 48 Total Shell Total D

573 575

Grog

Shell

Shell

E E rims (Elaine) E bodies (Eccles?, Emerson, Eugene?, Evansville/Evergreen?) Total Grog E E rims (Edna/Epic/Ezra?, Edward, Elaine, Enos?, Erie) E bodies (Eccles?, Edgecliffe/Eugene/Evansville/Evergreen) Total Shell Total E F F rims (Ferris, Franklin?, Fresno?) F bodies (Farmington?)

3 1 4 8 27 8 8 43 51

2 6 1 Total F

Shell

G

9

2 Total G

2

Grit Grog

H bodies (Harris?) H bodies (Harris?)

3 1

Shell

H H rims (Harding?) H bodies (Hampton?)

2 4 3 Total Shell Total H

9 13

Copyright © 2012. Nebraska Paperback. All rights reserved.

Note: Table does not include 492 sherds identified as A or B, one as A or C, two as B or C, and 47 as unidentified decorated. Sherds that show both rim and body decoration are counted twice in this table.

because of the difficulty in identifying designs on small sherds. The system is somewhat cumbersome with its profusion of pattern names. However, in a borderlands region, it has the advantage of description without presuming cultural affiliation by a quick assignment of ceramic types and varieties typical of either the lower Mississippi valley or Caddo areas. Relatively few sherds were identified as class A at the Hughes site. A larger proportion of class A sherds was tempered with grog (22 percent) compared to other classes. The most common class A patterns were Agnes and/or Alfred, with multiple rows or bands of diagonal incisions. While incised line decoration appears on many of the Hughes Caddo in the Saline River Valley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

301

10-7. Common ceramic patterns found in the Hughes site sherd assemblage.

Copyright © 2012. Nebraska Paperback. All rights reserved.

sherds, most were too small to identify as specifically class A or class B. Where enough of the design is visible, wide “trailed” curvilinear lines were often identified (class B), including body sherds with volutes or concentric circles (Babson and Baker patterns), with or without nodes. These are patterns commonly found on utilitarian Foster Trailed-Incised jars (e.g., Early 1988; Kelley 1997; Schambach and Miller 1984). Few fine ware sherds in the assemblage appeared to have trailed curvilinear decorations (e.g., the Belhaven pattern, as well as on some Bellarmine pattern sherds) that are seen on Keno Trailed bottles. Notably rare was the Bard body pattern (on two sherds tempered with shell and one with grog), a pattern associated with Military Road Incised jars in the middle Ouachita River valley (Early, ed., 1993). Horizontal lines on rim sherds (the Barnard rim pattern with broad trailed lines, the Bates pattern with well-done lines on firm paste, and the Butler pattern with irregular lines on wet paste) were identified on several rim sherds but may occur on several different types (Rolingson and Schambach 1981). Punctated decoration (class C) was fairly common (31 percent of the decorated sherds) in the Hughes sherd assemblage. Most frequent were pieces of shell-tempered pottery with rows of diagonal punctates formed with a tool (Chattanooga rim pattern). Some were narrower, some were deeper and wider, but most appeared to be tool formed rather than made with fingernails (Concord pattern rims and Coker pattern bodies). There were also a minority of the punctated class C sherds with vertical or horizontal tool punctates (Creighton or Cranbrook pattern rims). Agnes or 302

Trubitt

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Alfred or Chattanooga rims with Babson and Baker pattern bodies are commonly found on Foster Trailed-Incised tall rim jars (Kelley 1997; Schambach and Miller 1984). At the Cedar Grove site (3LA97), Schambach and Miller (1984) proposed a seriation of Foster Trailed-Incised varieties based on a change from incised to punctated rim decorations and from circular to volute trailed body decorations (see also Kelley, chap. 14, this volume). There, jars with incised Alfred rims and trailed Baker bodies are part of the Late Caddo Belcher phase component while punctated Chattanooga rims with trailed Babson bodies are a marker for the Historic Caddo Chakanina phase component. At Hughes, there were some Alfred rims and Baker bodies, but most of the trailed body sherds could not be identified as being specifically Baker or Babson patterns, and the Chattanooga rim pattern predominated (although different sorting criteria may have been used to distinguish incisions from punctations in the two analyses). The Chattanooga::Baker combination found at the Hughes site does not match any of the Cedar Grove varieties of Foster Trailed-Incised. There is a temporal trend at Hughes from wide-spaced to close-spaced rows or bands of diagonal punctates; widespaced bands of punctates are more often seen on sherds in the lower zones of unit N199E244 (both between features 2/3 and feature 7, and between feature 7 and feature 16), while Chattanooga sherds at or above features 2/3 more frequently had closely spaced rows of diagonal punctates (fig. 10-8). Several shell-tempered sherds were found that have a long narrow punctate decoration identified as the Chatham rim pattern (similar to designs 6–9). This decoration is seen on rims and bodies of some Cowhide Stamped vessels (Suhm and Jelks 1962:plate 15f, l). There are also several rims and body sherds of a distinctive shell-tempered vessel decorated with sharp vertical and diagonal trailed lines and triangular punctates (Chatham or Cisco rim patterns, and Cascade or Claflin body patterns). This particular vessel does not fall readily into a type but shows stylistic similarities with lower Mississippi valley ceramic types Barton Incised and Owens Punctated (Phillips 1970). Many of the Hughes site decorated sherds (43 percent) are class D specimens with brushed decoration, defined as lines impressed or brushed into a wet clay, spaced close together so that lines often touch or overlap (Rolingson and Schambach 1981:144). Two class D sherds are grog-tempered, but the majority are shell-tempered. Most are too Caddo in the Saline River Valley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

303

10-8. Class C rim patterns, Class B body patterns: top, Chattanooga close and widely spaced rows of punctations; middle, Chattanooga::Baker and Chattanooga::B rim/body sherds; lower, Baker/Babson body sherds (aas/hsu D4763).

Copyright © 2012. Nebraska Paperback. All rights reserved.

small to orient and are tabulated no further than to class. In the Red River valley, brushed pottery is found in the Middle and Late Caddo periods, but shell tempering comes in late in the sequence (see Girard, chap. 9, and Kelley, chap. 14, this volume). Identifiable patterns include Dana, Dawson, and perhaps Defiance rims (with horizontal brushing, diagonal brushing, and brushing with over-incising), and Danbury, Dartmouth, and Dakota/Delaware/Denver (with plain brushing, rectilinear brushing, and brushing with over-incising). Class D sherds and Dana, Dartmouth, and Danbury patterns may be from grog-tempered types such as Bossier Brushed, Plaquemine Brushed, or Pease BrushedIncised, or shell-tempered types such as Grace Brushed, Caney Punctated, or Karnack Brushed-Incised (Early, ed., 1993; Kelley 1997; Kidder 1986; Phillips 1970:153; Rolingson and Schambach 1981; Suhm and Jelks 1962; Webb 1963). Rim/body combination sherds include one with Chattanooga::Danbury patterns and one with fingernail punctates and brushing (Concord::Danbury/Drake patterns or a design within the Dana or Dwight patterns). 304

Trubitt

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

A number of the class D body sherds were identified as brushing in wide or narrow zones (Drake or Drury patterns); these are patterns seen on some Cowhide Stamped vessels (Rolingson and Schambach 1981; Suhm and Jelks 1962:plate 15b–c, e, m), and more recently included in the Mound Tract Incised and Brushed type by Kidder (1986; see also Kelley 1997). Three rim/body fragments have dentate-stamped rims with zoned brushing on the bodies (the Ferris/Fresno::Drake/Drury patterns) that also would fit into the Mound Tract type (Kidder 1992:fig. 8d). There is an example of a Barnard::Drake/Drury combination sherd in the Hughes assemblage. Engraved sherds (class E) are more commonly tempered with grog (16 percent) than are the other decorated pottery classes at the Hughes site. Most engraved sherds are small and difficult to identify to pattern. A few identifiable patterns, such as the Elaine, Edward, Ezra, Erie, and Eccles patterns (fig. 10-9) are associated with the Friendship Engraved, Cook Engraved, and Hardman Engraved bowl types in the middle Ouachita River valley (Early, ed., 1993; Lafferty et al. 2000; see Early, chap. 2, this volume). Others have parts of circles or swirls like in the Edgecliffe, Eugene, Evansville, and Evergreen patterns that are found on bottle types such as Bailey Engraved and Taylor Engraved in the Caddo area (Kelley 1997; Suhm and Jelks 1962). A rim fragment with engraved lines may be a nose from a Bailey Engraved human effigy bottle. Lacking here is evidence of post–sixteenth-century engraved types such as Hodges Engraved (in Deceiper phase contexts, Early, ed., 1993; in Caney Bayou phase contexts, Jeter and Early 1999; and in Belcher phase contexts, Kelley, chap. 14, this volume). Relatively few engraved sherds were found (less than 4 percent of the identified decorated sherds), reinforcing the observation that utilitarian vessels dominate the midden deposits adjacent to the mound. Stamped class F sherds include several with plain or zoned dentate stamping (Ferris or Fresno rim patterns and Farmington body pattern) and one with rocker stamping (the Franklin pattern?). The dentate stamped patterns are seen on Cowhide Stamped and Mound Tract Brushed-Incised (with zoned brushing) type vessels. Class G sherds are rare and are limited to two red-slipped and shell-tempered sherds that may be fragments of Old Town Red (Phillips 1970:144–147), a ceramic type that suggests Mississippi River valley connections. Several of the class H sherds in the Hughes site assemblage also suggest stylistic or cultural links to the east. A couple of rim sherds with appliquéd fillets Caddo in the Saline River Valley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

305

10-9. Class E sherds: top, Enos rim, Ezra rim, E effigy/tab rim; lower, Elaine rim pattern (grog), Edward rim pattern, Edgecliffe/Eugene, Evansville/Evergreen body patterns (aas/hsu D4767).

Copyright © 2012. Nebraska Paperback. All rights reserved.

or notched strips at or just below the lip (Harding or Cornell pattern?) resemble bowl forms found on Bell Plain and sometimes on Mississippi Plain (Moore et al. 2006; Phillips 1970:figs. 14i, 98b, and 99b). Several others had brushing with appliquéd notched strips (Hampton?) that are more than likely the Caddo type Pease Brushed-Incised. A relatively high proportion of the class H sherds are tempered with grit (23 percent), and have appliquéd elongated nodes (Harris pattern) similar to several Ouachita Plain sherds from the middle Ouachita River valley (Schambach 1998:91 and fig. 66l-n). In the type-variety system, the most common ceramic type identified at the Hughes site is Foster Trailed-Incised, and other utilitarian types such as Mound Tract Incised and Brushed/Cowhide Stamped, Grace Brushed, Pease Brushed-Incised, Caney Punctated, and Military Road Incised are present as well. Engraved wares, including Friendship Engraved, Cook Engraved, and perhaps Hardman Engraved and Bailey Engraved, are rare in the 2002 ceramic assemblage. Two sites recently excavated in central Arkansas (3PU111 and 3JE285 in Pulaski and Jefferson Counties) have ceramic assemblages that include Foster Trailed-Incised and shell-tempered brushed types and radiocarbon dates that cluster with those from Hughes. Kuykendall Brake (3PU111) is an Arkansas River valley site, east of Hughes (see fig. 10-1), where a burned rectangular structure covered with mound fill created 306

Trubitt

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

a low mound (House 1997). The excavated structure had a diverse array of 34 ceramic vessels recovered from the floor of the burned structure, including Caddo types with 9 to 13 Foster Trailed-Incised jars, a Belcher Engraved bottle, and a Hardman Engraved bowl, in addition to shell-tempered brushed jars and Mississippi Plain and Bell Plain bowls (at least one with appliquéd notched fillet at rim) and bottles. A burial feature with fragmentary remains of 17 people underlay the structure floor. Six radiocarbon dates have been run on charred timbers from the structure; the pooled assay has a calibrated intercept of A.D. 1483 and a 1-sigma calibrated age range of A.D. 1469–1617 (House 1997). The burned Kuykendall Brake structure is contemporaneous with the Hughes site feature 7 (see fig. 10-6), and the clear use of Foster TrailedIncised jars as utilitarian wares indicates stylistic similarities as well. The Eastwood Bayou site (3JE285) in Jefferson County, Arkansas, is a small site with a low mound that was investigated as part of a site testing project at the Pine Bluff Arsenal (House and Farmer 2001). One test unit identified a low mound that had a fragmentary Mound Tract Incised and Brushed vessel at the base of the mound deposits. The other test unit revealed a portion of a burned structure with mound fill on top of it. Pottery from the level associated with the structure included three Foster Trailed-Incised sherds. House and Farmer (2001:13–15) argue for a post– A.D. 1500 date for the pottery assemblage, even though the calibrated radiocarbon dates from charred wood posts fall in the late A.D. 1200s to early 1400s. The earlier dates at Eastwood Bayou overlap with the Hughes feature 8 and feature 16 radiocarbon assays (see table 10-1), and the ceramics associated with these features also seem stylistically later than the dates would otherwise indicate. Chipped Stone Lithic artifacts from the Hughes site include novaculite tools (points, fragments of bifaces, and utilized flakes) and chipping debris. Over 70 arrow points or point fragments have been catalogued. Many of these (fig. 10-10) are Maud points, small triangular points with concave or ushaped bases and sometimes side notches that are found on Late Caddo period sites in southwest Arkansas, east Texas, and northern Louisiana (Early 1988, Early, ed., 2000; Kelley 1997; Kidder 1992; Lafferty et al. 2000; Perttula 1992; Trubowitz, ed., 1984). In the Arkadelphia area, Maud arrow points as well as Foster Trailed-Incised jars are characteristic of the Social Hill phase, estimated to date from A.D. 1500–1650, Caddo in the Saline River Valley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

307

Copyright © 2012. Nebraska Paperback. All rights reserved.

although there are no radiocarbon dates to corroborate this (Early 2002). A corrected and uncalibrated radiocarbon assay of 430 ± 60 years B.P. from the Hardman site (3CL418, Beta-33546, Early, ed., 1993:47) was associated with a burial feature with engraved ceramics (Cook, Hardman, Friendship, Blakely, and Belcher Engraved types) and Bassett arrow points, assigned to the late Mid-Ouachita (or Social Hill, Lafferty et al. 2000:37) phase. Some basal notched arrow points with small stems similar to the Bassett or Perdiz types (Suhm and Jelks 1962) are also represented in the collection from Hughes. Bassett points are part of Belcher phase lithic assemblages (Webb 1959; see Kelley, chapter 14, this volume). For post-1430 Titus phase assemblages in east Texas (see chap. 13, this volume), a sequence of point types from Perdiz to Bassett to Maud, and then to Talco, has been well demonstrated (Perttula 1992:243–249; Turner 1978). At the Ouachita Mountains site of Standridge (3MN53), Early (1988: 107–109) defined new varieties of the Maud type based on a cluster of arrow points included as feature 9 grave goods. The feature also included pots typed as Keno Trailed, var. Curtis, Poteau Plain, Woodward Plain, Ashdown Engraved, and Foster Trailed-Incised. Maud, var. Maud refers to the triangular points with serrated blade edges and deep concave bases, and var. Hopper are points that also have small side notches on the blades. Early (1988:159) estimated the Standridge site occupation took place between the mid-1300s to early 1400s, although most of the radiocarbon assays gave aberrant results and only one (from feature 12, with a corrected age of A.D. 1326 ± 66) is included in figure 10-6 for comparison; archaeomagnetic dates from Standridge features ranged from A.D. 1350 to 1450. Both Maud and Washita (Washita is characterized by side notches and straight or slightly concave bases) points were associated with a burned structure from Winding Stair (3MN496) that was dated to around A.D. 1470 to 1500 (the two radiocarbon assays have measured ages about 100 years apart, but the 2-sigma ranges overlap; Early, ed., 2000:82). At the Hughes site, all the Maud points from unit N199E244 were found at or above feature 7, and stemmed, side-notched, and basalnotched forms were scattered in the unit. Highlighted by the comparisons of radiocarbon dates in figure 10-6 is a clustering of the probability curves into earlier (ca. A.D. 1250–1450) and later (ca. A.D. 1450–1650) groups. The later group includes dates from the Winding Stair, Hardman, Hughes, and Kuykendall Brake sites, 308

Trubitt

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

10-10. Arrow points from the Hughes site: top, Bassett, middle, Maud var. Maud, lower, Maud var. Hopper. (aas/hsu D4782.)

associated with artifact assemblages that have ceramic types in common such as Foster Trailed-Incised, shell-tempered brushed utilitarian wares, engraved types like Hardman Engraved, Blakely Engraved, and Belcher Engraved, and Bassett and Maud points. The earlier group includes dates from Eastwood Bayou, Standridge, and Hughes (as well as two of the Peeler Bend canoe dates); because the decorated ceramics and arrow point types from these sites also include Foster Trailed-Incised and Mound Tract Incised and Brushed jar fragments and Maud points, they have been interpreted as having stylistically later ceramics and Caddo in the Saline River Valley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

309

arrow points than their associated dates. The two radiocarbon assays from burned structures in a mound (3CL593) salvaged near Caddo Valley have been included in figure 10-6 as well. While the ceramic assemblage from the pre-mound midden related to an East phase occupation, the later structure at 3CL593 has clear Mid-Ouachita phase characteristics. There, grog tempering predominated, decoration included trailing, incising, punctating, and engraving, with brushing being rare, and Military Road Incised and Friendship Engraved were the most commonly recognized types (Brock and Reynolds 2007; Reynolds 2007; Trubitt 2009). A Regional View

Copyright © 2012. Nebraska Paperback. All rights reserved.

I started this chapter by pointing out that the Saline River valley lay at the eastern edge of the Caddo area. We began the research at the Hughes site with expectations of evidence in the material culture for interaction between Caddo and Mississippian neighbors. From the perspective of the middle Ouachita River valley to the west, the Hughes ceramic and lithic assemblages look Caddo and would fit comfortably as a Social Hill phase site (ca. A.D. 1500–1600) if it were in the Arkadelphia area. The late A.D. 1400s radiocarbon date associated with the Hughes feature 7 burned feature seems to reflect the temporal and stylistic character of the artifact assemblage better than the earlier dates obtained on the lower burned structure (features 8/16). The Hughes artifact suite also show similarities with sites in the Ouachita Mountains to the west. While there are overlapping types, the Hughes site ceramics are stylistically earlier than the Cedar Grove assemblage in the Red River valley (Schambach and Miller 1984; Kelley, this volume). There are also stylistic similarities between the Hughes assemblage and sites in the Arkansas River valley to the east, specifically with the unusual Kuykendall Brake site. While some brushed shell-tempered utilitarian wares are present in the middle Ouachita River valley (brushing occurs on some Military Road Incised, Caney Punctated, and Karnack Brushed-Incised vessels), a strong presence of brushed shell-tempered sherds at Hughes also suggests connections with sites in southeastern Arkansas and northern Louisiana such as Gordon (3AS152) and Jordan (16MO1) (Kidder 1986, 1992; White 1987). Ceramic characteristics such as appliquéd notched fillets on rims of otherwise plain vessels are unusual for the Caddo area but are often seen in ceramic assemblages in the Arkansas River valley, the Mississippi River valley, and further east. 310

Trubitt

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

The stone tools at Hughes were predominantly made of novaculite, and bedrock sources of this tool stone lie to the north and west in the Ouachita Mountains, although there is some noticeable use of novaculite gravels or cobbles as raw material at this site, so residents there may have used a local river gravel source as well. Obtaining novaculite from the eastern Ouachita Mountains and moving/exchanging it down the Ouachita and Saline Rivers has deep roots in the prehistory of this region. The arrow points found at Hughes—especially the Maud type— show similarities with arrow points found in the broader Caddo area, but I expect further documentation of artifact collections from the Saline River valley will show Maud varieties to have been common local arrow points there during the fifteenth and sixteenth centuries. The interactions and connections may have involved not only the movement of goods but travel by people between communities along the Saline River during the fifteenth and sixteen centuries. Water transport was important, as is highlighted by the finds of two wooden dugout canoes within several miles of each other on the Saline River, one of which was made from a tree felled in the thirteenth century (Trubitt 2002). Trails such as the one that later became the Old Military Road were important for transportation in the Caddo area as well. Overland travel may have brought the de Soto expedition close to the Hughes site during the sixteenth century (Hudson 1993), but our project in 2002 found no artifacts linked to this early Spanish presence. I think it is clear that the Hughes site was a Caddo community. Not only do we see these connections in the ceramic and lithic artifact styles, but they can be seen as well in the character of the features. The sequence of burned features—presumably parts of structures—at Hughes recalls the burning and burying of buildings at the 3CL593 mound near Caddo Valley (see Reynolds 2007) and at other documented Caddo sites in the area (Trubitt 2009). The features from the Hughes site and 3CL593 are dissimilar in that the fills covering the burned structures at 3CL593 were nearly sterile, but midden soils with lots of animal bones are associated with the Hughes structural features. Perhaps a better comparison for the activities adjacent to the Hughes mound come from the “cookhouse” structure filled with ceramic pots adjacent to the large mound at the Tom Jones site (3HE40; Schambach 2003) or perhaps the evidence of feasting debris in a structure adjacent to the Titus phase mound at the Whelan site (see chap. 13, this volume). Caddo in the Saline River Valley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

311

There are still gaps to be filled with further survey and excavation in Arkansas’s Saline River valley. The multi-stage mound at Hughes seems isolated and unique, but the two-stage mound form is commonly seen in the Caddo area of southwest Arkansas, especially in the Little Missouri River region. There are several other sites with mounds (3SA17, 3SA19, and 3SA73) recorded in the Saline River drainage around the Hughes site that are known only from brief site visits. There are also numerous non-mounded sites that may have archaeological evidence of contemporaneous smaller communities. It remains to be seen whether the paucity of mound sites further south in the Saline River valley (Lockhart 2007, see chap. 11, this volume) is due to a lack of survey coverage or a cultural boundary. More work must be done to understand the Hughes site Caddo community and its relationships with peoples living in the Saline River valley and the wider region, but the 2002 field work provides us with a baseline to characterize cultural affiliations along this “boundary” area. We have new radiocarbon dates and described artifact assemblages for excavated features that give us a glimpse of near-mound activities at one Caddo site in this region. Acknowledgments

Copyright © 2012. Nebraska Paperback. All rights reserved.

Excavations at the Hughes site (3SA11) were undertaken with students in the 2002 Henderson State University/University of Arkansas Archeological Field School (Laura Deneke, Jodi Dickerson, Jason Eads, Micah Gould, Jim McKenzie, Justin Pathe Miller, Anthony “Clay” Newton, Chris Page, Jessi Robkin, James Walden, and Brendon Ward). Arkansas Archeological Survey research assistants Kate McLaurin Wright and Matthew Reynolds aided me with field/lab supervision and with the preliminary analysis of ceramic sherds. Funding for radiocarbon dating samples from the Hughes site came from a Henderson State University Faculty Research Committee grant in 2006, and is greatly appreciated. I thank Bill Davenport for allowing access to the Hughes site for our research and for protecting this important prehistoric and cultural site. This work could not have been done without the assistance of the late Dr. Quin Baber, who served as the site steward for many years and who introduced me to the history of the Benton area. Finally I appreciate conversations with Tim Perttula, John House, Ann Early, Cricket Kelly, and Marvin Jeter about aspects of this research during the analysis and writing phases. 312

Trubitt

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

11. Spatial Patterns of Caddo Mound Sites in the West Gulf Coastal Plain of Arkansas jami j. lockhart

Copyright © 2012. Nebraska Paperback. All rights reserved.

Introduction During three weeks of each summer in 2001–2003, the Arkansas Game and Fish Commission’s Grandview Prairie Wildlife Management Area hosted the Arkansas Archeological Society’s Annual Field Training Program at a Caddo mound site called Tom Jones (3HE40). The annual program provides archaeological education and professionally guided excavation opportunities for amateurs. Due to amassed and organized resources, the concerted time and effort put forth has the potential to produce large amounts of site-specific information (Schambach 2003; Lockhart 2007). This chapter addresses two questions: what is the spatial pattern represented by the entire class of Caddo culture mound sites in Arkansas’s West Gulf Coastal Plain, and how does the Tom Jones site (3HE40) fit into that pattern? Although the general distribution of Caddo cultural and archaeological phenomena within Arkansas has been generally described, bounded, and summarized elsewhere (Schambach and Early 1982; Schambach 1990; Perttula 1992; Early 2000, 2004), specific relationships to physiographic characteristics have not been quantified using an entire class of sites. By extension, the results of such quantification have not yet been interpreted in terms of cultural decisions that may have underlain specific geographic distributions. This research has produced a set of simple indices of relatedness between specific cultural and natural features as a means of identifying and locating shared elements of Caddo cultural landscapes in prehistory. Statistical inferences are drawn from geo-cultural correlations, combined, and mapped as a step toward answering those deceptively simple questions. 313

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Cultural Framework As early as A.D. 800, people living within the West Gulf Coastal Plain physiographic region developed into complex societies linked by similarities in settlement pattern, adaptation, and cultural expression, as well as economic, religious, and political organization (Sabo 1998). These people are known to us historically and collectively as the Caddo. As an approximate temporal framework for archaeological remains, prehistoric Caddo cultures can be identified between ca. A.D. 800 and A.D. 1680 within the larger spatial bounds of the trans-Mississippi South (Schambach 1998), which encompasses forested lands west of the Mississippi River between the Southern Plains and the lower Mississippi valley, and can be broadly characterized as a southeastern U.S. woodland environment. According to ethnographic and archaeological evidence, the Caddo were mound builders, sedentary agriculturalists, and accomplished artisans and traders who were governed by complex sociopolitical and religious hierarchies (Swanton 1942; Griffith 1954; Wyckoff and Baugh 1980; Sabo 1998; see chap. 1, this volume). Theoretical Framework In A State Plan for the Conservation of Archeological Resources in Arkansas, Schambach and Early (1982:SW6) present three statements as fundamental components of modern archaeological research:

Copyright © 2012. Nebraska Paperback. All rights reserved.

1. Societies inhabit and adapt to particular and definable geographic areas. 2. The distribution of material remains associated with particular cultural systems delimits the boundaries of these extinct systems. 3. The pattern of distribution of cultural remains provides evidence of the mode of articulation between culture and environment. Guided by those statements, this study attempts to define and delimit the geographic areas and environments of prehistoric Caddo culture in Arkansas’s West Gulf Coastal Plain based on the most evident of extant material remains, which are composed of 114 mound sites. From approximately A.D. 800 to the present, civil and ceremonial life of Caddo communities has been organized around “centers” of some type and mound centers in particular until the eighteenth century (Schambach 1996). In this regard, the distribution of Caddo mounds will be used to 314

Lockhart

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

discover and examine expressions and interrelationships of elements from past cultural landscapes. From a statistically defensible standpoint, mound sites provide the most representative site type sample due to their visual prominence and enhanced recognition as archaeological sites. As compared with buried sites, mound sites are viewed as less affected by sampling biases attributable to narrowly focused survey designs and random/uneven site discovery. The primary aim in the decision to limit the study area to the West Gulf Coastal Plain was to examine spatial patterns and make comparisons between Tom Jones (3HE40) and 113 other proximate Caddo mound sites located in the same physiographic region composed of generally similar conditions. In terms of cohesion, the physical landscape of the West Gulf Coastal Plain is more homogenous within the region than between the other physiographic regions of Arkansas. Although there are at least 15 mound sites of Caddo cultural affiliation in the Ouachita Mountains, distance, geographic isolation, and differences in terrain, soils, geology, micro-climate, potential natural vegetation, etc., between the mountains and the West Gulf Coastal Plain are sufficient to have potentially led to altered adaptations and lifeways even among culturally related Caddo peoples.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Technological Framework Technological advances and the availability of high-resolution data allow researchers from different academic disciplines to explore interrelationships between the physical and cultural elements of a landscape at spatial and temporal resolutions that were previously prohibitive in terms of technological capability, time, and monetary cost. Natural properties of the physical environment as they would have related to settlement patterns of prehistoric cultures can be quantified, mapped, modeled, and analyzed as they relate to human lifeways through time. The present study describes a research approach for recognizing and examining integrated natural and cultural components of the prehistoric Caddo landscapes using digital tools and data sets that include a statewide database of archaeological sites, Geographic Information System (gis) technology, ethnographic and archaeological research, and exploratory data analysis techniques. Spatial Patterns of Mound Sites

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

315

Comprising a very useful set of tools for analyzing and synthesizing archaeological spatial information, gis technology (Farley et al. 1990) is now used by both the natural and social sciences to collect, store, retrieve, transform, and display spatial data for problem-solving applications (Burrough 1986). Kvamme (1999:154) has described gis as an “information visualization engine” that can be used to derive new information from existing primary data. Interoperability between a gis and attribute data contained in a relational database (rdbms) is now more common, and information can be subset by a variety of nonspatial attributes for very specific queries that increase the potential of multivariate analyses. The Data

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Tom Jones site is located in Hempstead County in southwestern Arkansas, some 20 miles northwest of the town of Hope. It lies within the West Gulf Coastal Plain physiographic subdivision, which encompasses approximately 24 percent of Arkansas and is bounded on the north by the Ouachita Mountains and on the east by the Mississippi Alluvial Valley. To the south and west, portions of the West Gulf Coastal Plain extend into Louisiana, Texas, and Oklahoma. Tom Jones is one of 795 Caddo sites of good cultural affiliation and locational reliability recorded in the Arkansas Archeological Survey’s Automated Management of Archeological Site Data in Arkansas (amasda) database (Hilliard and Riggs 2000). Of those 795 Caddo sites in Arkansas, 588 (74 percent) are located in the West Gulf Coastal Plain. Within that larger group of Caddo sites of all types, Tom Jones is 1 of 131 reliably encoded Caddo mound sites in Arkansas, of which 114 (87 percent) are located in the West Gulf Coastal Plain. Other data sets used in this research were obtained and derived from a variety of sources. gis data layers that have been compiled are elevation, hydrography, soils, geology, land use/land cover, modern and historic vegetation, watershed boundaries, physiographic subdivisions, slope, aspect, distance-to-water, space-based and aerial remote sensing, historic tribal boundaries, and other derived small-scale and large-scale themes (table 11-1). All of the gis data are georeferenced using the North American Datum 1927 (nad27), Universal Transverse Mercator (utm) Zone 15, which match the projection and datum of the archaeological site data in amasda. All data layers used in this research were compiled 316

Lockhart

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Scale/resolution/accuracy Plotted to the nearest 10 m 30 m horizontal, 1 m vertical 30 m horizontal, by degrees 30 m horizontal, by degrees 1:250,000-scale source 1:24,000-source 100 m horizontal 1:24,000-scale source 1:500,000-scale source 1:24,000-scale source 30 m horizontal 30 m horizontal Discrete location from survey description 1:250,000-scale 8-digit hydrologic basins 1:24,000-scale source 1:250,000-scale

GIS data layers

Mound Sites Elevation

Slope Aspect Hydrography Hydrography Soils

Soils Geology Geology Land use/land cover

Modern vegetation

Historic vegetation (trees) Watershed boundaries Watershed boundaries Physiographic subdivision

Table 11-1. Geographic Information System data layers

Copyright © 2012. Nebraska Paperback. All rights reserved.

amasda United States Geological Survey (usgs) National Elevation Database (ned) Derived from elevation (ned) Derived from elevation (ned) U.S. Soil Conservation Service (scs) usgs Digital Line Graphs (dlg) Natural Resource Conservation Service (nrcs) State Soil Geographic Database (statsgo) nrcs (scs) Arkansas Geological Commission usgs National Land Cover Data (nlcd) Landsat Thematic Mapper unsupervised classification ua Center for Advanced Spatial Technologies Landsat Thematic Mapper unsupervised classification General Land Office Surveys nrcs usgs scs

Source

or constructed using Arcgis (esri, Inc.), idrisi (Clark University), and grass (U.S. Army Corps of Engineers) software. The selection of data to be used in cultural and environmental modeling was not made randomly, nor did it constitute the first step in analyses of correlations. Instead, the process of data selection is conditioned by what Willey (1968:209) has termed “interpretive hypotheses.” The present research adheres to the idea that prehistoric Caddo people selected locations for mound construction because each landscape characteristic was of some importance to them (Story et al. 1990:5). Examples would include hydrography as a nearby water source and as transportation and communication routes; specific watersheds as recognizable demarcations of social/cultural inclusion and exclusion; local elevation for site drainage, intervisibility, and as a signal of elevated status within the larger social organization; slope as a practicality of stability; soils for constructed elements of the landscape, and as an essential part of a largely agrarian, sedentary life way; geologic features as a lithic resource and as physical touchstones of the conceptualized and ideational landscape (see Brooks, this volume); and open lands near forests of varying composition for diversity of necessary resources, such as timber for construction and grass for roof thatch. Environmental Similarity

Copyright © 2012. Nebraska Paperback. All rights reserved.

Although the clear majority of Caddo mound sites in Arkansas occur in the West Gulf Coastal Plain, site counts alone do not provide a measure of the statistical significance of the distribution. Table 11-2 quantifies the data using Chi-square (χ2) analysis to determine the similarity of the number of actual occurrences being investigated to the expected occurrences based on the amount of area comprising each physiographic subdivision and a random distribution of the total number of sites. Each category in table 11-2 is treated as an independent piece of information, so degrees of freedom equal one in each instance. The total number of Caddo mounds is known, so the number of sites not correlated with each category of a map theme is determined by the number that are correlated. Using the χ2 statistic within a gis, the proportion of areal coverage for each category of a spatial data theme is calculated. In the example shown in table 11-2, the data theme is physiographic regions of Arkansas. Then the statistically expected number of mound sites for each 318

Lockhart

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 11-2. Caddo mound coincidence with Arkansas physiography using χ2 statistic Physiographic region

Copyright © 2012. Nebraska Paperback. All rights reserved.

West Gulf Coastal Plain Central Ouachita Mountains Athens Plateau–Ouachitas Springfield Plateau–Ozarks Fourche Mountains–Ouachitas Salem Plateau–Ozarks Mississippi Alluvial Plain Crowley’s Ridge Boston Mountains–Ozarks Arkansas River Valley

% of state

Expected sites

Actual sites

χ2

24.6 3.9 2.4 7.6 7.1 7.9 27.5 1.4 7.7 9.8

32.3 5.1 3.2 9.9 9.3 10.4 36.1 1.9 10.1 12.8

114 11 3 2 1 0 0 0 0 0

207.227 6.936 0.012 6.323 7.447 10.403 36.059 1.898 10.078 12.790

physiographic region is calculated based on a hypothetically random statewide distribution of the total number of sites and the areal coverage for each physiographic region. A χ2 value is generated by comparing the actual number of sites in each physiographic region to the number of sites that would be expected given a hypothetically random distribution. Thus the χ2 test serves as a simple index that is useful in recognizing differences between the distribution of actual sites and random point distributions (Brandt et al. 1992). In this example, the χ2 scores are used to assess the probability that the distribution of Caddo mound sites within each physiographic region in Arkansas could be the result of mere chance (Hammond and McCullagh 1982). Generally speaking, the larger the χ2 value, the smaller the probability that the distribution is the result of mere chance. Specifically, in table 11-2 the West Gulf Coastal Plain receives the greatest χ2 value for having far more Caddo mound sites than would have been statistically expected. Conversely, the Mississippi Alluvial Plain also receives a relatively high χ2 score, but for having far fewer sites than would have been expected to occur given a normal distribution. Using the χ2 statistical technique, other elements of the Caddo natural world can be examined in order to suggest interrelationships between environmental characteristics and Caddo mound sites, first by quantifying the extent to which Caddo mound sites in the West Gulf Coastal Plain of Arkansas share individual environmental characteristics, and then by mapping the coincidences of those defining characteristics and extrapolating them across the entirety of the study area. The Spatial Patterns of Mound Sites

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

319

purpose of the resulting environmental similarity model is to quantitatively characterize the placement of the majority of Caddo mound sites in the West Gulf Coastal Plain of Arkansas, but also to identify any anomalous patterns in their distribution. Anomalous patterns will be recognizable by the presence of Caddo mound sites where none would be statistically expected and as large contiguous areas that are environmentally similar to the majority but that contain no Caddo mound sites. Caddo Mounds and Watersheds

Copyright © 2012. Nebraska Paperback. All rights reserved.

For more rigorous exploratory analysis, the spatial distribution of Caddo-encoded mound sites in Arkansas’s West Gulf Coastal Plain can be organized into more meaningful structures. In elemental terms, spatial distributions can be thought of as belonging to one of three types: uniform, random, and clustered (Hammond and McCullagh 1982). Caddo mound sites in Arkansas are not uniformly distributed among physiographic regions, stream courses, and landforms. Moreover, if we accept that humans do not select locations randomly in total disregard of their external environment and internal cognition, then human settlement patterns must be categorized as clustered distributions. Previous conceptualizations of Caddo culture within Arkansas by Early (Schambach and Early 1982:SW99–128; Early 2004:566) imply that proximity to specific streams may be a unifying cultural characteristic. To begin the more focused examination of environmental elements of the prehistoric Caddo world, watersheds are used to delimit natural areas that closely resemble the statistical clusters. One method of exploring the differential spatial distribution of the Caddo mound sites in Arkansas’s West Gulf Coastal Plain is to use the statistical technique cluster analysis (Tryon 1939; Kaufman and Rousseeuw 1990). Cluster analysis is defined as “the partitioning of data into meaningful subgroups, when the number of subgroups and other information about their composition may be unknown” (Fraley and Raftery 1998:579). Furthermore, cluster analysis describes a collection of algorithms well suited to data mining, which is defined as “the analysis of (often large) observational data sets to find unsuspected relationships and to summarize the data in novel ways that are both understandable and useful to the data owner” (Hand et al. 2001:1). There are several statistical methods that can be used for clustering. For this application, S-Plus statistical software was used with the clustering method known as Partitioning Around Medoids (pam). In short, 320

Lockhart

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

11-1. Statistical clusters of mound sites (left) compared with similar clusters grouped by watershed boundaries.

the pam method provides verification of statistically structured data, calculates the statistically optimum number of clusters for a data set, and minimizes the average dissimilarity of all objects within each cluster (Rousseeuw 1987). The spatial data used to quantify the correlations between Caddo mound sites and watersheds are the Natural Resource Conservation Service’s eight-digit hydrologic basins, which are geographic areas representing all or part of a surface drainage basin, a combination of basins, or a distinct hydrologic feature. As shown in figure 11-1, correspondence between the eight statistically optimized clusters of Caddo mound sites (left) and similar clusters bounded by natural watersheds (right) recommend watersheds as a viable component of environmental similarity. Watersheds containing more than the statistically expected number of sites within the Arkansas West Gulf Coastal Plain are the Ouachita (6), Little Missouri (5), Red (4), and Little (3) River drainages. The watersheds in the West Gulf Coastal Plain that contain fewer than statistically expected Caddo mound sites are interpreted here to represent cultural boundaries or buffers separating contemporaneous but separate cultural areas (Anderson 1994:39). Caddo Mounds and Hydrography Water is, of course, implicitly important to people of all cultures for sustenance. Streams in particular have fulfilled the practicalities of life in Spatial Patterns of Mound Sites

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

321

prehistoric times, but have also been used for cultural purposes, such as defense, transportation, and communication. The commonly accepted but previously unquantified perception concerning Caddo settlement patterns in the West Gulf Coastal Plain of Arkansas has held that the majority of mound sites were located along higher-order streams. Indeed, 78 of 114 (68 percent) Caddo mound sites in the West Gulf Coastal Plain of Arkansas occur within 1 km of a higher-order stream, and 92 (81 percent) occur within 2 km. The pattern most anomalous to this convention consists of Caddo mound sites that are located more than 8 km from a higher-order stream. These sites occur in cluster 5 (see fig. 11-1), and all are among the highest-elevation Caddo mound sites in the West Gulf Coastal Plain of Arkansas. The Tom Jones mound site is counted among this anomalous group and is the third highest in elevation. Caddo Mounds and Elevation Elevation, as a site selection criterion, has been important to people for practical reasons, such as flooding, but also for cultural reasons, such as status or intervisibility (see Vogel, this volume). Elevations in the West Gulf Coastal Plain of Arkansas range from 16 to 253 m above mean sea level (amsl). Seventy-eight percent (89 of 114) of the Caddo mound sites there occur below 85 m amsl. Consequently, there are more Caddo mound sites than would be statistically expected in the lowest elevation categories, and fewer than expected in the highest. Caddo Mounds and Slope

Copyright © 2012. Nebraska Paperback. All rights reserved.

Slope is another commonly used variable in archaeological modeling. Throughout time, people have located domestic and special use sites on the landscape in which slope is a consideration. Rock shelters containing rock art, for example, are commonly found along steep slopes that offer commanding and inspirational views. Alternatively, mounds are constructed on relatively level ground for purposes of stability. For this analysis, slope data for the West Gulf Coastal Plain are classified into nine 5-degree classes: 0–5 degrees, 6–10 degrees, etc. Based on these data, more than 97 percent of 114 Caddo mound sites in the West Gulf Coastal Plain are located on slopes of less than 5 degrees. Furthermore, 0 to 5 degrees was the only category containing more sites than would be statically expected by chance.

322

Lockhart

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Caddo Mounds and Aspect Aspect data were also derived from the usgs ned dem, and these data are useful in digital terrain modeling applications, such as line-of-site and intervisibility analysis. Aspect data are also useful in examining the distributions of other special use sites. However, the analysis of how Caddo mounds are correlated with slope in the West Gulf Coastal Plain has shown that the sites occur on flat to nearly flat ground in which aspect would play a diminished role. Specifically, 88 percent of 114 Caddo mound sites in the West Gulf Coastal Plain are characterized as having no slope and thereby no aspect.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Caddo Mounds and Geology Geology is considered here to be another potential factor in the location of Caddo mound sites. Beyond the practicality of rock shelters and other lithic resources in everyday life in prehistory, archaeological evidence suggests that geologic features played a role in the conceptualized and ideational landscapes of prehistoric cultures (Knapp and Ashmore 1999; Brooks, this volume). Rock art, for example, is among the most tangible evidence of abstract thinking in prehistory (Hilliard et al. 2004). An example within the Caddo archaeological record might include the apparent use of fossilized Cretaceous era ammonites, which have been found in archaeological contexts (Frank Schambach, personal communication 2003). Unlike elevation, geology is not a continuously mapped variable. The locations of boundaries separating different geologic categories are necessarily the result of interpretation based on surficial indicators and interpolation from discrete subsurface testing. Keeping in mind their limitations in large-scale and intra-site analyses, these small-scale geological data remain useful in broader regional applications, such as a study area the size of the West Gulf Coastal Plain of Arkansas. To summarize, three general geologic categories are predominant in Arkansas’ West Gulf Coastal Plain: the Claiborne group (19.4 percent), terrace deposits (25.2 percent), and Quaternary alluvium (Qal). Qal alluvium is notable because 74.5 percent of the Caddo mound sites occur on the 26.5 percent of Qal composing the West Gulf Coastal Plain of Arkansas, which further supports the commonly accepted model of Caddo mound

Spatial Patterns of Mound Sites

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

323

centers being located adjacent to higher-order streams. The Claiborne group and terrace deposits, alternatively, are notable for having fewer Caddo mound sites than would have been statistically expected. Caddo Mounds and Soils The West Gulf Coastal Plain is composed of 358 ncrs-statsgo soil categories. The most evident correlates in these data are the 57 (50 percent) Caddo mound sites located on the category 729 Guyton-Amy-Ouachita soil association, which accounts for less than 13 percent by area of Arkansas’s West Gulf Coastal Plain. The soil type occurs in alluvial settings along the Ouachita, Saline, and Little Missouri River valleys, which are composed of fertile bottomlands that would be conducive to a sedentary agrarian economy. Caddo Mounds and Vegetation

Copyright © 2012. Nebraska Paperback. All rights reserved.

Another source of environmental information used in the present research is modern land use/land cover from usgs National Land Cover Data (nlcd), which has been derived from Landsat Thematic Mapper (tm) satellite imagery and consists of 21 categories. The data layer is an unsupervised classification in which similar spectral signatures are clustered. The nlcd data merge satellite imagery from both leaf-on and leaf-off seasons. Summarizing 3 of the 21 West Gulf Coastal Plain land cover types (table 11-3), two vegetation categories show positive associations between Caddo mound sites and modern open-land uses such as pastures, hayfields, and row crops. These modern data were used because certain current land use may reflect modern analogues or even small vestiges of prehistoric landscapes. In some instances, modern land uses may also represent a temporal-spatial continuum based on environmental characteristics that humankind has long valued for particular purposes, such as farming. In addition to the positive patterns indicated in table 11-3, there is also a strong negative relationship between Caddo mound sites and predominantly evergreen forests of the eastern and southern West Gulf Coastal Plain of Arkansas. Today, many of these pine forest stands are managed by lumber companies, but it is likely that these areas were predominantly evergreen forest in late prehistoric times as well (Delcourt and Delcourt 1981). Largely monoculture evergreen forests would 324

Lockhart

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 11-3. Caddo mounds and land use/land cover in the West Gulf Coastal Plain WGCP land cover Evergreen Forest – 75 percent Pasture/Hay–grasses Row Crops–corn, soybeans

Percent of WGCP

Expected sites

Actual sites

χ2

40.6 8.6 3.7

63.4 13.4 5.8

18 38 20

32.48 45.22 35.25

Copyright © 2012. Nebraska Paperback. All rights reserved.

Table 11-4. Summary of environmental variables associated with most Caddo mounds in the West Gulf Coastal Plain Environmental data layer

Positive association with Caddo mound sites of Arkansas’ West Gulf Coastal Plain

Watersheds Elevation Slope Distance to Primary Stream Soil Association Geology Modern Land cover

Ouachita, Little Missouri, Red, and Little River drainages Below 85 m amsl Less than 5 degrees Less than 2,000 m Guyton-Amy-Ouachita Alluvium (Qal) Pasture, hay, grasses, row crops, woody wetlands

generally have presented decreased environmental variety, decreased resource availability, and fewer and smaller open areas when compared with the primary Caddo culture area to the north and west in the West Gulf Coastal Plain. Table 11-4 summarizes the combined results of the foregoing analysis of individual environmental correlations, which suggest that prehistoric Caddo people did not choose the location of their mound sites randomly. Instead, the data for Arkansas’ West Gulf Coastal Plain indicate a primary pattern of selecting locations in the northern and western half of the physiographic region subsuming the Ouachita, Little Missouri, Red, and Little River drainages. Most Caddo mound sites in Arkansas’s West Gulf Coastal Plain are located in the lowest one-third of the elevation range in the region, below 85 m amsl. The mounds are most commonly found within 2 km of a higher-order stream on alluvial sediments and soils, as opposed to terraces and highlands. The distribution of modern, possibly vestigial, land covers and related land use through time suggest a prehistoric preference for flat to nearly flat semi-open lands in proximity to the edge environments of wooded wetlands and floodplains. Spatial Patterns of Mound Sites

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

325

An Environmental Similarity Model

Copyright © 2012. Nebraska Paperback. All rights reserved.

This study affirms that mound sites and associated settlements were organized by the prehistoric Caddo into patterns that reflect an acute awareness of their environmental and cultural boundaries. These data, and those analyzed in other Caddo studies (see Brooks, this volume), suggest a pattern of shared environmental criteria among Caddo people. But spatial data in tabular form provide only part of the exploratory data analysis picture. Additional insights can be gained through extrapolation of those data to the entire study area to map all locales in which those same generally shared criteria coincide. A mapped model of environmental similarity based on the prehistoric placement of Caddo mound sites can be produced to (1) further define the general Caddo cultural landscape and settlement pattern in the West Gulf Coastal Plain of Arkansas; (2) assist in the identification of known but currently unaffiliated mound sites in the region; (3) identify Caddo mound sites that do not fit the principal spatial pattern; and (4) identify large contiguous areas that are environmentally similar to those of the Caddo mound site pattern but contain no Caddo mounds. The methodology for producing what is termed here an “environmental similarity model” is, in some ways, comparable to archaeological predictive modeling. However, the use of a new term for the model type to be detailed in this application is more than semantic in that predicting the location of undiscovered Caddo mound sites is not the purpose of the model. Nor is this model purely descriptive in that the majority and minority patterns represented were not known a priori through previous empirical quantification. To produce the environmental similarity model, a binary gis data layer for each of seven individual environmental characteristics (table 11-5) was generated. This was accomplished by reclassing the categories of the original data according to positive associations between Caddo mound sites and individual categories within each environmental theme, which is to say categories containing more sites than would be statistically expected given a hypothetically random distribution. In addition, to improve the statistical robustness of the final model, it was required that categories of positive correlation contain five or more expected and actual sites.

326

Lockhart

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Table 11-5. Statistics for the seven-class environmental similarity model Environmental similarity

Percent of West Gulf Coastal Plain

Expected sites

Actual sites

(Least) 1 2 3 4 5 6 (Greatest) 7

11.3 41.9 22.0 10.4 7.6 6.0 0.9

12.9 47.7 25.1 11.8 8.6 6.8 1.1

0 8 12 15 27 32 20

χ2 12.875 33.069 6.811 0.866 39.067 93.114 331.027

Degrees of freedom 1 1 1 1 1 1 1

As previously described, environmental themes were elevation, slope, geology, soils, modern land cover, watersheds, and distance to higher-order streams. For each of the individual data layers, all locations representing a category of positive correlation with Caddo mound sites and containing at least five expected sites received a value of 1. Locations composed of categories that were not correlated, negatively correlated, or containing fewer than five expected sites received a value of 0 (null). Using these rules, a seven-class environmental similarity model was produced by effectively summing the seven reclassed environmental gis data layers. Because the maximum value for each of the seven individual binary data layers was 1 and the minimum was 0, values in the resulting additive data layer theoretically range from 0 to 7 at any given location. Table 11-5 presents the statistics for the seven-class environmental similarity model in which categories 1–3 represent locations where one, two, or three of the environmental conditions were met, all of which contain fewer actual sites than would be statistically expected. Alternatively, categories 4–7 represent locations where four, five, six, or all of the environmental conditions (rules) were met, all of which contain more actual sites than would be statistically expected. Then, based on the seven-class environmental similarity data layer, categories 4–7 were combined to produce a binary model (fig. 11-2). Model Performance As previously stated, the intent of the model is not to predict sites. In fact, large areas of environmental similarity that contain no sites are considered at least as instructive as modeled areas containing many

Spatial Patterns of Mound Sites

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

327

11-2. Environmental similarity model for Caddo mound sites in the West Gulf Coastal Plain of Arkansas (model accuracy is 82 percent, model accuracy using independent sites is 80 percent).

Copyright © 2012. Nebraska Paperback. All rights reserved.

sites. Although the specific purpose of this model differs from that of a predictive model, similarities in methodology provide an opportunity to measure the extent to which the modeled areas coincide with the distribution of Caddo mound sites. To assess the performance of predictive models, Kvamme (2006:29) recommends calculations to measure precision and accuracy. Model precision is the ratio of the area outside the modeled high potential (environmental similarity) to the total study area: the higher the ratio (approaching 1:1), the greater the precision. The precision of the binary Caddo mound site environmental similarity model for the entire West Gulf Coastal Plain, as shown in figure 11-2, is 25,534 km2 to 33,936 km2 (0.75:1). Model accuracy, on the other hand, is the proportion of Caddo mound sites located in areas of modeled environmental similarity within the entire West Gulf Coastal Plain, and that proportion is 94 of 114 sites or 82 percent. Kvamme (2006:26) has further stated that the “ultimate” test of a model is to examine how well it performs using a group of sample sites that are independent of those used to develop the model. As previously described, the environmental similarity model is based on the loca328

Lockhart

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

tions of 114 Caddo mound sites with an amasda encoding of “good” for both cultural affiliation and locational reliability. In addition to those 114 sites, there are 30 additional Caddo mound sites in the West Gulf Coastal Plain that are encoded as having “good” cultural affiliation reliability, but only “fair” location reliability, meaning that the encoded utm location is considered to be accurate to within 40 acres. Twentyfour of those 30 Caddo mound sites are located within the area of modeled environmental similarity, yielding a comparable model accuracy value of 80 percent.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Caddo Cultural Buffers in the West Gulf Coastal Plain The environmental similarity model is based on correlations between Caddo mound sites and elements of the physical environment. Cultural elements of Caddo mound site distributions are more difficult to measure. The model does, however, implicitly suggest cultural manifestations of the site distribution by delineating contiguous areas of environmental similarity where no Caddo mound sites are recorded, as well as Caddo mound sites that are located outside the modeled area. As shown in figure 11-2, in the easternmost third of the West Gulf Coastal Plain in Arkansas (east of the Ouachita River) there are extensive contiguous areas with similar environmental characteristics to those that the prehistoric Caddo mound builders apparently selected for in other parts of the study area. However, there are few mound sites of assigned Caddo cultural affiliation in those eastern areas of modeled environmental similarity, such as the lands along the eastern Saline River and its tributaries, as well as the lower Ouachita River drainage in Arkansas. Indeed, there is a contiguous expanse of more than 5,800 km2 between the Saline River and Bayou Bartholomew containing only two Caddo sites of reliable cultural affiliation and location, one Caddo mound site, and no reliably non-Caddo and contemporaneous Mississippi period mound sites. That same expanse contains only 22 recorded Mississippi period sites of any type, resulting in an average site density of one site per 263 km2. If environmental conditions within these more sparsely populated areas were potentially conducive to the lifeways of the Caddo, the paucity of sites may imply the existence of cultural boundaries. Although few large areas in Arkansas have been systematically and intensively surveyed, the paucity of mound sites of any cultural affiliation in the eastern portions of the West Gulf Coastal Plain (Lockhart 2007:fig. 2.1) Spatial Patterns of Mound Sites

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

329

may provide evidence suggesting an intentionally maintained buffer separating the prehistoric Caddo from mound-building cultures living in the adjacent Mississippi River alluvial valley. This hypothesis is further supported by the crowding of Caddo mound sites in proximity to the Ouachita River on the west and those mound sites of contemporaneous cultures in proximity to Bayou Bartholomew to the east, with few contemporaneous mound sites between. To the south (see fig. 11-2), another paucity of Caddo mound sites (or those of any other affiliation) in the large expanse between the Ouachita and Red River drainages is mirrored in the diminished area of modeled environmental similarity. This pattern suggests a reduced population density in an area that may have been too environmentally dissimilar to Caddo-populated and favored areas to the north and west. Much of this less populous area is part of the Felsenthal region, which Schambach (1989:14) excludes from the Caddo homeland partially because of its greater resemblance to the lower Mississippi valley environment than to that of the West Gulf Coastal Plain. The possibility remains, however, that the scarcity of mound sites of any cultural affiliation in this area could have cultural/territorial implications as well. A Portion of the Prehistoric Caddo Homeland within Arkansas’ West Gulf Coastal Plain

Copyright © 2012. Nebraska Paperback. All rights reserved.

By calculating the density of Caddo mound sites in addition to modeled environmental similarity, specific areas can be delimited of greatest potential population and land use to represent the habitation, activity, and reasonably accessible spaces of prehistoric Caddo culture within the West Gulf Coastal Plain of Arkansas. In figure 11-3, the outermost extent of the area that subsumes the modeled environmental similarity is the statistical surface contour at which estimated Caddo mound density reaches zero as calculated using the Kernel Density estimation module within esri Spatial Analyst. Based on an arbitrary travel approximation of a maximum one-day round-trip walk, 20 km (12.5 miles each way) was specified as the search radius. The density value of each cell in the output raster data layer is weighted according to its distance from incidences of Caddo mound sites, thus creating a statistical surface bounded by the areal extent of estimated Caddo mound site density greater than zero. Within that general boundary of reasonable accessibility, the extrapolation of areas most 330

Lockhart

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

11-3. Habitation, activity, and accessibility spaces of prehistoric Caddo culture in the West Gulf Coastal Plain of Arkansas based on Caddo mound site density and environmental similarity.

environmentally similar to the majority of Caddo mound sites defines the “habitation” and “activity” spaces (Kvamme 2006:16) of prehistoric Caddo cultural groups in the West Gulf Coastal Plain of Arkansas. The environmental similarity model comprises just 29 percent of the area of Caddo mound density greater than 0, so the ratio for model precision is 0.71:1 (71 percent). Encoded in amasda, there are 588 “good” Caddoaffiliated sites of all types, and 544 (92.5 percent) of those sites occur within the modeled accessibility space, which is only 49 percent of the West Gulf Coastal Plain (see fig. 11-3) in Arkansas. The delimitation of the habitation, activity, and accessibility space of Caddo cultural groups within the West Gulf Coastal Plain of Arkansas is a statistical interpolation and a cartographic representation of a realistically nondiscrete boundary. Because the general prehistoric settlement pattern of the Caddo has been associated ethnographically and archaeologically with mound construction and ceremonialism, it was hypothesized that the areas for which Caddo mound site density is estimated to be greater than 0 (which in turn subsumes the areas of greatest Spatial Patterns of Mound Sites

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

331

environmental similarity containing the greatest number of Caddo mound sites) would also encompass the largest and densest prehistoric Caddo populations in Arkansas’s West Gulf Coastal Plain during prehistoric and early historic times. The statistics reported in the preceding paragraph, while not conclusive, support that hypothesis. Additional support for the hypothesis and the resulting Caddo habitation/ activity/accessibility model can be inferred from Brooks’s research (this volume). In his examination of the distribution of residential and ceremonial sites in the West Gulf Coastal Plain of southeastern Oklahoma, Brooks concludes that there is no difference in site placement based on site function. The model boundaries established in this research do not preclude culturally related populations in other physiographic subdivisions of Arkansas. Likewise, the model boundaries do not preclude populations within the same physiographic region that fit neither the traditional nor statistical models of Caddo settlement patterns based on environmental similarity. Neither do the boundaries preclude a wider distribution of Caddo-related special use and transient site types. Instead, the boundaries are intended to focus attention on an area of cultural zenith and social organization composed of the greatest Caddo populations, population densities, and concentrated land use based on the best information that is currently available. Conclusions

Copyright © 2012. Nebraska Paperback. All rights reserved.

“What is the spatial pattern represented by the entire class of Caddo culture mound sites in Arkansas’s West Gulf Coastal Plain, and how does the Tom Jones (3HE40) site fit into that pattern?” In this chapter, a statewide archaeological database, gis, and statistical analyses have been used to locate environments similar to those that typified the majority of Caddo mound sites in the West Gulf Coastal Plain of Arkansas. These analyses have empirically confirmed a general pattern characterized by flat, semi-open areas at lower elevations of the West Gulf Coastal Plain in proximity to wooded wetlands and high-order stream valleys composed of alluvial soils. Extrapolating those coincident characteristics to the entire West Gulf Coastal Plain of Arkansas provides a modeled cartographic approximation of common Caddo habitation and activity spaces, which 332

Lockhart

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

were bounded by a larger, statistically derived region that would have been reasonably accessible to the Caddo populations affiliated with the mounds. The modeled region of accessibility encompassed 544 of 588 (92.5 percent) Caddo sites of good cultural and locational reliability in Arkansas’ West Gulf Coastal Plain. Beyond the areas reasonably accessible from the Caddo mounds discussed herein (and where fewer Caddo sites and artifacts have been discovered), Caddo occupation and population of Arkansas’ West Gulf Coastal Plain would logically have become increasingly sparse. In addition to quantifying a general spatial pattern of Caddo sites, results of this analysis have identified large contiguous areas of environmental similarity within the eastern half of Arkansas’s West Gulf Coastal Plain for which there are no Caddo mound sites and only two Caddo sites of any type encoded in the amasda database. Moreover, based on amasda data, there is another 4,000 km2 with a dearth of mound sites between the lower Ouachita and Red Rivers in Arkansas, areas which are adjacent but environmentally dissimilar to the Caddo habitation, activity, and accessibility spaces defined by the model. The environmental similarity model has also been used to identify Caddo mound sites in the West Gulf Coastal Plain of Arkansas that do not fit the principal spatial pattern. Thirteen anomalous Caddo mound sites, including Tom Jones, that do not fit the general pattern of river valley settlement are located in the highest elevations. They were sited near and, in some instances, precisely on the watershed divides of the Little, Little Missouri, and Red Rivers. The environment that was chosen for these atypically situated Caddo mound sites is unique, not just for its maximal elevations but for its Blackland Prairie ecosystem (Lockhart 2007). This study employs the more statistically defensible class of mound sites to examine areas of prehistoric Caddo population and culture, but the prospect also exists for examining the distribution of other key, but less statistically robust, site types, such as habitation sites, in order to gain a fuller perspective on Caddo settlement patterns. Potential benefits from such an examination could include a more detailed understanding of the size, density, and pattern of populations that occupied Arkansas’s West Gulf Coastal Plain, as well as other physiographic regions. Although some records are available, the amasda database will require additional vetting in order to produce reliably representative and Spatial Patterns of Mound Sites

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

333

statistically viable site counts necessary for studies involving large geographic areas and finer temporal distinctions. Also, more detailed information relating to widely dispersed prehistoric Caddo residential sites will require additional location, identification, and subsurface testing of diagnostic archaeological features. For this research, the more readily recognizable and quantifiable feature characteristics of mound sites were thought to ensure a more representative data set as a class than lesser known and necessarily estimated subsurface features. Although outside the scope of this chapter, it can hardly be overemphasized that the best understanding of the larger prehistoric Caddo cultural landscape will be most fully realized through cooperation, coordination, and data sharing among agencies and Caddo archaeological researchers in surrounding states in order to encompass the entire Caddo homeland. Acknowledgments

Copyright © 2012. Nebraska Paperback. All rights reserved.

I gratefully acknowledge the Caddo Nation of Oklahoma, the Arkansas Archeological Survey and the Arkansas Archeological Society, University of Arkansas Departments of Environmental Dynamics, Anthropology, Geosciences, the ua Center for Advanced Spatial Technologies, and the Arkansas Game and Fish Commission’s Rick Evans/Grandview Prairie Wildlife Management Area. I would also like to thank Drs. Thomas J. Green, George Sabo III, Frank F. Schambach, Kenneth L. Kvamme, Timothy K. Perttula, and Stephen K. Boss. My thanks also to Dorothy G. Neely, Trew W. Lockhart II, and Barbara F. Lockhart.

334

Lockhart

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

12. Decisions in Landscape Setting Selection of the Prehistoric Caddo of Southeastern Oklahoma A GIS Analysis robert l . brooks

Copyright © 2012. Nebraska Paperback. All rights reserved.

Introduction Since the 1950s archaeologists have been interested in the question of how sites are spatially distributed. In a seminal study Gordon Willey (1953) examined patterns in the distribution of prehistoric settlements in the Viru Valley in Peru. Willey argued that the patterns in settlement and their changes through time reflected decisions on the part of the prehistoric societies in terms of where to settle as well as responses to changes in environmental, economic, political, and social conditions. Willey’s work also served as an avenue for examining a basic issue in archaeology: culture change. The practice of examining settlement distributions and changes in their patterns soon became widespread if not pervasive in archaeology (cf. Gumerman 1971; Smith 1978a). The study of “settlement patterns” became a fundamental part of basic archaeological analysis in the 1970s and 1980s. During this time, there were also ancillary types of analyses ranging from the examination of the circumscribed natural surroundings of a site-catchment (Findlow and Ericson 1980) to various statistical means of examining patterns in the spatial distribution of sites on the landscape (Hodder and Orton 1976). While settlement pattern analysis was widely embraced in the processual era of the 1970s and 1980s, it came under increasing criticism in the post-processual period that began in the 1990s. Settlement pattern studies were criticized for focusing too much attention on the natural/physical attributes in settlement placement choice with too little focus on a particular group’s decision-making processes in respect to locational

335

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

choice (Hirsch and O’Hanlon 1993). It was argued that many factors such as cultural preferences, political pressures, and religious considerations were not addressed through existing analytic procedures. There was also a concern that settlement pattern analysis was a static/passive perception of a group’s relationship to their natural and cultural surroundings. It was suggested instead that this relationship should be perceived as a dynamic or active one between the group and the natural and cultural world in which they lived. This more interactive dynamic view of a group’s behavior in a spatial sense became known as “landscape archaeology” (cf. Bender 1993). While some landscape archaeology studies differed little from previous considerations of settlement patterns, others made a clear effort to distance themselves from this convention, examining issues such as the use of the landscape in a religious context (Tilley 1994). A Conceptual Model for the Examination of the Caddo Landscape

Copyright © 2012. Nebraska Paperback. All rights reserved.

This chapter focuses on an examination of the Caddo cultural landscape in one part of the Caddo area with the premise that landscape is a cultural construct. Within this context, Knapp and Ashmore (1999:10–13) describe constructed, conceptualized, and ideational landscapes. Constructed landscapes are a built environment representing an active intervention by a group in the form that the landscape takes. In other cases, the natural landscape may have associative religious, artistic, or social meaning. This is a conceptualized landscape. An ideational landscape is one that takes on the role of serving as a symbolic reference for mythical histories or other symbolic messages. The implication is that the landscape cannot be viewed as simply a physical feature upon which cultures construct a variety of built environments. Lewis et al. (1998) have proposed that the design of prehistoric Mississippian towns is ritually prescribed: that there is an architectural grammar for the construction of these places on the landscape that also represents sacred spaces. Obviously, an even stronger argument can also be made for the building of mounds and other ceremonial features on the landscape. However, if such an architectural grammar is present, is it consistently applied to different hierarchies within the built environment? The focus here is to examine Caddo residential and ceremonial constructions on the southeastern Oklahoma landscape and determine 336

Brooks

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

whether there are distinctions in their placement that may reflect an active structuring of the landscape. Are places with mounds situated differently than residential places? Are there other landscape distinctions between larger ceremonial centers (places with multiple constructed mounds) and single mound locations? Such an analysis obviously must be accomplished through a study of attributes of the constructed landscape and their relationship to attributes of the natural environmental setting (i.e., soils, landform, elevation, distance to water). Through this examination it may be possible to gain some insights into the conceptual/ideational landscape of Caddo peoples living for hundreds of years along the Red River. The conceptual and ideational landscape may be realized through determinations of the distances between towns or centers. These distance measures have been frequently associated with the exercise of control of a center over the surrounding towns. But it may also be that this distance reflects the associative meaning for a particular center’s conceptual landscape. Such an idea bears greater weight when fictive or real kinship ties broadly relate the groups in question. The ideational landscape can best be understood, however, through an examination of Caddo history and mythology. The relationships that exist between the lower world, middle world, and upper world (Lankford 2004) and the ideational landscape may be understood through Caddo stories of mythical beings and their travels between these worlds (cf. Dorsey 1905).

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Natural Landscape Choctaw and McCurtain Counties in the extreme southeastern corner of Oklahoma are part of the Gulf Coastal Plain geographic province and contain the lowest point in Oklahoma (Johnson 2006a:7) (fig. 12-1). Immediately to the north of the coastal plain are the foothills and higher elevations of the Ouachita Mountains, extending to a maximum of roughly 2,000 ft at their highest elevation. These differences provide for some pronounced relief where the Gulf Coastal Plain and the Ouachita Mountains are intermeshed. Geologically, the Gulf Coastal Plain is made up of sand, silt, clay, and gravel of Quaternary age in the floodplains and terraces of the Red River valley. To the north lie Cretaceous age sand, clay, and marine limestone. The Ouachita Mountains contain deposits ranging in age from the Cambrian through Pennsylvanian times. The Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

337

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

12-1. Study area of Choctaw and McCurtain Counties and streams and rivers of southeastern Oklahoma.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Copyright © 2012. Nebraska Paperback. All rights reserved.

mountains are composed of limestone, chert, sandstone, and shale (Johnson 2006b:9). Soils in the two counties are represented by vertisols with high clay content along some portions of the Little River and the Red River valley. Small areas of alfisols also extend along Little River in central and eastern McCurtain County. However, vast portions of the two counties are composed of ultisols, which are acid and highly leached soils formed under an oak-hickory-pine forest (Hoagland 2006:16–17). Oak-hickorypine forests occur in the uplands with hardwood forests in the valleys except where the hardwoods have been replaced in modern times by pine plantations. Fauna are those typically associated with lowland hardwood and oakhickory-pine forests. Among the fauna that would have served as food for the prehistoric Caddo are white-tailed deer, black bear, and smaller game such as raccoon, beaver, rabbit, and squirrel. Birds present include wild turkey as well as migratory waterfowl. There would also have been an abundance of fish species in the streams and rivers of Choctaw and McCurtain Counties. The area receives from 46 to more than 50 inches of precipitation annually. Combined with the small number of days with 32 degrees or less (< 70), McCurtain County in particular has a growing season of 215–235 days. Such conditions undoubtedly served as favorable settings for agriculture during the time the Caddo lived in southeastern Oklahoma. There are five major stream drainages within the two counties that figure prominently in this chapter (see fig. 12-1). They all play critical roles in the distribution of Caddo sites within the region. The Kiamichi River originates in southern Le Flore County and drains in a southwesterly direction in Pushmataha County where it is dammed to form Sardis Lake (not depicted in fig. 12-1). Leaving Sardis Lake, the Kiamichi flows in a southerly direction through Choctaw County, where it is damned to create Hugo Lake. Exiting Hugo Lake, the Kiamichi continues in a southerly direction and enters the Red River in eastern Choctaw County. Little River originates in extreme southwestern Le Flore County, follows a southerly trend through Pushmataha County, and enters western McCurtain County where it is dammed to form Pine Creek Lake. A short distance after exiting the lake (near Wright City), Little River takes an abrupt eastward course, paralleling Red River, and drains most of the northern two-thirds of McCurtain County. The Glover River (labeled as Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

339

Glover Creek in fig. 12-1) originates in central McCurtain County, flows in a southerly direction, and has its confluence with Little River in the southwestern part of McCurtain County. Mountain Fork River is the most easterly of these drainage systems. It originates in southeastern Le Flore County, flows through the Ouachita Mountains, is dammed to form Broken Bow Lake, and then maintains a southern course before its confluence with Little River about 8 km west of the Arkansas border. The Red River, originating in the Texas Panhandle, forms the southern border of Oklahoma and flows in an east-southeasterly direction. It is dammed to form Lake Texoma, to the west of the Caddo area (near Denison, Texas). The Red River forms the southern boundary of Choctaw and McCurtain Counties, as well as the boundary between Arkansas and Texas, and flows through Louisiana to its confluence with the Mississippi River in the lower Mississippi valley. It is the largest ranked stream of the region, and the previously discussed rivers all eventually drain to the Red. Previous Archaeological Study of Caddo Sites in Southeastern Oklahoma

Copyright © 2012. Nebraska Paperback. All rights reserved.

Archaeological research in the two southeasternmost counties bordering on the Red River (Choctaw and McCurtain Counties) can be divided into five distinct categories. These are projects conducted by the Works Progress Administration (wpa), investigations in advance of lake construction by the Oklahoma River Basin Survey (orbs), surveys undertaken by the University of Oklahoma and Museum of the Red River for National Park Service–funded Survey and Planning efforts, work conducted for recent federally or permitted-funded construction projects, and volunteered site reports. The wpa, orbs, Survey and Planning work, and contractual studies fall within specific temporal boundaries; each of these also represent different methodologies that affected how they have contributed to our understanding of southeastern Oklahoma Caddo archaeology. The volunteered site reports, on the other hand, are distributed across a wide span of time and also have different and sometimes greater constraints in the building of a database of Caddo archaeology. Works Progress Administration The purpose of the wpa’s involvement in archaeology was to provide a source of income for unemployed rural laborers during the Great De340

Brooks

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

pression (Lyon 1996). Between 1936 and 1942, the wpa, using graduate students under the supervision of Dr. Forrest Clements from the University of Oklahoma, excavated more than 100 archaeological sites, including numerous mounds and mound centers, in the eastern part of the state (Rogers 1978). Investigations in southeastern Oklahoma were less intensive, with only a few sites studied, than that conducted in the Arkansas River valley or in northeastern Oklahoma. In Choctaw County, the wpa conducted excavations at the Nelson (34CH7) and Cook (34CH8) sites in 1940. These were village sites that appear to represent some of the westernmost sites of the prehistoric Caddo tradition. Although excavated in the early 1940s, materials from the Cook and Nelson sites have neither been fully analyzed nor written up for publication. In McCurtain County, the wpa conducted excavations at the Clement mound and village areas (34MC8–10), the MacDonald I and II sites (34MC11 and 34MC12), and the Pitman site (34MC14, a single mound). The MacDonald sites were excavated in December 1941; Clement and Pitman also were excavated earlier in December 1941. As in the case of the Choctaw County sites, the excavations and materials from the Clement site have never been thoroughly studied, and only limited parts of the ceramic assemblage have been published (e.g., Flynn 1976). The Pitman site has experienced much the same treatment with no examination since its excavation in 1941. Most of the wpa documentation for these sites is sparse or, in the case of the Pitman site, absent except for a short note accompanying the artifacts. In most cases, the wpa records were sufficient for subsequent relocation of the site. However, the Pitman site, because of discrepancies between the legal location and the physical description, has not been relocated. Oklahoma River Basin Survey Following World War II the United States began a vast public works program to construct roads and lakes. The lake construction served not only to provide flood control, water supply, and recreation for a number of regions of the United States but also to employ large numbers of men returning from the war. As part of this massive construction program, archaeological investigations were also undertaken. The archaeological portion of the lake building was overseen by the National Park Service and referred to as the River Basin Survey (rbs) program. The Oklahoma Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

341

Copyright © 2012. Nebraska Paperback. All rights reserved.

rbs conducted surveys and excavations at some 20 proposed lake projects throughout the state. Graduate students at the University of Oklahoma under the supervision of Dr. Robert E. Bell were responsible for virtually all of the investigations during this period. Three of these lake projects were located in southeastern Oklahoma: Hugo Reservoir, Broken Bow Reservoir, and Pine Creek Reservoir (see fig. 12-1), as was the proposed but not constructed Lukfata Reservoir. Hugo Reservoir was built on the Kiamichi River in eastern Choctaw County and southeastern Pushmataha County. The reservoir was initially surveyed by avocational archaeologist Sherman Lawton (Lawton 1960). Work was carried out by the Oklahoma rbs in the 1970s and included excavations at Mahaffey (34CH1), Moore 6 (34CH37), Spencer Spring (34CH43), Montgomery (34CH70), Pate-Roden (34CH90), Hugo Dam (34CH112), and Deaton (34PU82) (Burton 1970; Rohrbaugh et al. 1971; Lewis 1972; Rohrbaugh 1973). Many of these contained Caddo occupations, although none were mound sites. Pine Creek Reservoir involved the damming of the upper reach of Little River in western McCurtain County. Don Wyckoff conducted the initial survey of Pine Creek (Wyckoff 1963) and also excavated two Caddo residential sites, Bell (34MC96) and Gregory (34MC98), some five years after the initial survey (Wyckoff 1968b). Three additional Caddo village or town sites were also investigated by Barr (1965). There were no mound sites or ceremonial centers within the Pine Creek Reservoir project area, although the Grobin Davis Mounds site (34MC253) is located downstream. The proposed area of Lukfata Reservoir in McCurtain County was investigated by Gettys (1975) and resulted in the documentation of a number of Caddo residential sites as well as the Pine Creek Mounds (34MC146), a multiple mound site at the upper end of the proposed reservoir. The last of the reservoirs for the area was Broken Bow, built on the Mountain Fork River in eastern McCurtain County. A number of Caddo period sites were present in the proposed project area, including Beaver (34MC1; Wyckoff 1968a), E. Johnson (34MC52; Wyckoff 1967a), and Biggham Creek (34MC105; Wyckoff 1965b). The Biggham Creek site contained a residential occupation and two earthen mounds. The Woods Mound Group (34MC104), a multiple mound center dating to Late Caddo period (Wyckoff 1967b), was also in the Broken Bow Reservoir area. Investigations undertaken during the Oklahoma rbs era are well documented, and all were published shortly after completion of the 342

Brooks

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

fieldwork and analysis. Although 1:24,000 scale usgs topographic maps were not available for many of the lake construction projects, detailed project maps had been prepared by the U.S. Army Corps of Engineers, who were responsible for lake construction. Thus accurate locations can be established for most of the sites. This is especially relevant for this chapter, as most of the Caddo sites excavated by rbs personnel are now inundated under Hugo, Pine Creek, and Broken Bow lakes. Material recovered during the work of the Oklahoma rbs is curated at the Sam Noble Oklahoma Museum of Natural History.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Survey and Planning The Survey and Planning era refers to projects funded through the National Park Service and the State Historic Preservation Office at the Oklahoma Historical Society from roughly 1980 through the present. This work represents surveys of targeted areas or selected themes (e.g., the Red River valley/Post-Removal Choctaw homesteads). This type of activity has been relatively limited for extreme southeastern Oklahoma. There have been inventories of areas affected by timber clear-cutting (Neal 1988a, 1988b), surveys of the Kiamichi River valley (Neal 1989, 1990), a study of early Choctaw homesteads (Neal et al. 1991), and an inventory of Caddo sites along Red River conducted by Jack Bennett and the Museum of the Red River. All of these projects resulted in the documentation of prehistoric Caddo sites, but only the survey by Bennett specifically targeted Caddo sites. There was also a specific survey and planning project designed to establish National Register of Historic Places eligibility for the Grobin Davis Mounds (Wyckoff and Ragland Fisher 1985). Federal Regulatory Work With the passage of the National Historic Preservation Act in 1966, federal agencies became responsible for the effects of their actions on cultural resources, in particular those considered to be eligible for inclusion in the National Register of Historic Places. In the late 1970s, federal agencies began to responsively consider this charge and literally hundreds of small and moderately sized survey and site testing programs were undertaken by agencies or their delegated authorities, including the U.S. Army Corps of Engineers, U.S. Forest Service, Oklahoma Department of Transportation/Federal Highway Administration, and the Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

343

Natural Resource Conservation Service (nrcs, formerly the Soil Conservation Service). Despite the large number of projects that have been undertaken, only a few of these merit attention here because of their investigation of Caddo sites. The nrcs conducted excavations at two Late Caddo period village sites in the Waterfall-Gilford watershed (Prewitt et al. 1982) and later investigated Harkey Mound (34MC206; Kraft 1997); Harkey Mound is a naturally occurring landform with a Caddo occupation on it rather than being a built mound. The Ouachita National Forest conducted surveys and investigated sites as part of a massive land exchange with the Weyerhaeuser Corporation involving thousands of acres in McCurtain County. Many of the documented sites contained Caddo occupations as well as four that were single mound or multiple mound centers (Etchieson n.d.). Last, the U.S. Army Corps of Engineers, Tulsa District, funded investigations at Broken Bow Lake for sites to be affected by increasing the lake level for hydropower generation (Perttula et al. 1998; Perttula and Nelson 2004b; Sundermeyer et al. 2004). These investigations yielded information on a number of Caddo hamlets and villages. Caddo Cultural Periods

Copyright © 2012. Nebraska Paperback. All rights reserved.

From studies conducted in the region over the past 75 years, a number of schemes have been devised to establish a Caddo cultural sequence. Some used the Midwestern Taxonomic System to establish a series of foci for Caddo occupation of the four corners region of southeastern Oklahoma, southwestern Arkansas, northwestern Louisiana, and northeastern Texas (cf. Bell and Baerreis 1951; Krieger 1946). Later efforts used more of a stage approach to establish a cultural sequence (see Davis 1970). Dee Ann Story et al. (1990) presents a thorough and thoughtful overview of the efforts. In this analysis of Caddo landscapes, the periods I employ are based primarily on Wyckoff and Ragland Fisher’s (1985:8–18) chronological model of Caddo prehistory in southeastern Oklahoma. The period from ca. A.D. 800–1000 can be regarded as the formative period of development for Caddo culture in southeastern Oklahoma (see also table 1-1, Formative Caddo period). It is at this time that an increasing sophistication in ceramic technology is observed with the production of decorated wares, constricted neck bottles, and carinated bowls. There also 344

Brooks

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

appears to be a greater emphasis on plant processing, if not cultivation of domesticated plants, than was the case during the earlier Woodland period (ca. 500 B.C. to A.D. 800). Accompanying these material developments are increasing sedentism, larger settlements, and evidence of a greater sociopolitical complexity. This complexity may be expressed in the evidence of mound construction, although this was limited in southeastern Oklahoma then. Sites identified for this formative time include Mahaffey, Beaver, E. Johnson, and Biggham Creek. The period from ca. A.D. 1000 through 1200 (table 1-1, Early Caddo period) is clearly a time of growth and “settling-in” to the area. There is an increase in the numbers of residential sites as well as the first appearance of constructed earthen mounds. This is accompanied by greater stylistic variability in ceramic wares, both in form and decoration. With the population increases there was undoubtedly a greater emphasis on farming and agricultural production. The presence of mounds and ceremonial centers also points to some level of community efforts being directed to non-economic ends and the increasing importance of political/religious leaders. Residential hamlets or towns (in the Caddo sense) were Bill Hughes, Beaver, E. Johnson, Gregory, Bell, Payne, Pat Boyd, Nelson, and Cook. Mound centers included Clement, Grobin Davis, and Pine Creek. Thus it appears that mound centers occur at the same time as significant increases in population, rather than following afterwards. The time from ca. A.D. 1200–1450 (see table 1-1, Middle Caddo period; in other regions and localities, this period ends at ca. 1400) is one of continued development and maturation of Caddo societies in southeastern Oklahoma. This period is witness to a continued population increase as well as the establishment of a greater number of sites with mounds or ceremonial centers. This era is also noteworthy for the evidence of interregional contact, especially with “Caddoan” neighbors in the Arkansas River valley north of the Ouachita Mountains. (Here, “Caddoan” refers to groups who share an ancestry as Caddoan-language speakers, but does not reflect the historical continuity of the Caddo witnessed in the Red River valley.) This interaction appears to have been reciprocal, with material goods (red-filmed or slipped pottery) flowing from Caddo sites north to the Harlan, Norman, and Spiro mound centers in the Arkansas River basin. In turn, items appearing to emanate from the Arkansas River valley (cf. Schambach 1991) perhaps are present Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

345

Copyright © 2012. Nebraska Paperback. All rights reserved.

at sites in southeastern Oklahoma as well. However, it may be that some of these items (relics?) did not originate in the Arkansas River valley but were following a pathway through a Mississippian-based religious system (Brooks 2006). Most of the sites occupied in early Caddo times were also in use during the thirteenth and fourteenth centuries. Important new settlements were the Bohannon, Hugo Dam, Roden, McDonald, McKinney, and Young sites. The period from ca. 1450 through 1750 is marked by a diminished Caddo presence in southeastern Oklahoma. Whether this represents an emigration of Caddo populations from the region or a realignment of settlements and political structure is uncertain. Clearly, the sociopolitical-religious complexity of Middle Caddo period times (ca. 1200–1450) is much reduced in Late Caddo societies. Mound construction appears to be limited to a simple sequence of construction, destruction, and burying of religious structures rather than the construction of larger tumuli at centers such as Grobin Davis and Clement. Multiple mounds may be present at Late Caddo sites, but they are the product of reiterative events dictated perhaps by calendrical cycles or deaths of Caddo leaders. There are significantly fewer identified Caddo sites for this period, among them the Baldwin, Gregory, McKinney, Woods Mound Group, Beaver, E. Johnson, Biggham Creek, Bohannon, and Roden sites. Unfortunately, many of the Caddo towns and mounds used in this analysis cannot be identified with a particular Caddo time period. This is because their documentation has been derived from surface survey, they were informant-reported sites with little specific archaeological information, or their periods of occupation cannot be further refined because of a mixture of archaeological deposits from several occupations. While the Caddo culture sequence discussed above should ultimately prove to be useful in examining long-term changes in prehistoric Caddo culture, this analysis is limited to a less-refined approach. Caddo Sites Some 279 Caddo sites in Choctaw and McCurtain Counties, Oklahoma, are included in this landscape analysis. Information on these sites has been obtained from the state of Oklahoma archaeological site files as well as the aforementioned reports resulting from previously discussed investigations as well as volunteered reports. Data collected for these sites include UTMs, landform, soils, elevation, slope, distance to water, 346

Brooks

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

principal water source, built environment type, and period of occupation during the Caddo settlement of southeastern Oklahoma. Built environment refers to the character of site construction by the Caddo. There are three categories of the built environment: (1) residential locations (hamlets or villages); (2) single mound sites (with or without accompanying residences); and (3) multiple mound sites, including those with either the same mound type or those centers with multiple but structurally different mound types. The last attribute (time period) refers to the period (or periods) when a site was occupied during the almost 1,000 years of Caddo settlement in the region. Regrettably, only 21 percent of the Caddo sites could be linked to a specific time of occupation. This low percentage has undoubtedly hampered refined distributional analysis, but for this study, a somewhat time-held-constant perspective will be maintained. Where temporal refinements permit, the more detailed results will be integrated into the analysis. Many of the data attributes have been coded from site form data or reports and may contain source errors. However, the site files have been manually examined as a form of quality control of the Caddo site data.

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Distribution of Caddo Sites The 279 Caddo sites in southeast Oklahoma have been encoded as arcgis and spss datasets to allow some elements of statistical and geographic analysis. As discussed earlier, many of the sites were identified by Don Wyckoff, working for the Oklahoma River Basin Survey. Numerous sites were excavated and subsequently inundated by construction of Broken Bow and Pine Creek reservoirs in the 1960s and early 1970s. Their inundation resulted in the loss of data pertaining to specific landscape characteristics. The landscape analysis presented here is of a more general nature and less statistically based than that presented by Jami Lockhart (see chap. 11, this volume) on Caddo sites in southwest Arkansas. Of these Caddo sites, 259 (93 percent) are considered to be residential expressions of a constructed landscape, 8 (2.9 percent) represent a single mound with or without an accompanying residential occupation, 9 (3.2 percent) are multiple mound sites having the same structural form of mound present, and 2 (0.7 percent) are multiple mound sites with different structural mound forms. The numbers of mounds is a bit misleading in that there should be considerably more single mound sites than multiple mound centers. However, as many of the multiple mound Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

347

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

12-2. Distribution of Caddo settlements in Choctaw and McCurtain Counties, southeastern Oklahoma.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Copyright © 2012. Nebraska Paperback. All rights reserved.

sites represent the sequential building of a single mound structure at a given location, these are only quantitatively rather than qualitatively different from the single mound sites. The constructed cultural landscape has two different patterns in the built environment (fig. 12-2). In the northern portion of the study area (most of which is in McCurtain County), the sites are linearly distributed along the principal drainages of the Little, Glover, and Mountain Fork Rivers. In this portion of the region, these rivers are cutting through the Ouachita Mountains, forming rather circumscribed valleys with pronounced relief. Thus the valley floors are modest to small in extent and would have served as a constraint because of limited arable land for these farming people. Tributary streams to the principal drainages would have even more limited valleys and soils suitable for farming. Consequently there appear to be few Caddo sites in these tributary settings, and these may be the product of hunting or specialized use activities rather than farming. This cultural landscape is markedly different from that seen throughout much of the southeastern United States, where river systems have created wide alluvial valleys with soil fertility annually recharged by seasonal flooding (Smith 1978b). This pattern holds for some 42 percent (n = 117) of the Caddo sites identified in this study. The second pattern in the built environment occurs in the southern portion of McCurtain County where sites are more broadly dispersed across the natural landscape (fig. 12-2). This pattern is similar to that present in the southeastern United States. Most of the sites represent residential sites with at least one multiple mound location present. In this area, streams meander across a rather level landscape. As a consequence, soils as well as fauna and flora tend to be homogenously distributed, perhaps contributing to the prehistoric Caddo being less selective in their landscape choices for settlement. Roughly 58 percent (n = 162) of the sites occur within this area. Where the Little River turns markedly eastward (fig. 12-1) serves as the boundary between the two Caddo built environment patterns. Two factors account for this. The first is that this represents the ecological transition from the Ouachita Mountains to the Gulf Coastal Plain physiographic region, with less pronounced relief than that present in the mountains to the north. Associated with this change in landforms, the Glover and Mountain Fork Rivers and smaller tributaries flow into the Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

349

Copyright © 2012. Nebraska Paperback. All rights reserved.

Little River, forming a much wider floodplain valley. Even the smaller streams collect greater amounts of water and also form larger stream valleys in this area. This same phenomenon also holds true for the Red River on the other side of a divide or interfluve from the Little River. These larger alluvial valleys provided the Caddo farming communities with greater access to suitable farming soils and also more flexibility in the construction of their built environment. The more circumscribed valleys to the north may also have limited the ability of mountain-dwelling Caddo groups to construct mounds or mound centers. Multiple mounds as well as the layout of mounds around a central ceremonial space (such as at Grobin Davis) would require a wider stream valley to be practical. Building of larger mounds as well as greater numbers of mounds might also merit construction in valleys with more readily available alluvium for such construction. There are, however, contradictions to this suggestion. The Pine Creek Mounds site contains four house mounds of similar size, although an open ceremonial space is absent. This site is far up the Glover drainage where Pine Creek meets the Glover. The valley at this location is quite narrow, and the alluvial soils are relatively shallow. Perhaps such as consideration is only favored where multiple mounds of different structural types as well as ceremonial spaces are constructed. These two patterns, which conform to the Ouachita Mountains and Gulf Coastal Plain, mark a distinct departure from the study undertaken by Jami Lockhart in southwestern Arkansas (see chap. 11, this volume). In Lockhart’s study, the focus was on differential distributions of Caddo mound sites on the landscape within the Coastal Plain. Caddo sites in the Ouachita Mountains were not part of his analysis. Here, both physiographic regions were included to potentially contrast the distinctions that might exist in how the Caddo responded to different water, soils, and landform characteristics of the two regions. It also means that our analysis results are not directly comparable. However, in some cases, we did find similar patterning on the landscape. This means that it was likely that the Caddo selected a particular setting because it was readily available in both regions or because it was a more homogenous landscape setting regardless of other factors. The distinction in landforms also clearly has implications for social, political, and religious integration of the Caddo “polities” living in both 350

Brooks

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

environmental settings. The two multiple mound sites with multiple mound types, Grobin Davis and Clement, are both situated where they can serve as gateways (or gatekeepers) to the sites north of them on the Little and Glover Rivers, respectively. This is partially a function of the circumscribed valleys that exist on these rivers above the two mound centers. It can be argued that Grobin Davis and Clement served as the controlling centers for these valleys, at least from ca. 1000 until 1450. Sabo (chap. 15, this volume) examines the cosmology of the Caddo from the perspective of the 1691–1692 Terán map and suggests that the platform mound and the residence of the Xinesi served as an entrance point to the Nasoni village, at least for some religious-political events. Perhaps the Grobin Davis and Clement mound centers performed a similar function on a larger scale, serving as religious/political entrance points for those visiting the upstream polity. The absence of such a center on the Mountain Fork is perplexing. Perhaps the presence of the “narrows” disrupted the need for such a pattern. It is more likely, though, that such a center has escaped detection, as this area has not been extensively examined. Here the model proposed by Hally (1993) of mound centers spaced some 8 km apart is inappropriate for such a linear distribution of Caddo settlements and mound centers. To the south of where these centers are located, the more open and less circumscribed landforms were not as conducive to this type of control. This may be why there are no multiple mound sites with different types of structural mounds in the southern part of the area. This could also reflect social, political, and religious changes among the Caddo, as many of the sites there were occupied later (from ca. 1200 until 1750). Certainly residential sites, sites with single mounds, as well as those with multiple structural mounds appear to be more uniformly distributed across the landscape, although there are some instances where a multiple mound site served as the locus for a notable concentration of Caddo residential sites. The Association of Soils and the Caddo Cultural Landscape Soils serve an important role in the decisions of prehistoric societies concerning the constructed cultural landscape. The permeability of soils can determine whether a society establishes their settlement in damp, marshy locations or one that is dry and well drained. Marshy soils would also encourage other less than desirable conditions, such as increased Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

351

Copyright © 2012. Nebraska Paperback. All rights reserved.

numbers of pests, mold, and mildew, and a suitable medium for disease. Soils high in clay content have a low permeability index and would not be selected for settlement based strictly on residential usage criteria. However, subsoils with high clay content are often brightly colored (red, orange, and yellow). These soils would be sought for use in construction of nonresidential landscapes such as mounds and ceremonial grounds (plazas). Use of clay subsoils in the manufacture of ceramic goods is an additional factor that may have influenced settlement locational choices. Another important consideration is soil fertility. Obviously, prehistoric farming societies desire fertile soils, but nonfarming groups may also have sought out more fertile soils due to their increased natural productivity, leading to a high diversity in plant communities. The plant communities would attract a variety of fauna, many of which would have been hunted for food by those prehistoric societies. In this study, the value of fertile soils is thought to have been an important factor in the choice of location of the constructed cultural landscape, especially residential locations. The presence of soils with a high suitability index for agriculture and also having an extensive distribution along the stream and river valleys should have been a deciding locational factor for prehistoric Caddo farmers. These soils would also have been annually renewed by seasonal floods (cf. Smith 1978b). There were also potential choices made in the selection of soils where their agricultural potential was of secondary importance. If the constructed landscape was principally ceremonial, then agriculturally fertile soils would be of less interest. Perhaps ease in the excavation of borrow for mound construction or certain soil color schemes may have played a role in constructed landscape choices. Or perhaps soils themselves were not as important for the location of a ceremonial place in comparison with other natural and cultural landscape features. It may have been situated strategically to provide the elites in residence more of a commanding presence (physically or ritually) or placed near a prominent feature in a conceptualized or ideational landscape. The only means of discerning whether one or another of these factors may be playing a role in the Caddo’s selection of specific soils is by a study of the distribution of the constructed landscape against the backdrop of soils found in southeastern Oklahoma. The 279 sites are found on seven different soil associations in McCurtain and Choctaw Counties 352

Brooks

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 12-1. Distribution of Caddo sites by soil association Soil association Missing AP BCB DST ECH HP KCB MYT

Copyright © 2012. Nebraska Paperback. All rights reserved.

Total

Residential

Single mound

Multiple mound

Total

14 30 33 10 8 78 37 49

1 1 2 0 0 4 1 1

0 0 3 0 0 2 3 1

15 31 38 10 8 83 41 51

259

10

9

278

(table 12-1). Two soil associations are poorly represented: the DurantSan Saba-Tarrant (dst) and the Ender-Hector-Conway (ech). The Durant–San Saba–Tarrant soils occur on the divide between the Little River and the Red River, and the Ender-Hector-Conway soil association is present in the northernmost portions of McCurtain County. Neither of these soil associations was selected for mound construction, and there were only a small number of residential sites present on these soils. The remainder of the soil associations appear to have been widely selected for the Caddo constructed landscape. However, for my purposes, these associations can be further combined into four groups. The Atkins-Pope (ap) and Miller-Yahola-Teller (myt) associations represent nearly level, deep sandy to clayey bottomland soils developed under a lowland hardwood forest. These soils occur along eastern portions of the Little River valley and along the Red River. The Bowie-Caddo-Boswell (bcb) and Kirvin-Cuthbert-Bowie (kcb) associations are light colored, acidic, and sandy soils developed under a pine-oak forest. These soils cover an extensive area in the Little River drainage basin, and a large number of Caddo residences and mound sites are situated on this soil group. The two multiple mound sites that contain structurally different mounds (Grobin Davis and Clement) are situated on kcb soils. EndersConway-Hector (ech) and Hector-Pottsville (hp) soil associations represent light-colored acid sandy to loamy soils developed on sandstone and shale under a pine-oak-hickory forest. These soils are most prevalent in the upland valleys along the Glover and Mountain Fork Rivers above their confluence with the Little River. The remaining soil association, Durant-San Saba-Tarrant (dst), is composed of dark loamy to Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

353

Table 12-2. Soil associations and the constructed landscape Soil associations AP/MYT BCB/KCB DST ECH/HP Missing Total

Residential

Single mound

Multiple mound

Total

79 70 10 86 14

2 3 0 4 1

1 6 0 2 0

82 79 10 92 15

259

10

9

278

Copyright © 2012. Nebraska Paperback. All rights reserved.

clayey soils that developed on marine clay/limestone under tall grasses. As these soils have poor permeability and are also difficult to work when dry, it is perhaps not surprising that they were not highly selected by southeastern Oklahoma Caddo farmers. These combined soil association categories and the tallies for the different kinds of constructed landscape features are presented in table 12-2. A χ2 analysis of these distributions yielded a statistically significant difference at the .001 significance level. There are some obvious soil preferences if table 12-2 is deconstructed. Almost 70 percent of the multiple mound sites are situated on bcb/kcb soils, including the previously mentioned Grobin Davis and Clement mound centers. Approximately 50 percent of all the Caddo mound sites are also located within this group. The soils closer to the Red River (ap/myt) have a large number of residential sites but only a few mound sites (30.5 percent of all the sites occur within the soil group, and 15.8 percent of all the mound sites in southeastern Oklahoma). This is also true for the ech/hp soil associations. These soils contain 33 percent of the residential sites, 31.5 percent of all the known mounds, and only 22 percent of the multiple mound sites. From these data, it would appear that the bcb/kcb soil associations certainly figure more prominently in the constructed landscape of the Caddo in southeastern Oklahoma, at least with respect to the construction of single and multiple mound centers. This may strictly be a function of soil qualities, or there may be other factors that, combined with the soils, made these settings more attractive to Caddo peoples. These results closely follow the soil preferences noted by Lockhart (see chap. 11, this volume) on mound sites in southwestern Arkansas. One notable difference here is the selection of soils associated with an oak-hickory-pine forest along the Mountain Fork and Glover 354

Brooks

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 12-3. Distribution of the Caddo constructed landscape by landform Landform

Residential

Single mound

Multiple mound

Total

Floodplain Terrace Hillside Dissected Undissected Missing

32 187 11 17 1 10

1 6 0 2 0

4 5 1 0 1

37 198 12 19 2

Total

248

9

11

268

Rivers for both residential places and mound centers. But his data set excluded the Ouachita Mountains Caddo sites, and this study uses data from both regions. To determine whether the Oklahoma and Arkansas data are comparable would require either isolation of one segment of the Oklahoma data or having the Arkansas dataset include the Ouachita Mountain Caddo sites.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Landform and the Caddo Cultural Landscape As the Caddo were a farming people, it would be expected that they would occupy the valley lowlands where fertile agricultural soils are present. However, this is somewhat complicated by the more circumscribed valleys present along the north-south flowing portions of the Little River and on the Glover and Mountain Fork Rivers. A landscape with fewer prominent terraces or other arable landforms (i.e., natural levees) may have been less favored for a constructed landscape. The nature of the constructed landscape also may have played a role in landform selection. Residential sites with mounds or multiple mound sites that were ceremonial centers may have required different, religious-based landscape settings. Mound placement within such a religious context would also be expressed through the viewscape of the site residents (e.g., Vogel 2005b). Landform categories used in this landscape analysis include floodplain, terrace, hillside, dissected uplands, and undissected uplands. The use of these more general categories was necessary because of the nature of the available data in the state site files; 219 of the residential sites (88 percent) are situated on terrace or floodplain settings, while an additional 16 single or multiple mound sites (80 percent of the mound Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

355

FIGURE CAPTIONS

12-3. Elevation of Caddo settlements by built environment.

Copyright © 2012. Nebraska Paperback. All rights reserved.

sites) are found on these landforms (table 12-3). Approximately 91 percent of the Caddo constructed landscape is situated in the stream valleys based on these tabulations. There does not appear to be a distinction in this regard between residential sites, residential sites with a single mound, or multiple mound sites. The two sites that contain multiple but structurally different kinds of mounds, Clement and Grobin Davis, are on the floodplain and a terrace, respectively. However, the setting of the Clement site could easily be seen as a terrace; here we may be dealing with a landform coding error. It is likely that much of the Caddo constructed landscape is homogeneous by landform choice, with perhaps a critical but unrecorded variable being the average elevation relief of particular sites above the stream or river. A number of single and multiple mound sites are situated on higher terrace elevations relative to the floodplain (e.g., Woods Mound Group). However, this is not true of all mound sites. For example, the Pine Creek Mounds are situated in the floodplain immediately adjacent to Pine Creek, with only modest relief from the stream channel (although two of the three mounds are on the higher terrace area). The 356

Brooks

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 12-4. Distribution of Caddo sites by slope Slope (%) 0-5 6-10 11-25 26-50 >50

Copyright © 2012. Nebraska Paperback. All rights reserved.

Missing

Frequency

Percentage

46 19 6 4 0

16.5 6.8 2.2 1.4 0

204

73.1

kind of landform, it would appear, was a critical consideration for the Caddo in the placement of their constructed landscape, but they did not selectively choose a certain landform for ceremonial use with mounds and another for residential settings. This is somewhat confirmed by the Terán map of 1691–1692. The map illustrates an upper Nasoni Caddo village where the common residences, the Caddi’s residence, nonresidential features, and the mound all appear to be on the same landform (although the mound is separated from the residential area of the community), namely, a natural levee/point bar along the Red River (see Sabo, this volume; Perttula 2005). The distribution of the Caddo cultural landscape by elevation also demonstrates that the number of residential occupations and mound sites increases as elevation decreases (fig. 12-3). The higher elevations are present in the upper reaches of the Glover and Mountain Fork Rivers. While there are Caddo residential and mound sites present in these areas (e.g., Pine Creek Mound Group), there was probably a greater interest in establishing a presence at lower elevations where the valleys were wider and there were deeper and more fertile soils. Another feature closely related to landform is the degree of slope of the landform and the cardinal directional orientation of the landform. Numbers available for this analysis (ca. 75; table 12-4) are substantially smaller than the total Caddo dataset because most of the sites recorded prior to the 1980s did not have such information noted and it cannot be easily derived from topographic maps. Additionally, many of the sites are submerged beneath man-made lakes in Choctaw and McCurtain Counties. An analysis of the constructed landscape in respect to the degree of slope revealed no pattern of significance (i.e., the χ2 test for their being a difference in the slope direction yielded a significance Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

357

Table 12-5. Distribution of Caddo constructed landscape by slope direction Slope direction

Residential

North Northeast East Southeast South Southwest West Northwest All directions

4 5 19 12 18 8 11 4 6

Total

87

Single mound

Multiple mound 1

1 1 1

1 1

4

Copyright © 2012. Nebraska Paperback. All rights reserved.

value of 0.995, meaning that there is virtually no association between the degree of slope and whether the site was a residential occupation, a single mound site, or a multiple mound site), as 86 percent of the Caddo sites where the degree of slope was recorded were on a landform where the inclination/declination was 10 percent or less. This is consistent with these sites being situated on a floodplain or terrace within an alluvial valley. The analysis of the relationship between the constructed landscape and slope direction was equally unrevealing. Although the difference in numbers by directional orientation is relatively minor, it would appear that there is a slightly higher preference for landforms with slopes oriented to the east and to the south (table 12-5). A χ2 significance score of 0.734 demonstrated no association between residential, single mound, and multiple mound sites and the slope direction. Based on these results, there do not appear to be particular aspects of placement of a town, village, or mound where the lay of the land is an important consideration. This consistency, however, may be reflective of a conceptual/ ideational perception of natural landscape features where a preference for specific landforms within the valley as well as directional view is valued equally. This conforms closely to Lockhart’s results in southwestern Arkansas (see chap. 11, this volume), where slope is often nearly level. Keep in mind that slope for the Oklahoma Caddo sites is generated from inundated sites in many cases where slope direction is a function of the trend surface of the landscape on the topographic map rather than a calculated value from a recent surface examination. Even when the Okla358

Brooks

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 12-6. Distribution of Caddo constructed landscape by nearest water Water Source

Residential

Single Mounds

Multiple Mounds

Permanent stream Intermittent stream Permanent spring Seep Natural Lake River Slough Relic stream Well/cistern*

102 47

4 2

7

Total

242

5 1 51 33 2 1

3

1 1

9

9

* Coding a prehistoric site for use as an historic well is an error although it does not significantly affect the outcomes of the analysis.

homa sites have a slope direction, this is frequently derived from a slope of 5 percent or less and should be viewed as essentially level.

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Caddo Constructed Landscape and Water Water undoubtedly played a critical role in the distribution of Caddo places on the natural landscape. Availability of potable drinking water was a critical resource for towns, villages, and ceremonial centers. However, the distance that people are willing to travel for water varies depending on a number of factors, including but not limited to the natural abundance of rainfall, landform character, water quality, and potential hazards. Beyond the use of water for drinking and bathing, larger streams serve as means of transportation. The Caddo are known to have used canoes for travel, and the streams and rivers in southeastern Oklahoma would have served as a reliable means of travel except during the driest periods and times of extreme flooding. An analysis of the nature of the nearest water source relative to the constructed landscapes of the Caddo illustrates that there was an emphasis on permanent streams and rivers (64 percent), while intermittent streams and sloughs were also widely used (table 12-6). There does not, however, appear to be a distinction between residential sites and single or multiple mound locations in the selection of water resources (χ2 significance = .091). It is interesting that many more of the multiple mound sites were situated near permanent streams than adjacent to the larger rivers. However, many of the multiple mound sites were situated where Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

359

a permanent stream joins with a larger river. In these cases, the mound site may be nearest to the permanent stream, but the river is also in close proximity. This pattern may reflect a conscious effort on the part of the Caddo to avoid flood situations, especially on the lower, more eastern portion of Little River and in most areas adjacent to the Red River. These settings are known for flooding and hold the flood deposits for some time. Distance to the nearest water source was calculated for recently recorded Caddo sites in the southeastern Oklahoma study sample. Thus mean distances are highly skewed and are not representative of the total sample. Residential sites are approximately 100 m from their nearest source of water, single mound sites are 120 m from water, and multiple mound sites are 180 m distant. Although these data may not be representative of the entire Caddo site database in the region, this pattern is intriguing. Multiple mound centers are larger sites and tend to be located in wider alluvial valleys. So it is not unreasonable that larger multiple mound-constructed landscapes would be more distant from the nearest water source. There are exceptions as well: the Pine Creek Mounds are located directly adjacent to Pine Creek within a relatively narrow valley segment, although the mound center is not far from the confluence of Pine Creek and the Glover River. Conclusions

Copyright © 2012. Nebraska Paperback. All rights reserved.

This chapter has considered whether the study of the constructed landscape of the prehistoric Caddo can detect differences in setting based on whether the site was a residential settlement (farmstead, hamlet, and town/village), a single mound accompanying a residential community, or a multiple mound center. A distributional study of prehistoric Caddo sites on the natural landscape of Choctaw and McCurtain County indicates that sites are more broadly dispersed along tributary streams as well as on higher ground separating the Little River drainage and the Red River drainage. The linearity in the distribution of these Caddo sites is highly reminiscent of the settlement pattern in Mississippian polities in the southeastern United States (Hally 1993). Of particular note is the strategic placement of the Grobin Davis and Clement sites where they could effectively mediate the flow of information and goods on the Little and Glover Rivers, respectively. 360

Brooks

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Examination for differences in the placement or setting of residential places versus the constructed landscape of mounds and multiple mound centers offered few significant distinctions. Features of the natural landscape such as soils, landform, nearest source of water, elevation, and slope direction were used in this analysis. The Caddo constructed landscape is homogeneous with no difference in site placement based on their function across the landscape. However, it has been demonstrated that the Caddo favored certain soils that were agriculturally productive, lived on the floodplain and terraces of major streams and rivers, and were more or less linearly distributed along these streams. This is a pattern of Caddo settlement that endured well into the seventeenth century. Although no revealing distinctions were noted in the patterns of the Caddo constructed landscape, does this mean that none existed? This is probably not the case. As noted elsewhere, this study has been sorely limited by the absence of temporal refinement in assessing the periods and span of occupation at many of the Caddo sites; this has hampered component period identifications. Distinctions may lie hidden in these refinements that we cannot currently explore. This is especially true for the periods of significant change that took place from ca. 1200 to 1750. Such refinements, unfortunately, must await considerably greater knowledge of the Caddo sites in southeastern Oklahoma. It is also possible that the “special” setting of the Caddo constructed landscape was applied to both residential and ceremonial places. There may be additional factors beyond the conceptual/ideational view of the landscape that resulted in some places existing as residential locales and others as mound centers. As a number of sites exhibit evidence of residential use as well as mound construction, this further corroborates some concept of religious practices that did not serve to distinguish a particular natural setting for one set of practices as opposed to more mundane practices. There are undoubtedly ceremonial centers such as Grobin Davis and Clement that functioned strictly as ceremonial places and were probably occupied only by priests and their attendants, but apparently these are limited in number among the Caddo societies that lived in southeastern Oklahoma. They also appear not to function in this fashion during the latter portion of the Late Caddo period (Perttula 1992). In fact, Caddo ceremonialism as reflected in the constructed cultural landscape seems to have become less dynamic and less pervasive in historic times. But, Landscape Selection of the Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

361

Copyright © 2012. Nebraska Paperback. All rights reserved.

as demonstrated in Sabo’s insightful study (chap. 15, this volume) of the Caddo settlement (the Hatchel site) depicted in the 1691–1692 Terán map, a Caddo community’s cosmological orientation reads much like a constructed sacred landscape. This structure is undoubtedly present at all Caddo residential and ceremonial places and may explain why there is little to distinguish residential places from those that also hold mounds or multiple mounds.

362

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

13. The Character of Fifteenthto Seventeenth-Century Caddo Communities in the Big Cypress Creek Basin of Northeast Texas timo thy k . perttula

Copyright © 2012. Nebraska Paperback. All rights reserved.

Introduction The Caddo peoples living in a number of communities in the Big Cypress Creek basin in northeastern Texas (fig. 13-1a) during the fifteenth through seventeenth centuries were a strong and socially powerful group in their region. They were farmers who lived in dispersed communities of individual to multiple farmstead compounds, and they were active traders. These groups were among the most populous and socially complex of the many Caddo societies living at that time in the Caddo archaeological area (e.g., Early 2004:573), as suggested by the earthen mounds they built within a number of political communities in the basin, and they were also the westernmost aboriginal group in Texas that was in any sense sociopolitically akin to middle and late Mississippian polities living in the Southeast and the Midwest (Milner et al. 2001; Butler and Welch 2006; Blitz 2010). These Caddo communities living in the heartland of the Big Cypress Creek basin (fig. 13-1b) in the Pineywoods (Diggs et al. 2006) experienced rapid and sustained population growth during times of fluctuating climatic conditions. These dynamic farming communities dealt with local climatic and subsistence stresses by effecting new means of holding their societies together, successfully and boldly forming several stronger communities within a larger regional political system than had existed before in the region, centered around the establishment of a series of small mound centers, community cemeteries, and villages.

363

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

13-1a 13-1. The distribution of the Titus phase: (above) a. extent of the Titus phase in northeast Texas, and (right) b. Titus phase components and the location of other Late Caddo period phases.

The Titus Phase

Copyright © 2012. Nebraska Paperback. All rights reserved.

Archaeological evidence indicate that many of the components represent permanent, year-round, settlements of horticultural or agricultural peoples. These Caddo sites belong to the Titus phase, an archaeological unit of long-standing taxonomic definition (i.e., Suhm et al. 1954; Turner 1978; Thurmond 1990; Perttula 1992) with temporal and spatial contiguity, as well as temporal depth (ca. 250 years). The fifteenth- to seventeenth-century archaeological record in the basin “refers to a number of distinctive socio-cultural groups, not a single Caddo group; these groups or communities were surely related and/or affiliated by kinship, marriage, and social interaction” (Perttula 2005, ed.:401). The social and cultural diversity that probably existed among Titus phase cultural groups is matched by the stylistic and functional diversity in Titus phase material culture, particularly in the manufacture and use of fine ware and utility 364

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

13-1b

ware ceramics, ceramic pipes, and arrow points. (The ceramic tradition is the surest reason for evaluating attribution of archaeological components to the Titus phase.) It is the character of their stylistically unique material culture (see Parsons 2010; Perttula 1998, 2000, 2005, ed.:401–410; Thurmond 1990), coupled with the development of distinctive mortuary rituals and social and religious practices centered on the widespread use of community cemeteries and mound ceremonialism as means to mark social identities, that most readily sets these Caddo groups apart from their neighbors in east Texas and in the Red River basin to the north and east (Early 2004:568), as well as to non-Caddo peoples living in the Blackland Prairie to the west and Gulf Coastal Plain to the south. Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

365

Community Setting

Copyright © 2012. Nebraska Paperback. All rights reserved.

These east Texas Caddo peoples chose to settle and build structures and other facilities (i.e., granaries and open-air ramadas and arbors) in habitats where suitable sandy soils could be worked with simple wood and bone digging tools. The land where they built and maintained their homesteads had to be well drained and elevated above the annual floods along Big Cypress Creek and its tributaries. They had to live in areas where wood and grass were plentiful for house construction and refurbishing, alluvial clays were abundant for ceramic vessel manufacture, and fresh drinking water was available. The fact that most Titus phase settlements are not found in any notable spatial clusters or aggregated farmstead compounds (as is seen on the Red River at this time, cf. Perttula et al. 2008; see also Walker and McKinnon, chap. 7, this volume) suggests that the many resources that were needed by the sedentary Caddo populations to sustain themselves in the Big Cypress Creek valley could best be exploited by dispersing the groups in a variety of settings. This dispersed settlement arrangement would have helped lessen the competition for available plant, animal, and mineral resources and would not have allowed for the environmental degradation of those suitable habitats by a single large community. This would have been of critical concern in droughty periods for Caddo peoples living along the margins of the Pineywoods, in environmentally risky areas. This dispersion of population would also have permitted the Caddo peoples to take advantage of the diversity in habitats to exploit a number of them from season to season and year to year, thus ensuring that the overall community or communities of interacting farmsteads and individuals living in those farmsteads could survive if there were economic difficulties or failures (i.e., local droughts, flooding, fires) in some habitats but not in most of the others. In Late Caddo Titus phase times, when the Caddo peoples had a diet that primarily (i.e., possibly as much as 50 percent of the diet) consisted of cultivated plants like maize, beans, and squash (see Wilson, chap. 4, this volume; Perttula 2008), agricultural pursuits must have been of particular importance in determining the location of individual farmsteads and hamlets, more so than they were in the Early or Middle Caddo periods (ca. A.D.1000–1400) when Caddo peoples were not apparently quite so dependent upon cultivated plants for their diet (or at least that seems to be the case until ca. A.D. 1300–1350, see discus366

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

sions in Perttula [1996, 2008] and Dering [2004]). How are these constraints reflected in the spatial distribution of Late Caddo Titus phase sites? What is seen in this part of the Big Cypress Creek basin is that the overall settlement pattern was dispersed (with generational residential mobility), and this dispersion occurred in conjunction with a heightened emphasis on situating sites along the secondary streams and the spring-fed branches (see fig. 13-1b). These areas may have had more dependable water, or more accessible water, at least at that time, and it is also likely that fields would have been easier to clear along the more open upland forests than if fields had to be located in the more mesic and thickly wooded valleys. There are more Late Caddo Titus phase sites in this part of the Big Cypress Creek basin than earlier sites (cf. Thurmond 1990:table 63; Perttula 2005). This suggests that the regional population was quite a bit higher during the Late Caddo period (all things being equal, especially the length of time each community may have been occupied), and there are several clusters of settlements that apparently represent parts of contemporaneous small communities. A political community as used here refers to a cluster of interrelated settlements and associated cemeteries that are centered on a key site or group of sites distinguished by public architecture (i.e., earthen mounds) and large domestic village areas. The key sites within these Titus phase political communities include places such as Lower Peach Orchard (41CP17) on Big Cypress Creek in the vicinity of the Sandlin Dam community cemetery; Pilgrim’s Pride (41CP304) on Walkers Creek; Tom Hanks (41CP239), Harold Williams (41CP10) and Tuck Carpenter (41CP5) on Dry Creek (Turner 1978, 1992); Sam Roberts (41CP8), P. S. Cash (41CP2), and the Shelby (41CP71) sites on Greasy and Prairie creeks; Harroun (41UR10), Dalton (41UR11), Chastain (41UR18), and Camp Joy Mound (41UR144) on Big Cypress Creek and Meddlin Creek; and the Whelan (41MR2), H. R. Taylor (41HS3), and Pea Patch (41HS825) sites on Big Cypress Creek and Arms Creek (fig. 13-2; see also Perttula 2004:figs. 13.30, 13.31). In the upper part of the Big Cypress Creek basin, there are several clusters of Titus phase settlements within political communities. The Lower Peach Orchard site (41CP17) is probably one of the most important in the mid-reaches of the valley, since it had a number of deep shaft tombs of the social and political elite as well as extensive settlement deposits (Thurmond 1990; Perttula 1998, 2004); it appears to be a village Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

367

Copyright © 2012. Nebraska Paperback. All rights reserved.

13-2. Titus phase Caddo political communities in the Big Cypress Creek basin of northeast Texas.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

nexus comparable to that documented around the Pilgrim’s Pride site, about 10 miles downstream. Other concentrations of Titus phase sites in this area are noted along Brushy Creek upstream from its confluence with Big Cypress Creek and in upland/valley margin settings (Perttula and Nelson 2003c). Parsons (2010) has identified a similar community in the Little Cypress Creek basin, south of the Titus phase heartland. There are other important villages in Titus phase political communities in the Titus phase heartland, and they appear to be situated in similar topographic settings, namely, along tributary streams near their confluence with Big Cypress Creek. They are marked by a higher density of permanent settlements around one premier community. In the case of the Shelby site, the social and political center of the community stretches for several hundred meters along Greasy Creek and a small tributary, with an earthen mound at the northern end of the village and a large cemetery at its southern end (fig. 13-3). Domestic village areas occur between the one mound and the cemetery and cover at least 10–15 acres (Perttula et al. 2004). The Titus phase earthen mound covered a burned structure at the base, and a second structure had been built that stood on the mound itself and then was burned and capped with a final sandy fill. The arrangement of mound, domestic areas, and planned cemetery here is essentially duplicated at the Pilgrim’s Pride site (Perttula 2005:fig. 11-2), although the village areas and the size of the cemetery at the Shelby site are considerably larger. Based on work at the site in 2002, the north levee area at the Shelby site has thick midden deposits and evidence of several burned structures, implying the existence of an intensive occupation here. Another important community nexus in the Titus phase heartland includes the Tuck Carpenter and Harold Williams community cemeteries (see Turner 1978, 1992; Turner et al. 2002) and various domestic settlements (marked by midden deposits and habitation debris) on the lower part of Dry Creek and Swauano Creek (see fig. 13-2). No Titus mounds have been found in this area yet, although the Tom Hanks mound site (41CP239) is located in the upper part of the Dry Creek basin and may be part of this particular political community. This area is on the first important set of eastward-flowing tributaries to Big Cypress Creek that are downstream from Walkers Creek. The next downstream Titus phase village community may be centered at the afore-mentioned Greasy Creek Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

369

Copyright © 2012. Nebraska Paperback. All rights reserved.

13-3. The important Titus phase village and political community center at the Shelby site (41CP71) on Greasy Creek.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

community at the Shelby site; there are numerous Titus phase settlements and large cemeteries on Greasy Creek and its tributaries. Also probably part of this particular Caddo community are the Titus phase mound sites (single mounds) on Prairie Creek—only a few miles away from the Shelby site—at the P. S. Cash and Sam Roberts sites. The next community nexus is in the Meddlin Creek and Big Cypress Creek areas, midway between Greasy Creek and Arms Creek. It includes three or four mound sites, namely Harroun (41UR10), Dalton (41UR11), Chastain (41UR18), and Camp Joy (41UR144), various small domestic settlements in valley and upland settings, and several large community cemeteries (see fig. 13-2). As evidence of the integration of these mound centers, there are extended entranceway structures at the Harroun site that point toward each other (in the case of mounds B and D), and another (on mound C) is oriented to face an extended entranceway structure at the nearby Dalton site (Thurmond 1990:fig. 27). In turn, the extended entranceway structure at the Dalton site points southwest toward a mound platform (with burned structures and burials) at the nearby Camp Joy Mound site (41UR144, Perttula and Nelson 2001). The Camp Joy, Dalton, Chastain, and Harroun sites may in fact form individual portions of a single multiple mound center used by a larger community in the Titus phase heartland. Each of these sites have mounds built over wooden structures that probably had a special religious and ritual purpose, and the earthen mounds were built up over the structures after they had been burned. Here the community cemeteries are not found in close association with the mound centers, as they are in the Greasy Creek Caddo community, but instead they are situated along Big Cypress Creek and its tributary streams, presumably in general proximity to the many farmsteads that must have been dispersed across the countryside at that time. The last recognizable Titus phase political community in the Big Cypress Creek heartland is along Arms Creek and Big Cypress Creek. It includes the community nexus at the Whelan (41MR2) mound site (see fig. 13-2). The Whelan site, with four earthen mounds (Davis 1958), is on the floodplain of Big Cypress Creek, and there are farmstead settlements along the creek as well as along Arms Creek. The larger community cemeteries at the H. R. Taylor (41HS3) and Pea Patch (41HS825) sites are situated a few miles from the Whelan site, near the headwaters of Arms Creek. Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

371

Copyright © 2012. Nebraska Paperback. All rights reserved.

Other than the village centers, how were other Titus phase domestic settlements (i.e., the small settlements recognized by Thurmond 1990) organized and laid out spatially? Fortunately, recent excavations at two Titus phase settlements—the Ear Spool and Rookery Ridge sites (see Galan 1998; Parsons 1998, 2010; Perttula and Sherman 2009)—have provided solid archaeological evidence of their character. The settlements appear to have been composed of one to several family units with house midden/daub concentrations and trash midden deposits; many activities occurred outside the house areas, resulting in trash-filled pits, hearths, and posts in these areas. At the Ear Spool site (41TT653) there were four circular structures (probably thatched and daub-covered), one with an extended entranceway pointing toward the other three houses, that may have been from two temporally different and sequential Titus phase occupations (see Perttula and Sherman 2009:fig. 6-1). There was a broad, open area between the houses that may have been a small plaza or courtyard, as there are marker poles at its center, and there were clusters of pits along its margins. Three burials were present at the site: one child burial inside structure 2 and adults in two widely separated areas. There apparently was no large family cemetery (i.e., 10–20 burials) at the Ear Spool site. The Rookery Ridge site (41UR133) excavations exposed two early Titus phase circular structures and extensive midden deposits. The middens were about 15 m south of the one structure with an extended entranceway; the entranceway faced north, suggesting that other habitation features besides those excavated by Parsons (1998) were present on the northern part of the alluvial landform along Kelsey Creek (Parsons 2010). Child and adult burials were present either inside a structure or immediately outside, in an earthen embankment along the structure walls, but again, there was no larger family cemetery at the site. Using several lines of evidence, particularly the occurrence of well-preserved burned house debris and the placement of several child burials just outside the walls of the extended entranceway structure, Parsons (2010) suggests that this structure had a soil berm around it, and when the structure was burned, it was covered with a very low mound of relatively clean sediments. Perttula and Nelson (2003c:34) have noted that Late Caddo Titus phase sites are more common to the south of Big Cypress Creek than they are on the north side of the basin or in other stream valleys (such 372

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

as Little Cypress Creek, White Oak Creek, or streams in the Lake Fork Creek basin) in the Big Cypress, Sulphur River, and Sabine River basins. Poorly drained and steeper, rockier landforms on the east side of Big Cypress Creek were also apparently not heavily settled by these Caddo peoples. Regional settlement data for the Titus phase does suggest that this pattern in the spatial distribution of sites may be part of a much broader trend in the overall density of Late Caddo sites between the Titus phase heartland and outlying areas also occupied by Titus phase Caddo peoples (see Perttula 1998, 2004; Nelson and Perttula 2003a). That trend indicates that Titus phase sites, as well as Titus phase sites with mounds and large community cemeteries, are more common across the landscape and through time from the Brushy Creek to Lake Bob Sandlin dam area downstream along Big Cypress Creek than they are in the Post Oak Savannah immediately north and northwest of Big Cypress Creek. This distribution of settlements, mounds, and community cemeteries strongly suggests that the density of Caddo peoples during the Titus phase was more concentrated in the Big Cypress Creek heartland, including its many southward-flowing and eastward-flowing tributaries, than it was elsewhere across the landscape. Nevertheless, other parts of the Titus phase Caddo homeland were also well settled: two such areas are the Dry Creek and Caney Creek valleys in the upper Lake Fork Creek valley in the Post Oak Savannah (see also Bruseth 1987; Bruseth and Perttula 1981; Perttula et al. 1993). Here at sites like Burks (41WD52), Steck (41WD529), Spoonbill (41WD109), or Goldsmith (41WD208), the farmstead occupations have house and trash midden deposits, apparently from two to four structures and nearby family cemeteries with roughly 5–15 individual interments (Perttula 2004:fig. 13.28). The density of Titus phase settlements along Caney Creek is impressive, with more than 50 components on a ca. 7 km stretch of the creek and the adjacent upland landforms. The majority of the sites are in the uplands, rather than in the Caney Creek valley or on its smaller tributaries (Perttula, ed. 2005:fig. 11-6). Habitation sites are well dispersed across the landscape, as are the habitation sites with reported cemeteries. The village–mound–large cemetery association noted for Titus phase heartland community centers, however, seems to be absent in the Caney Creek cluster and other nonheartland settings, although overall population densities from one locale to another may have been comparable. A third area on Kelsey Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

373

Creek, in the Little Cypress Creek basin, was thickly settled by the Caddo in the fifteenth century A.D. (Parsons 2010), but not afterward. Age of the Titus Phase

Copyright © 2012. Nebraska Paperback. All rights reserved.

At the present time, there are only a few Titus phase components that are well dated by absolute dating techniques. There are 17 Titus phase components that have a reliable set of radiocarbon and Oxidizable Carbon Ratio (ocr, see Frink 1992; Feathers 2008:178–179) dates, and in each case, the maximum age ranges extend from ca. A.D. 1400 to 1650. In 90 percent of the sites, the age ranges of the best-dated Titus phase components actually extend from ca. A.D. 1430 to 1640.1 As table 13-1 shows, radiocarbon dates from Titus phase components consistently range from the early fifteenth century to the late seventeenth century. The most reasonable (i.e., with probability distributions greater than 0.76) of the numerous recent calibrated radiocarbon dates from a variety of Titus phase domestic, mortuary, and mound contexts consistently span the period from cal A.D. 1431 to 1680. Besides the Pilgrim’s Pride site (Perttula 2005, ed.), the betterdated Titus phase domestic contexts include habitation areas on the north levee at the Shelby site (1390–1635) (see fig. 13-3), the Ear Spool components (1415–1660), 41UR118 (1430–1679), and Rookery Ridge (1435–1679). With respect to the dating of Titus phase burials (primarily the dating of organic residues on ceramic vessels placed as funerary objects in the burials), a burial from Tuck Carpenter (Turner 1978) dates at one sigma to cal A.D. 1536–1635, while one dated burial at the Harold Williams site (see Turner et al. 2002) has a calibrated date range of 1460–1630, and several from the Mockingbird site (Perttula et al. 1998) range from 1437 to 1667. At the Alex Justiss site (41TT13), two of three dates on vessel residues range from 1430 to 1640, although on other grounds, the cemetery there may date from the late sixteenth century to the mid-seventeenth century (Rogers et al. 2003). A calibrated 2 sigma age range of 1470–1630 has been obtained on the clay paste of a red-slipped engraved deep bowl from the qlbs site in the upper Sabine River basin (Cast et al. 2006:table 1). Recently obtained dates from village areas at the Shelby mound span the period from 1322 to 1631 (with the early end of the temporal span being dubious, but coming from a sample of charred material from a context below the mound itself ), and those from the Pilgrim’s Pride site range from 1414 to 1636. The Camp 374

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 13-1. Titus phase radiocarbon dates Site name/site number Tuck Carpenter/41CP5 Sam Roberts/41CP8 Harold Williams/41CP10 Shelby Mound/41CP71

Kitchen Branch/41CP220 Underwood/41CP230

Copyright © 2012. Nebraska Paperback. All rights reserved.

Pilgrim’s Pride/41CP304

41CP313 41CP316

Conventional Calibrated age age range* (A.D.) 360 ± 70 320 ± 60 240 ± 90 180 ± 40+ 390 ± 60 540 ± 60 510 ± 50 350 ± 50 410 ± 50 380 ± 50 470 ± 40++ 430 ± 40 460 ± 70 320 ± 50 340 ± 50 360 ± 40 430 ± 90 410 ± 30 180 ± 70++ 320 ± 50 330 ± 50 140 ± 60++ 60 ± 60++ 80 ± 60++ 230 ± 50 540 ± 40 400 ± 50 470 ± 40 380 ± 60 480 ± 90 530 ± 70 370 ± 80 220 ± 70++ 440 ± 60 560 ± 60 370 ± 80 410 ± 70 370 ± 70 380 ± 60 310 ± 70 470 ± 60 310 ± 60

1473–1635 1490–1649 1515–1696 1460–1630 1465–1631 1322–1449 1390–1455 1440–1650 1420–1635 1450–1630 1420–1450 1430–1470 1436–1625 1516–1669 1485–1633 1498–1636 1419–1627 1442–1613 1431–1626 1513–1652 1509–1642 1447–1626 1508–1602 1507–1654 1650–1680 1401–1435 1443–1623 1414–1449 1495–1636 1400–1640 1405–1445 1450–1650 1420–1480 1430–1500 1325–1430 1445–1645 1440–1635 1440–1635 1450–1635 1480–1655 1425–1490 1480–1650

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Assay no. Tx-666 Tx-199 Tx-202 Beta-152353 Beta-132852 Beta-132853 Beta-181792 Beta-181793 Beta-181795 Beta-244934 Beta-204251 Beta-204250 Beta-120069 Beta-120070 Beta-125985 Beta-125986 Beta-125987 Beta-132239 Beta-132240 Beta-132241 Beta-132242 Beta-132243 Beta-132244 Beta-132245 Beta-132246 Beta-133239 Beta-133240 Beta-138850 Beta-138851 Beta-138852 Beta-138853 Beta-138855 Beta-138856 Beta-138857 Beta-138858 Beta-138859 Beta-138862 Beta-138863 Beta-138867 Beta-132855 Beta-129980 Beta-129981

Context C M M C M M H H H C H H H H H H H H H H H H H H H C C M M H H H H H H H H H H H H H (continued)

Table 13-1. Titus phase radiocarbon dates (continued) Site name/site number Red Honeysuckle/41CP335 Alex Justiss/41TT13

41TT182 41TT373 41TT392 41TT406 Mockingbird/41TT550

Copyright © 2012. Nebraska Paperback. All rights reserved.

Ear Spool/41TT653

41TT672 William A. Ford/41TT852

Jas. E. Richey 41TT853

Conventional Calibrated age age range* (A.D.) 410 ± 80 400 ± 40 380 ± 40 550 ± 40 290 ± 120 320 ± 70 440 ± 80 320 ± 80 470 ± 60 350 ± 60 460 ± 60 390 ± 50 330 ± 60 450 ± 60 380 ± 50 360 ± 50 460 ± 50 320 ± 60 400 ± 40 440 ± 50 320 ± 50 320 ± 50 420 ± 50 360 ± 50 280 ± 40 380 ± 40 210 ± 40 400 ± 40 140 ± 40 150 ± 40 290 ± 40 190 ± 40 430 ± 50 400 ± 40 350 ± 40 410 ± 40 400 ± 40 410 ± 40 350 ± 40 260 ± 40 320 ± 40 250 ± 40 380 ± 40

1435–1635 1430–1630 1440–1640 1310–1430 1465–1680 1492–1649 1411–1625 1483–1666 1404–1486 1507–1654 1433–1622 1509–1642 1508–1667 1437–1623 1449–1626 1485–1633 1415–1484 1497–1600 1442–1620 1422–1615 1516–1647 1516–1647 1434–1619 1485–1633 1523–1667 1450–1624 1650–1950 1440–1490 1670–1950 1670–1950 1520–1650 1660–1950 1431–1615 1432–1632 1455–1637 1427–1631 1432–1632 1427–1631 1455–1637 1491–1681 1469–1648 1514–1683 1442–1634

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Assay no.

Context

Beta-133241 Beta-170437 Beta-170438 Beta-170439 Beta-44787 Beta-44789 Beta-48886 Beta-64977 Beta-64982 Beta-99688 Beta-99689 Beta-99690 Beta-99692 Beta-99693 Beta-105530 Beta-105531 Beta-117271 Beta-117273 Beta-117275 Beta-119001 Beta-119002 Beta-119003 Beta-119006 Beta-119623 Beta-119624 Beta-119625 Beta-229318 Beta-229319 Beta-229320 Beta-229321 Beta-229322 Beta-229323 Beta-80432 Beta-242378 Beta-242380 Beta-242381 Beta-242382 Beta-242384 Beta-242385 Beta-242386 Beta-242388 Beta-242389 Beta-242390

H C C C H H H H H C C C C C H H H H H H H H H H H H H H H H H H H H H H H H H H H H H

Table 13-1. Titus phase radiocarbon dates (continued) Site name/site number S. Stockade/41TT865 Harroun/41UR10

Dalton/41UR11 41UR118

Verado/41UR129*** Rookery Ridge/41UR133

Camp Joy/41UR144

Copyright © 2012. Nebraska Paperback. All rights reserved.

S. Lily Creek/41UR279 QLBS/41WD60 Steck/41WD529

Conventional Calibrated age age range* (A.D.) 380 ± 40 490 ± 100 265 ± 65 330 ± 110 555 ± 70 345 ± 75 480 ± 110 300 ± 60 440 ± 40 400 ± 60 460 ± 40 266 ± 42 360 ± 40 420 ± 50 540 ± 80 300 ± 50 550 ± 50 430 ± 60 260 ± 60 390 ± 60 310 ± 60 330 ± 70 460 ± 50 350 ± 40 480 ± 80

1442–1634 1386–1620 1511–1680 1444–1668 1387–1432 1479–1641 1391–1626 1518–1679 1430–1483 1442–1625 1400–1510 1250–1640 1480–1630 1435–1495 1315–1440 1515–1655 1395–1425 1427–1620 1518–1679 1495–1636 1515–1675 1502–1673 1410–1500 1470–1630 1393–1621

Assay no.

Context

Beta-242372 Tx-84 Tx-238 Tx-239 Tx-240 Tx-241 Tx-83 Beta-72372 Beta-90532 Beta-132010 Tx-7990 Tx-7989*** Beta-90534 Beta-117740 Beta-117741 Beta-117742 Beta-117744 Beta-132011 Beta-132012 Beta-84435 Beta-84436 Beta-145232 Beta-183858 Beta-210579 Tx-3473

H M M M M M M H H H H H H H H H H H H M M M H C H

* 2 sigma (see Stuiver et al. 1998); 41TT852, 41TT853, and 41TT865 calibrated following Reimer et al. (2004) ** H = habitation; C = cemetery; M = mound *** 2 sigma calibration by OxCal (Bronk Ramsey 2001) + radiocarbon date on human remains; C13/C12 value of -14.0 o/oo; ++=sample on corn

Joy Mound has three calibrated dates that span the interval from 1495 to 1673 (table 13-2), and this may have been one of the latest Titus phase mound constructions. It is interesting that the radiocarbon dates obtained in the 1960s from what have been characterized as Whelan phase mound sites (e.g., Thurmond 1990) fall into two clusters: one spanning the period from cal A.D. 1382 to 1520, and the other spanning the period from cal A.D. 1444 to 1668. The latter cluster of dates, with two dates from Harroun and one from Sam Roberts, is contemporaneous with those mentioned Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

377

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

UR10

CP304

Tx-84

Tx-238

Md. B fill

Beta-138851

Unit 7-01, 70-80 cm

House 4 under Md. D

Beta-138850

265 ± 65

490 ± 100

380 ± 60

470 ± 40

540 ± 60

Beta-132853

98.5-98.04 Elevation

Fea. 71

390 ± 60

Beta-132852

240 ± 90

320 ± 60

14 C age (B.P.)

98.85-98.55 Elev., burned Structure

Tx-202

submound structure, charred pole in md. fill

CP71

Tx-199

submound structure, charred pole in md. fill

CP8

Lab

Provenience

Site

265 ± 76

490 ± 108

340 ± 60 (-27.5 o/oo)

480 ± 40 (-24.4 o/oo)

520 ± 60 (-26.0 o/oo)

370 ± 60 (-26.3 o/oo)

240 ± 99

320 ± 72

Corrected age (B.P.)*

AD 1511–1599 (0.39) AD 1616–1680 (0.34) AD 1756–1804 (0.20)

AD 1385–1515 (0.72) AD 1311–1352 (0.17) AD 1593–1620 (0.10)

AD 1495–1605 (0.83) AD 1613–1636 (0.17)

AD 1414–1449 (1.00)

AD 1343–1449 (0.86) AD 1322–1340 (0.14)

AD 1559–1631 (0.54) AD 1465–1524 (0.46)

AD 1724–1816 (0.34) AD 1621–1696 (0.29) AD 1515–1592 (0.24)

AD 1490–1605 (0.76) AD 1613–1649 (0.24)

Calibrated age (1-sigma)**

Jelks and Tunnell 1959; Pearson et al. 1966

Jelks and Tunnell 1959; Tamers et al. 1964

Perttula 2005

Perttula 2005

Perttula et al. 2004

Perttula et al. 2004

Tunnell 1959; Pearson et al. 1965

Tunnell 1959; Pearson et al. 1965

Reference

Table 13-2. Titus phase radiocarbon dates from earthen mounds in the Pineywoods and Post Oak Savannah of northeast Texas

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

310 ± 60 330 ± 70

Beta-84436

Beta-145232 Beta-145233

Fea. 1 burned Lens (Zone D)

Zone A, 84-86 cm

Zone D, 71-74 cm

130 ± 60 (-28.0 o/oo)

280 ± 70 (-27.9 o/oo)

270 ± 60 (-27.4 o/oo)

340 ± 60 (-28.3 o/oo)

480 ± 117

345 ± 85

555 ± 81

330 ± 117

AD 1683–1745 (0.37)

AD 1502–1603 (0.53) AD 1614–1673 (0.35)

AD 1515–1592 (0.42) AD 1621–1675 (0.39)

AD 1495–1605 (0.83) AD 1613–1636 (0.17)

AD 1391–1520 (0.68) AD 1571–1626 (0.21) AD 1316–1346 (0.11)

AD 1479–1641 (1.00)

AD 1382–1437 (0.56) AD 1308–1357 (0.44)

AD 1444–1668 (0.98)

Perttula and Nelson 2001

Perttula and Nelson 2001

Perttula and Nelson 2001

Perttula and Nelson 2001

Davis and Gipson 1960; Tamers et al. 1964

Jelks and Tunnell 1959; Pearson et al. 1966

Jelks and Tunnell 1959; Pearson et al. 1966

Jelks and Tunnell 1959; Pearson et al. 1966

* Age not calibrated; delta 13C values in parenthesis. Assays on nutshell and wood charcoal lacking delta 13C values use the value estimates for fractionation correction, namely -25.0 o/oo. These particular assays have standard deviations that include an error in the estimated delta 13C. ** Calibrations use bidecadal record of Stuiver and Reimer (1993) and Stuiver et al. (1998); probability distributions are in parentheses.

180 ± 60

390 ± 60

Beta-84435

Fea. 1, burned lens at contact between md. fills (Zone D)

480 ± 110

345 ± 75

UR144

Tx-83

Tx-241

House 4, beam above floor, Md. D

555 ± 70

latest of 2 structures under Md [House B]

Tx-240

Md. C fill

330 ± 110

UR11

Tx-239

Md. C fill

Copyright © 2012. Nebraska Paperback. All rights reserved.

Copyright © 2012. Nebraska Paperback. All rights reserved.

above from Titus phase domestic and mound-building contexts. At Sam Roberts, although there is a Titus phase component in one area of the site and the calibrated date from the sub-mound structure dates to the same period, Thurmond (1990:144) has argued that the mound was built during the preceding Whelan phase because “there is no clearly demonstrated instance of mound building in a Titus phase context.” The sixteenth- and seventeenth-century radiocarbon dates from a burned structure in the Camp Joy mound and from several contexts at the Shelby site indicate, on the grounds of both ceramic and radiocarbon associations, however, that the Sam Roberts mound was built and used by Titus phase Caddo peoples. The few earlier dates (i.e., with beginning ages in the fourteenth century) may be from Whelan phase components, including two radiocarbon dates from a probable Whelan phase domestic context (cal A.D. 1315–1440 and cal A.D. 1395–1425) from the Rookery Ridge site (41UR133). The terrace area at the Rookery Ridge site has a buried, single-component occupation with Pease Brushed-Incised jars, Ripley Engraved vessel sherds with the continuous scroll motif, and a Perdiz arrow point (Nichols et al. 1997:table 17-2). These are characteristic of Period 1 (ca. 1350–1450) occupations in the Cypress Cluster (Perttula 1992:248 and table A-2), but could just as easily be included in the early part of the Titus phase on these same stylistic grounds alone, since all three ceramic and lithic types were made and used well past ca. 1450, including well-dated contexts at the Mockingbird, Ear Spool, and Pilgrim’s Pride sites. In addition to radiocarbon dates (n = 109), 16 Titus phase sites have ocr dates from columns and features in habitation areas or columns in mound deposits at the Pilgrim’s Pride, Tiddle Lake (41CP246), and Camp Joy mound sites (Perttula, ed. 2005:table 11-3). Components are dated as early as the first part of the fifteenth century, with ending dates on pedogenic markers that fall as late as 1642 at the Shelby site, 1655 at 41CP313, and 1660–1683 at the Camp Joy Mound. ocr dates from mound contexts at the Tiddle Lake, Pilgrim’s Pride, and Camp Joy mound sites compare favorably with those obtained from radiocarbon dating of Titus phase mound sites (see table 13-1): ca. 1477–1539 at Tiddle Lake, 1341–1478 from the Area VII mound at the Pilgrim’s Pride site, and 1520–1683 from Camp Joy mound. The ocr dates of 1411–1595 from the Frank Benson site (41TT310) are from a thick de380

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

posit of daub and burned clay that is apparently from a burned structure capped with a sandy mound fill (see Perttula and Nelson 2002). The general contemporaneity of the ocr and radiocarbon samples from these mound sites provides further confirmation of the fact that Titus phase Caddo peoples (at least in the Big Cypress Creek heartland) were actively engaged in the construction and use of mounds as places of power and ritual during most, if not all, of the Titus phase.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Mound-Building: Construction and Use of Mounds by the Titus Phase Social Elite In a general summary of mound building by Caddo peoples before ca. 1680, Barnes and Perttula (1999:6) noted that construction of ceremonial mound centers by the Caddo began about A.D. 800, and over the next 900 years “these mound centers became increasingly larger and more complex, under the apparent direction of a Caddoan chiefdom elite.” During the sixteenth and seventeenth centuries, population losses due to the introduction of European epidemic diseases may have led to the gradual cessation of ceremonial constructions in some parts of the Caddo area, and by ca. 1700, the vast majority of Caddo ceremonial mound centers were abandoned or were no longer being built and used by the Caddo peoples. Caddo mounds were apparently erected usually near or at the center of the larger villages or political communities, and on the larger rivers (such as the Red River) the mounds were integral parts of what Perttula (1992) called Caddo “towns.” As Early (2000:126) notes, the mounds and the community centers they were found on were the focus of mortuary and ritual activities for Caddo peoples. The mounds became permanent markers on the cultural landscape, because they were associated with specific rituals, events, and peoples, and they made “a powerful social and political statement. They serve equally well to reinforce the positions of important people and social groups, to mark territories, [and] to underscore common group identity” (Milner 2004:305). How the mounds were constructed and used in Late Caddo times by Caddo communities on smaller streams and tributaries to the larger rivers, such as Big Cypress Creek, particularly among the Titus phase Caddo, is not as well known because it was once thought that mound building in the Pineywoods and Post Oak Savanna of northeast Texas had ceased between 1400 and 1500/1550 (Thurmond 1990; Perttula Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

381

Copyright © 2012. Nebraska Paperback. All rights reserved.

1989, 1993, 1994, 1995). With continued archaeological research, and the expanded dating of mound features and archaeological deposits, there are currently 11 known Titus phase mounds in the Big Cypress Creek basin and the heartland of the Titus phase (Perttula 2004:fig. 1330). The Pine Tree Mound site (41HS15) on Potters Creek in the Sabine River basin (see fig. 13-2) appears to be another Titus phase ceremonial center and village with at least three mounds, a plaza, five associated habitation areas, shaft tombs, and at least one cemetery (Fields 2008). The radiocarbon dates from the Harroun, Dalton, Sam Roberts, Shelby, Pilgrim’s Pride, and Camp Joy mounds cover a broader span of time, however, than 1400–1500/1550, but fall into two clusters, one (at 1-sigma) ranging from cal. A.D. 1385 to 1520, and the other between cal A.D. 1444 and 1668. On this basis, it is suspected that most of the mound-building activities took place during most, if not all, of the Titus phase. In general, the calibrated radiocarbon ages from these Titus phase mounds indicate they are broadly contemporaneous with the construction of the other known mounds. At the Harroun site, for instance, House 4 (Mound D), the most probable calibrated age ranges of this burned house (Jelks and Tunnell 1959:fig. 10) are cal A.D. 1386–1515 and 1479–1641. The uppermost burned structure at the Shelby Mound has a calibrated age range of 1465–1631, approximately the same age range as the burned structure in the mound at the Pilgrim’s Pride site. The calibrated age ranges from the Dalton and Sam Roberts sites are comparable to the others. Only the one date (Tx-240) from Mound C at Harroun, on woody charcoal from mound fill (Thurmond 1990:table 59), suggests that the mound was built earlier than the Titus phase. Mound C is the only one of the four mounds at Harroun that has evidence of two sequent structures, suggesting a lengthier period of use than the other mounds and a period of construction and use that perhaps began in the late fourteenth to early fifteenth century. However, the other assay from Mound C has a calibrated range of 1444–1668, solidly in the Titus phase. The small number of Titus phase mound sites known in the region are unlike the types of mound complexes typically constructed in the major river valleys at this time (Story 1990). Mounds built by the Caddo that lived in the Pineywoods were substructural mounds; no pyramidal platform or burial mounds are known for this period. Substructural mounds are restricted to mounds that cap a burned circular structure (fig. 13-4a) 382

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

that was constructed on the ground surface or in a small, shallow pit. In at least two instances, the mounds contained sequent structures, but the “structures originated at higher levels in the mound[s] due to occupational accumulations of soil and ash, and not the result of any deliberate capping” (Thurmond 1990:168). At the Camp Joy mound, the 2.3 m high mound apparently had two tiers or platforms (Turner 1993), the latest tier capping a burned structure (marked by a 7 cm thick charcoal lens) dated to cal A.D. 1495–1605 (see Perttula and Nelson 2001). The structures that were capped by a mound deposit, or built at higher levels in the mound itself (as at Harroun), were circular, with extended entranceways facing west and with central hearths (fig. 134b). They were partially dismantled and burned, then capped with sediments. Again at the Harroun site, the structures were built inside large circular pits, and there were obvious soil berms around the enclosing pit and the structure (see fig. 13-4a). A standing structure with berms around it would look like the structure was buried (or partially buried) in the mound itself (cf. Schambach 1996). At the Dalton site, two temporally sequent circular structures (with clay-lined floors) of slightly different sizes were both built and used within the same shallow pit; when the second one was burned and destroyed, it and the surrounding pit were buried by a sandy mound fill to a depth of 80 cm (Davis and Gipson 1960:17–19; Thurmond 1990:210). At the Whelan site, one of the mounds buried four temporally sequent circular structures that ranged from 5.2 to 6.4 m in diameter (Thurmond 1990:168). Exactly what triggered the dismantling and burning of the structures, or their capping with mound sediments, at the Titus phase community centers is not clear. However, given the generally close association between the mound places and the community cemeteries (many of which held the burials of members of the social elite, including lineage heads), it seems plausible that the house destruction and mound-building episodes occurred after the death and burial of a leader or a member of the social elite. However, these elite individuals were buried with their peers and kin-affiliated relations in the community cemetery, not in the mound itself (as was the case in earlier Caddo times in east Texas, see Story 1997, 1998), and thus in essence the mound-building rituals of the Titus phase Caddo consisted of “public building–oriented ceremonialism” (Schambach 1996:41), such that the mounds “contain the remains of important buildings rather than important people.” Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

383

13-4a 13-4. Mound C at the Harroun site (41UR10): (above) a. profile, and (right) b. plan of the extended entrance structures in Mound C.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Thurmond (1990:234–235) suggests that the locations of Late Caddo period mounds in the Big Cypress Creek basin appear to be associated with clusters of contemporaneous settlements, cemeteries, and limited use areas, “and it is therefore possible that these concentrations of components represent the archeological manifestation of . . . Cypress cluster constituent groups during the [preceding] Whelan phase.” A similar association has been noted for Late Caddo period mounds and settlements in the middle Sabine River basin (Perttula 1989, 1994; Rogers and Perttula 2004). As I discussed above with respect to the identification of different political communities within the Titus phase, Thurmond’s 1990 suggestion appears to be a good one, with the exception that the mound building in the Big Cypress Creek basin primarily—if not exclusively—took place in Titus phase times rather than in the Whelan phase (ca. A.D. 1350–1450). 384

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

13-4b

With respect to the different political communities or constituent groups that may have existed among the Caddo after ca. A.D. 1430 in the Big Cypress basin, there are several key differences in the mound sites in the “heartland” of the Titus phase that hint at the sociopolitical diversity that was present in the various communities at the time. The first obvious difference is in the number of mounds on a site: the Harroun and Whelan sites each have four mounds, while the other sites have only one. We should note, however, that the proximity of the Chastain, Dalton, and Camp Joy mound sites to each other on a prominent upland landform (the three mound sites are within 250–500 m of each other) overlooking Meddlin Creek and Big Cypress Creek probably is evidence of a single large Titus phase mound and community center, not three disparate and unrelated sites. It is especially notable that the only multiple mound centers known in the Titus phase are all situated in the lower reaches of Big Cypress Creek, Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

385

Copyright © 2012. Nebraska Paperback. All rights reserved.

in the Lake O’ the Pines area (see fig. 13-2), while the other single mound sites are found some distance upstream in different political communities. These differences in the complexity of the various mound sites are likely to be a reflection of differences in the power and authority that each political community had within the Titus phase heartland, as the construction of earthen mounds express that power and authority in visible and tangible ways (cf. Payne 2002:195). It may well also be the case that the Lake O’ the Pines area was home to the highest densities of Titus phase Caddo peoples, as the size of each political community’s population would also be evidence of their chiefly power. Another difference between the mound centers is whether or not they were a discrete part of a larger planned village. In the case of the Pilgrim’s Pride and Shelby mound sites, the one earthen mound at each site was situated at the northern end of a large (+10 acres) village community (see fig. 13-3). At the P. S. Cash site, the one mound was 400 m north of a small associated cemetery, and no obvious habitation areas were noted by Jackson (1931), but there are extensive Titus phase habitation deposits immediately across Greasy Creek at the E. S. Dooley Farm (41CP4) (see Thurmond 1990:139 and fig. 17). The Sam Roberts site had a single 1.1 m tall mound that capped a burned circular structure (Tunnell 1959:4–7). There was a ca. 200 m2 midden deposit on the northeastern side of the mound, but there were much more extensive habitation areas 200 m east of the mound itself. According to Thurmond (1990:144), these habitation areas had “dark brown greasy soil . . . [and large dark outlines associated with concentrations of wattle-impressed daub [that] may mark the locations of structures.” The two habitation areas were each 70–100 m in diameter. In the immediate area of the Dalton, Chastain, and Camp Joy mound sites, there are extensive Titus phase habitation areas. These occur either on the mound sites themselves (as with the Chastain site, with a 15,000 m2 habitation area, see Thurmond 1990:212) or in nearby and associated domestic settlements. These include Sam Gray #1 (41UR9), Jake Martin (41UR12, Titus phase burials have been looted from the site in recent years), Watkins (41UR13), Sam Gray #2 (41UR14), Cunliffe (41UR15), and Mosquito Hollow (41UR16) sites (Thurmond 1990:fig. 27; Burson and Cliff 2000). The single mound at the Tiddle Lake mound site has no obvious as386

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

sociated habitation deposits, although there are numerous Titus phase settlements on nearby upland landforms (Perttula 2005:table 11-1). The only habitation area at the Whelan site was a ca. 2,000 m2 archaeological deposit southeast of Mound A and near Big Cypress Creek (Thurmond 1990:16). The other three mounds lay to the east and northwest. The small habitation area contained one 7.9 m household structure and an elevated 3.1 m granary structure, but other habitation features probably existed in the many areas not subjected to excavations (Davis 1958). No habitation deposits were identified at the Harroun site, but rather only four small mounds spread out along Big Cypress Creek (Jelks and Tunnell 1959; Thurmond 1990:fig. 27). With the main multi-mound centers at the Harroun and Whelan sites having little if no associated habitation debris, while most of the other mounds did, this dichotomy between mound and habitation associations on Titus phase mounds leads to the conclusion that there were fundamental and measurable differences in the complexity (if not wealth and prestige) of the different political communities in the heartland. For most of the communities, the ritual, power, and authority of the elite leaders had not been divorced from the populations living in the communities, especially those living in and near the largest villages. This intimate relationship—as seen by the placement of mounds across the landscape—between the community and its leaders was not duplicated in Titus phase Caddo communities living along the lower Big Cypress Creek basin and near the eastern margins of the settlement distribution of the Titus phase (see figs. 13-1b and 13-2). Here the community mound centers were basically kept separate from domestic affairs and were focused more exclusively on ritual activities and the control of ritual knowledge. Those leaders who lived at the Harroun and Whelan sites may have gained their prestige and authority through their control of ritual affairs (cf. Potter 2000:301). Although the Whelan site had only a small habitation area, a very large assemblage of pottery sherds (more than 14,000 in all) were recovered in the excavations, including a dense midden with thousands of sherds and numerous animal bone fragments inside the one non-mound structure (Davis 1958:42–48, 65–66; Thurmond 1990:168). This particular structure (structure 2) was just southeast of Mound A, and next to a probable granary perhaps under elite control. Thurmond suggests that the midden “derives from use of the structure for trash disposal after its Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

387

abandonment (but while it was yet standing), as the deposit was entirely within its perimeter.” These archaeological deposits have the markings of public feasting activities (see Hayden 2001:table 2.1), and ritual feasting events have been noted at Early and Middle Caddo period mounds in east Texas and southwest Arkansas (Perttula 2004:386; Scott and Jackson 1998, see also Jackson et al., chap. 3, this volume). A series of repetitive rituals centered on communal feasting took place here (with the end result being a large midden refuse deposit filled with broken ceramic vessels and animal bones), and probably also in the series of four structures eventually buried in nearby Mound A. Such feasting activities would certainly have served to help establish and preserve Titus phase intercommunity alliances and integration, at least at this particular political community. Furthermore, the Caddo community that was hosting the public feasting activities could well have “gained a measure of prestige that could be translated as a source of political influence for individual village headmen” (Knight 2001:327). Visualized in this matter, it would seem to be no coincidence that the scene of repeated public feasting activities by the Titus phase Caddo would be situated in the largest community mound centers. Mortuary Rituals and Burial Features

Copyright © 2012. Nebraska Paperback. All rights reserved.

There are 133 known Titus phase cemeteries in northeast Texas. These range in size from fewer than five individuals to as many as 200 burials, as far as can be determined from the available evidence, and they occur in a variety of settings within the Titus phase area, with a considerable diversity in burial treatment and mortuary offerings (see Perttula and Nelson 1998:328–401; Rogers et al. 2003:19–22). The majority of the known Titus phase burials are from large community cemeteries; these all had more than 70 interments (table 13-3). Cemeteries represent distinctive and socially significant places on the Caddo landscape. They are places where past Caddo societies created and renewed a sense of place and developed a sense of social memory (e.g., Cannon 2002:193–194; Goldstein 2006:380), leading to an appreciation of the social and spatial dimensions of mortuary practices. Ashmore and Geller (2005:87–88) have noted that cemeteries are appropriate places for transforming the dead kin people through life and death cycles (cf. Silverman 2002:7) and commemorating ancestral continuities through ritual action. Burials and cemeteries, whether small family cemeteries 388

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 13-3. Titus phase cemeteries

Copyright © 2012. Nebraska Paperback. All rights reserved.

# 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

Site name/site number Mockingbird, (41TT550) Horton (41CP20) A. P. Williams (41TT4) Thomas P. Caldwell (41TT6) Alex Justiss (41TT13) Joe Justiss (41MX2) Starrett (41MX4) Russell Brothers Farm (41TT7) Harold Williams (41CP10) Mattie Gandy (41FK4) Culpepper (41HP1) P. S. Cash Farm (41CP2) G. W. Rumsey (41CP3) E. S. Dooley (41CP4) Tuck Carpenter (41CP5) Johns (41CP12) R. A. Watts #2 (41CP14) Lower Peach Orchard (41CP17) 41CP22 Shelby Mound (41CP71) 41CP76 Guest (41CP78) Jonah C. Atkinson (41FK1) J. E. Galt (41FK2) B. J. Connally (41FK5) S. P. Brown (41FK6) P. G. Hightower (41FK7) W. A. Ford (41TT2) W. L. Ormes (41HS1) H. R. Taylor (41HS3) Keasler (41HS235) Attaway (41HP15) 41HP26 Ben McKinney (41MR12) W. M. Dickson (41MR13) Pleasure Point (41MR63) Mims Chapel (41MR74) R. L. Cason (41MX1) Hancock Farm (41MR24) Lone Star Steel (41MX9) Camp Joy Mound (41UR144) Pea Patch (41HS825)

Interments 11+ 19 10 10 26+ 11 1+ 25-30+ 95+ 11 22 9 5-6+ 2+ 71+ 30 1+ 35-45 8+ 119+ 3 7 5 5 2 1 3+ 3+ 1+ 71 35+ 1 1 11+ ? 120-140 20+ 4 4 1+ 6+ 88+

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Drainage basin Hayes Creek Big Cypress Creek Dragoo Creek Tankersley Creek Swauano Creek Boggy Creek Black Cypress Creek Big Cypress Creek Dry Creek Coon Creek Stouts Creek Greasy Creek Greasy Creek Greasy Creek Dry Creek Prairie Creek Big Cypress Creek Big Cypress Creek Brushy Creek Greasy Creek Dry Creek Big Cypress Creek White Oak Creek Brushy Creek Big Cypress Creek Brushy Creek Big Cypress Creek Sulphur River Arms Creek Arms Creek Little Cypress Creek Caney Creek Caney Creek Big Cypress Creek Arms Creek Big Cypress Creek Big Cypress Creek Boggy Creek French’s Creek Big Cypress Creek Big Cypress Creek Arms Creek (continued)

Table 13-3. Titus phase cemeteries (continued)

Copyright © 2012. Nebraska Paperback. All rights reserved.

# 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85

Site name/site number George L. Keith (41TT11) Wofford (41TT62) 41TT111 Fourth Tower B (41TT125) Covey (41TT320) W. O. Reed (41UR1) J. M. Riley (41UR2) Henderson-Southall (41UR3) S. Blackstone Farm (41UR31) Gold Star Ballroom (41UR107) J. H. Reese (41WD2) Minnie Garrison (41WD16) 41WD19 1972S (41WD44) YBS (41WD45) M. W. Burks (41WD52) Sandhill (41WD108) Spoonbill (41WD109) 1973S (41WD206) Goldsmith (41WD208) Jessie Rogers (41WD533) Gilbreath (41WD538) Big Oaks (41MR4) Lasco (41UR106) 41UR109 West Island (41MX65) Holly Lake Ranch (41WD57) A. C. Riley (41CP7) R. A. Watts #1 (41CP13) J. H. Chadd (41HS2) Scotty Scott (41TT119) 41UR32 Spider Lilly (41UR143) Sandy Creek (41MR122) Cedar Springs (41UR137) 41MR110 Lost Indian (41TT643) QLBS (41WD60) 41TT716 Stephens Barn (41TT717) D.C. Washout (41HS577) CK 1 (Upshur Co.) Sandlin Dam (41TT726)

Interments 7 1+ 6-11+ 1+ 6 5 18+ 200+ 5 53 5 1+ 15 2 3 1+ 1 19 2+ 3 1+ 5 80-120 30+ 2+ 2+ 1+ 2+ 1 1+ 2+ ? 60+ 67 ? 45 33 20+ 21+ 27+ 6+ 42 150+

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Drainage basin Hart Creek Swauano Creek Big Cypress Creek Blundell Creek White Oak Creek Greasy Creek Little Creek Meddlin Creek Glade Creek Big Cypress Creek Little Dry Creek Brushy Creek Dry Creek Dry Creek Caney Creek Little Dry Creek Caney Creek Caney Creek Dry Creek Dry Creek Caney Creek Glade Creek Big Cypress Creek Kelsey Creek Kelsey Creek Ellison Creek Holly Creek Big Cypress Creek Big Cypress Creek Arms Creek Big Cypress Creek Big Cypress Creek Walnut Creek Sandy Creek Big Cypress Creek Big Cypress Creek East Piney Creek Dry Creek Big Cypress Creek Hart Creek Little Cypress Creek Lilly Creek Big Cypress Creek

Copyright © 2012. Nebraska Paperback. All rights reserved.

# 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127

Site name/site number J. McKay (41TT730) Power Plant (41TT727) J. Wright (41FK96) 41TT96 French-Daily (41CP238) Jack Morris (41TT723) Sugar Creek (41UR207) State Park (41TT725) Oil-topped (41CP225) Meddlin Creek (Upshur Co.) Arms Creek (Marion Co.) Mutt McGrede (41GG53) Teneryville #1 (41GG50) Teneryville #2 (41GG51) Mud Creek (41GG56) Daingerfield Water Intake (41MX28) Peterson Ranch (41HS253) Maughan (Upshur Co.) Keeling (Upshur Co.) 41HS5 Couch Mountain (Camp Co.) Turbeville (41WD20) Tarp (41TT729) Dickson (41GG55) Whatley (41GG54) Woody Michener (41CP244) Polk Estates (41CP245) Sword (41UR208) Hoosier School (41WD524) Tina Resch (41HS14) Tom Byrd (Morris Co.) Jake Martin (41UR12) 41CP317 Pilgrim’s Pride (41CP304) 41CP350 Rookery Ridge (41UR133) Ear Spool (41TT653) Pine Grove (41HS826) W-S (41TT741) Harold Nix (Morris Co.) Dixon Creek (41UR273) Redware (41UR272)

Interments 17+ 27+ 31+ 1+ 30+ 40+ 40+ 30+ 40+ 115+ 80-120+ 25+ 12+ 23+ 20+ 20 15-20+ 3 15 5+ 115+ 2 30+ 8 14 11+ 20-30+ 47+ 1+ 20-30+ 1+ 3-4+ 1 30 2+ 4 3+ 10+ 118 20 20+ 30+

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Drainage basin Hart Creek Swauano Creek Brushy Creek Tankersley Creek Walkers Creek White Oak Creek Sugar Creek Big Cypress Creek Big Cypress Creek Meddlin Creek Arms Creek Hawkins Creek Hawkins Creek Hawkins Creek Sabine River Big Cypress Creek Grays Creek Little Cypress Creek S. Greasy Creek Arms Creek Greasy Creek Lake Fork Creek Swauano Creek Hawkins Creek Hawkins Creek Big Cypress Creek Big Cypress Creek Gum Creek Caney Creek Potters Creek Boggy Creek Big Cypress Creek Walkers Creek Big Cypress Creek Big Cypress Creek Kelsey Creek Piney Creek Little Cypress Creek Swauano Creek Swauano Creek Greasy Creek Greasy Creek (continued)

Table 13-3. Titus phase cemeteries (continued) # 128 129 130 131 132 133

Site name/site number

Interments

Heart Attack Hill (41HS828) 41WD511 Gus Bogan (41WD25) Turquoise (41WD586) Peach Orchard Overlook (41CP25) Tuinier Farm (41HP237)

1 1 ? 9 11+ 3

Drainage basin Little Cypress Creek Lake Fork Creek Sabine River Caney Creek Big Cypress Creek Stouts Creek

Note: See Perttula (2005:Table 11-4) for complete references for each of the cemeteries. The sites without trinomial numbers have not been recorded due to lack of exact locational information.

or the larger community cemeteries, reflect a continuity of kin and community, ethnicity and identity, in their location relative to the inhabited and habitable landscape and in their location relative to other deceased Caddo peoples. As Chesson (2001:110) has noted: On a regional scale, the people who buried their dead . . . followed a particular set of standards in their mortuary rites, forging a sense of similarity between people in [the] region, as well as nurturing a sense of tradition in the treatment of [the] dead. These communities shared a set of principles and local knowledge that instructed them how to treat their dead, where to commemorate them, and the importance of maintaining these practices. This knowledge, and the resulting practices, may reflect a regional identity. [emphasis added]

Copyright © 2012. Nebraska Paperback. All rights reserved.

From this perspective, mortuary practices of the Titus phase Caddo peoples are believed to be guided by the cultural and social beliefs, ideas, and traditions of the living Caddo people concerning the afterlife as well as the relationship of the living with their ancestors (cf. Sullivan and Mainfort 2010:7–12). These ideas and relationships—essentially social ties expressed in shared mortuary rituals—can endure for generations. Shared ritual mortuary practices among the Caddo living in the Big Cypress Creek basin in the Pineywoods lasted as least 250–300 years (10–15 generations). Because these mortuary practices are based on the active beliefs of the living Caddo who laid the deceased in graves in certain sacred areas, with specific burial pit forms and orientations, as well as certain kinds and quantities of grave furniture (cf. Carr 1995:table III), mortuary practices provide key insights into Caddo cosmology and ideology, particularly beliefs on the afterlife, the nature of the soul, and steps necessary to protect and save the souls of the deceased. 392

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

The setting of cemeteries across the Big Cypress Creek basin region is not haphazard, but has a spatial and historical context that can be related to traditions and beliefs shared by Caddo peoples about universal orders and the structure of the cosmos (e.g., Carr 1995:table XIV; Charles 2005:23–24; Kay and Sabo 2006; see Sabo, chap. 15, this volume). The demographic dispersion of agricultural Caddo peoples, and the apparent belief that the dead needed to be kept near the living, led to the accompanying dispersion of cemetery sites in proximity to habitation areas and settlements. These Caddo households, or groups of households comprising a community of kin-related groups, defined their own sense of place and who they were by the unique juxtaposition of the living and the dead in circumscribed areas within larger farming communities. An emphasis on affinity and kinship within and between different households that created these sacred cemetery places was apparently more important to household members than were their broader ties to the larger communities of which they were a part. Although these Caddo shared similar views on social memory and the place of the dead in their lives, it was their households and associated cemeteries (i.e., their ritual space) that were the center of their universe, in juxtaposition to the larger and socially more diverse communities that they shared cultural traditions with. Titus phase cemeteries are particularly common in sites located along Big Cypress Creek itself, followed by other cemeteries on a series of tributaries to Big Cypress Creek (fig. 13-5). The principal tributary creek settings for Titus phase cemeteries are Swauano Creek, Boggy Creek, Dry Creek, Arms Creek, Meddlin Creek, and Greasy Creek (see fig. 13-2). With the exception of the Boggy Creek cemeteries, the Titus phase communities living on the other tributaries appear to represent recognizable concentrations of settlements, mounds, and community cemeteries that constitute distinct political communities. The average Titus phase cemetery contained about 26 individuals, arranged in a number of rows, with little if any overlapping of graves (Perttula and Nelson 1998:375). Titus phase burials were interred in an extended supine position with the head of the deceased facing west (towards the House of Death in the sky), and usually they were accompanied by a variety of grave goods, including ceramic vessels to hold food and liquids on the six-day journey to the House of Death, stone tools, as well as perishables such as baskets and wooden implements. In early Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

393

Copyright © 2012. Nebraska Paperback. All rights reserved.

13-5. Number of recorded burials from Titus phase cemetery sites on different drainages in northeast Texas.

eighteenth-century observations of Hasinai Caddo mortuary practices by Fray Espinosa (Hatcher 1927:162), he noted that “as soon as the souls leave the body they travel towards the west and from there they rise once more into the air and go close to the presence of the great captain whom they call caddi ayo. From there they go to wait in a house located towards the south, called the House of Death.” For the Caddo, the soul’s travel to the west is also associated with the setting of the sun. An east-west body orientation was likely primarily due to a near universal belief in the direction of the afterlife and the soul’s journey to the afterlife. There is a considerable variation in the size of the cemeteries, which 394

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

13-6. Variation in the size of Titus phase cemeteries.

must relate not only to the length of time they were used by different families and lineages as places for burial of the dead but also apparently depending upon if the cemeteries were used by a larger community and not just extended family members. More than 40 percent of the Titus phase Caddo cemeteries with reliable information on the numbers of burials they contained had less than 5 interments, and another 24 percent had between 6–20 separate burials (fig. 13-6). Approximately 17 percent had 21–40 burials, and 9 percent had 41–80 graves. Finally, almost another 9 percent were very large cemeteries with either 81–160 or 160+ separate Caddo burial features. The two largest known Titus phase cemeteries are the Shelby site on Greasy Creek (Thurmond 1990; Mitchell 2000; Perttula et al. 2004) and the Henderson-Southall site (41UR3) on Meddlin Creek (Perttula et al. 1998); both cemeteries are situated a few miles above the confluence of tributary streams with Big Cypress Creek (see fig. 13-2). Titus phase Caddo cemeteries fall readily into two groups: community and family (see Perttula and Nelson 1998). The community cemeteries are large and well planned, with more than 70–80 individuals, and often must have been the final resting place of individuals of higher social status or rank within various communities; in fact, their burials may have been the impetus for the creation of many community cemeteries. These very large cemeteries appear to be the result of burials Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

395

Copyright © 2012. Nebraska Paperback. All rights reserved.

from a number of farmsteads and small village compounds within each of the political communities mentioned above, and their existence—and persistence in use for a number of generations—show that there was a community-wide participation in ceremonial and mortuary rituals. There are 13 known Titus phase community cemeteries, 12 on Big Cypress Creek and its principal tributaries (i.e., Arms, Meddlin, Greasy, Boggy, and Dry creeks), and the thirteenth on Walnut Creek, a tributary to Little Cypress Creek (see fig. 13-2). From north to south, the community cemeteries include the Sandlin Dam, Tuck Carpenter, Harold Williams (41CP10), W-S (41TT741, no. 124), Shelby/Tracy, Gold Star Ballroom (41UR107), Pleasure Point (41MR63), Henderson-Southall, Big Oaks (41MR4), Sandy Creek (41MR122), Pea Patch (41HS825), H. R. Taylor (41HS3), and Spider Lilly (41UR143) on Walnut Creek. The community cemeteries are internally organized by space. They may also be differentiated by the placement of high status or socially ranked individuals near the center of various clusters of burials or in the central part of the cemetery itself (see Perttula 2004b:fig. 13.33), as if that individual was a founding member of one of the important Caddo families in the larger community that created and maintained the large cemeteries. At the W-S community cemetery site on Swauano Creek (fig. 13-7a), there were two periods when the cemetery was apparently being used (Perttula 2006b). This is evident by the different orientation of certain burials and burial rows; the burials from each period did not overlap. Ripley Engraved vessels are common throughout the cemetery, including a few with the post-1600 pendant triangle motif. Other recognizable vessel forms in the northern area include Simms Engraved, Taylor Engraved, and Cass Appliqued. One distinctive vessel (burial 68) was a double-tiered engraved vessel with double rows of suspension holes. This vessel may be a form of Natchitoches Engraved; similar vessels have been documented in a collection from the J. M. Riley site (41UR2) and from a Titus phase site in the Little Cypress Creek basin (Perttula 2006a). In the southern part of the cemetery (south of the dashed line on fig. 13-7a), probably the earliest episode of cemetery use, the burials were oriented northeast-southwest. The remainder of the younger burials in the cemetery are in rough and closely packed east-west aligned rows. Although the available documentation on the burials is somewhat scanty (Perttula 2006b), information on the range of funerary offerings 396

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

from the two periods of cemetery use does shed some light on differences in mortuary ritual at a large Titus phase community cemetery. First, the earlier burials in the southern part of the W-S cemetery tended to have more ceramic vessels and caches of Talco arrow points than did the later and larger northern part of the cemetery. Three burials (1, 6, and 90) had more than 14 vessels placed as offerings with the deceased; none of the burials on the northern part of the cemetery had as many. Three other graves (8, 26, and 71) each contained more than 30 Talco arrow points, compared with only one grave (47) in the northern part of the cemetery. However, Talco points themselves are found in more graves in the youngest part of the cemetery, whether they occurred in large caches or not. Second, the one burial (16) with marine shell beads and ceramic ear spools is also found in the southern and earlier cemetery area. Third, five burials (37, 44, 47, 63, and 99) at the W-S cemetery had clay pipes as funerary offerings, and 80 percent of them are in the northern cemetery area. Finally, four burials (55, 62, 74, and 90) contain large chipped blades, probably Galt bifaces, and three of these are in the northern and younger cemetery area. These large bifacially chipped blades may represent warrior symbols, used as badges or items of display (cf. Hally 2008:465). At the Shelby site, the available documentation (see Mitchell 2000; Perttula et al. 2004) indicates that the many burial rows of interments had a roughly east-west orientation, similar to the youngest burials in the northern part of the W-S site community cemetery. Four burials (116–119) are more northeast-southwest in orientation. A total of 119 burial pits had been mapped by one of the diggers when the Shelby site was being looted, and apparently more burials had been excavated in the cemetery besides those mapped in figure 13-7b (Bo Nelson, 2002 personal communication), but the spatial patterning in this large community cemetery is clear. Mitchell (2000) indicates that not all the burials at the Shelby site were single, extended interments. Some apparently held multiple individuals, and several of the Caddo burials had been placed in deep and probable shaft tombs; signs of the looted shaft tombs were still visible in 2004 (Perttula et al. 2004). Notes on file at the Texas Archeological Research Laboratory at the University of Texas at Austin also indicate that there was a large concentration of daub in the area of the later plotting of burial 117 (see fig. 13-7b), and it is possible that this particular burial had Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

397

Copyright © 2012. Nebraska Paperback. All rights reserved.

13-7a 13-7. The structure of Titus phase community cemeteries: (above) a. the community cemetery at the W-S site on Swauano Creek in Titus County, Texas; (right) b. the Titus phase community cemetery at the Shelby site (41CP71).

been placed in the floor of a deliberately burned and wattle- and daubcovered Caddo structure. This burial is also significant because in it were two 9 ft red cedar poles, a rare find indeed. The cedar poles may be the preserved remnants of a wood litter upon which burial 117 would have rested. One cannot think of a wood litter on a pre-1680 Caddo site without considering the cedar pole burial litters in the Great Mortuary at the 398

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

13-7b

Spiro site (see Brown 1996), as these were among the most laden in mortuary contexts there with exotic marine shell artifacts, embossed copper plates, and other rare Southeastern Ceremonial Complex artifacts. While not as spectacular in that sense as the litter burials from Spiro, this litter burial at the Shelby site was said to contain five whole ceramic vessels, 22 arrow points, and many sherds from an uncertain number of other vessels (Mitchell 2000). One of the vessels was described as a red, yellow, and tan bowl, which may be a Hatinu Engraved vessel, similar to Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

399

Copyright © 2012. Nebraska Paperback. All rights reserved.

those seen in collections from the Hatchel (41BW3), Clements (41CS25) (Gonzalez et al. 2005:30 and fig. 4.5), and Foster (3LA27) (see Weinstein et al. 2003:figs. 106 and 107) sites in the Red River basin. Other notable burials at the Shelby site had ceramic ear spools, rarely found in contexts other than those associated with the social elite, and at least three large (20 cm) black Big Fork chert bifacially knapped blades came from burials (Mitchell 2000). Finally, one burial (not identified by burial no.) was accompanied by a very large cache of 120 arrow points, 117 of which were of the finely chipped Talco type, along with celts, hammerstones, axes, and a probable heir-loomed bannerstone (Mitchell 2000). These kinds of funerary objects placed by the Caddo with the deceased at the Shelby site are very much indicative of high-status Titus phase burials (see Perttula 2004:401; Thurmond 1990:235). The family cemeteries on Titus phase sites are generally found in the immediate proximity to a farmstead or hamlet, and they contain fewer interments than the much larger community cemeteries. They seem to have had 10–20 individuals in cemeteries along the western margins of the Titus phase area and between 20–40 individuals in the Titus phase heartland along Big Cypress Creek, suggesting that there were intraareal differences in population densities and the social organization of extended families and lineages. Burials within the family cemeteries generally included single extended inhumations within a patterned arrangement of burials in rows, sometimes aligned east-west and other times in roughly north-south rows. Burial pit sizes, funerary offerings, and preserved human remains suggest the majority of the burials in the family cemeteries are adults. Grave good associations and burial treatment of Caddo peoples in Titus phase family cemeteries do not show much evidence of differential status or social rank. According to Thurmond (1990:235–236), artifact associations in family cemeteries differed primarily only by age and sex: adolescents tended to be buried with more funerary offerings than children or infants, but would have had fewer funerary offerings than adults, whether they were male or female. The graves of males often had clusters of arrow points in patterns suggesting they represented quivers of arrows, and females commonly had polishing stones or more numerous pottery vessels. Items of non-local origin (either vessels or stone tools) were extremely rare. Finally, graves with large numbers of funerary offerings were limited in Titus phase family cemeteries. This limited diversity in burial treatment is part of the evidence preserved in 400

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

the mortuary practices of these Caddo peoples that strongly hints that there were not any considerable social organizational differences between affiliated and kin-related farmsteads that created these cemeteries (cf. Carr 1995:157 and tables XIII and XIV). Carr (1995:tables XI and XIV) further suggests that body treatment, preparation, and orientation closely track beliefs about the soul and the afterlife more than they do the social position of the deceased. An example of a large family cemetery at the Pilgrim’s Pride village had 30 interments (Perttula 2005:221–294). Its planned location within the village community, and its internal character, suggest that the Pilgrim’s Pride site cemetery was probably used for several generations by the different families and lineages that lived throughout the large village. Although all the Caddo burials in the cemetery are single individuals placed in an extended supine position, with their heads facing west, six burial groupings were recognized. These were defined on the basis of burial treatment (i.e., distinct burial fills), grave pit size, and the kind and amount of funerary offerings placed with the deceased (Perttula 2005:table 11-5). The six burial groupings sorted readily by character as well as spatially (see Perttula 2005:fig. 6-26) into two larger groups, based primarily on the use of red clay for burial fill only in Groups I–III. More important adult family members or lineage heads are included in Groups I and IV, and they had the highest numbers of burial offerings, including celts, arrow points, and other chipped and ground stone tools. Based on the relative position of the different burial groups in the several rows of burials at the Pilgrim’s Pride site, the Group IV–VI burials were later (perhaps more than a generation later) than the Group I–III burials. In other ways, however, the mortuary rituals used by the different groups were very much the same through time. That is, the deceased—whether adult or juvenile—were placed on their back in a burial pit that was excavated into the red clay B-horizon, with their heads facing west, toward the sunset (Perttula and Nelson 1998:378; Turner 1978:105). The burials were accompanied primarily by ceramic vessels, placed by the head and along one or both sides of the body (bowls, carinated bowls, jars, and bottles), that probably held foods and liquids. A few individuals— probably adult males and senior family members—had caches of stonetipped arrow points also placed with them in the grave, along with an occasional ground stone celt or a clay elbow pipe. Other funerary offerings included clay pigments and a few chipped stone tools (such as a beveled Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

401

Table 13-4. Frequencies of funerary objects in selected Titus phase cemeteries Site name

Ceramic vessels (mean)

R. L. Cason W. A. Ford Thomas P. Caldwell Alex Justiss J. M. Riley J. E. Galt Tuck Carpenter H. R. Taylor Ben McKinney Goldsmith B. J. Horton A. P. Williams W-S Spoonbill P. S. Cash Mattie Gandy Mockingbird Joe Justiss W. O. Reed Pilgrim’s Pride Harold Nix Turquoise Culpepper

9.3 9.3 9.5 7.2 9.9 9.8 9.2 8.3 11.0 8.3 7.5 8.8 7.1 8.1 5.7 8.2 8.1 6.2 7.0 6.6 6.0 4.2 4.9

Arrow points (mean)

Total specimens

12.8 10.3 8.7 9.4 5.1 0.4 4.3 5.1 0.8 3.3 3.5 2.9 4.1 2.6 4.7 1.4 1.0 2.0 1.4 0.9 1.3 1.7 0.4

24.5 20.0 18.9 16.8 16.1 15.2 14.8 14.5 13.8 13.7 13.0 12.6 11.6 11.4 10.9 10.3 10.1 8.5 8.4 8.3 7.5 6.3 5.4

No. of burials 4 3 10 28 18 5 45 71 8 3 11 10 118 7 9 11 11 11 5 19 20 9 8

Note: See Perttula (2005:table 11-6) for references for each of these cemeteries.

Copyright © 2012. Nebraska Paperback. All rights reserved.

knife made from Florence A chert that originated in south Kansas) and caches of lithic debris from non-local raw materials. The 19 burials in the Pilgrim’s Pride cemetery had a mean of 8.3 funerary objects per burial. Approximately 80 percent of these offerings were ceramic vessels. The frequency of funerary objects at this cemetery is on the lower end of the scale of burial furnishings when compared with the other documented Titus phase cemeteries listed in table 13-4. Those cemeteries at the top of the list, including R. L. Cason (41MX1), W. A. Ford (41TT2), Thomas P. Caldwell (41TT6), and Alex Justiss, had between 16.8 and 24.5 funerary objects per burial, with numerous distinctive caches of arrow points accounting for 50–60 percent of the items placed with the deceased. In the rest of the 23 cemeteries listed in table 13-4, 60–90 percent of 402

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

the funerary offerings are ceramic vessels, and the absolute quantity of funerary offerings is as much as four times lower than has been noted at the four cemeteries dominated by offerings of arrow point caches/quivers. Pilgrim’s Pride is at the low end of the scale of total specimens per burial, and it also has one of the lowest number of arrow point cache/ quiver offerings, only slightly higher than at the Culpepper (41HP1) and J. E. Galt (41FK2) cemeteries, and about the same as at the Mockingbird (41TT550) cemetery. There are a number of burials found in Titus phase cemeteries that are clearly those of important people, either adult members of the social elite, individual community leaders, or even heads of paramount lineages (as suggested by Rogers et al. 2003:21–22). They received special treatment at death, including having rare artifacts placed with them in the burial pits, just as they must have owned during their lives (table 13-5). It does appear to be the case with Caddo burials in Titus phase sites that only a few individuals in these communities had access (either in life or death) to the kinds, numbers, and variety of funerary offerings that have been documented in Caddo cemeteries and individual burial features in mound contexts, in shaft tombs, or in other special mortuary contexts from east Texas where individuals had distinctive exotic and non-local funerary offerings. These burial features appear to be those of individuals of high social status, probably men, that are found in contextually unique settings, with high numbers of ceramic vessels and/or arrow point quivers and pipes, as well as a diverse array of other sorts of funerary offerings. These individuals may represent those with a higher social position, likely the heads of different lineages or extended kin groups that resided in farmsteads comprising the community, and ended up being interred in associated cemeteries amidst their kin. Notable Caddo burials in Titus phase contexts include shaft tombs at three sites, two on Big Cypress Creek and the third on Greasy Creek; burials placed in two mounds on Big Cypress Creek; at least seven sites on Big Cypress Creek and its tributaries with large chipped Galt-style bifaces; 17 burials at nine sites on tributaries of Big Cypress Creek with individuals buried with large numbers of grave goods; and four burials from three sites on tributaries to Big Cypress Creek and White Oak Creek that had double extended supine burials with quantities of grave goods.2 Taking into account the fact that cemeteries contain the burials of both adult men and women, as well as some children, it is likely that Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

403

Table 13-5. Notable burials from Titus phase cemeteries and kind of mortuary treatment Site*

Burial No. Burial Treatment

Total Total Total Spec. Arrow Vessels

Burial in shaft tomb Lower Peach Orchard

N/A

Pleasure Point

N/A

Shelby/Tracy

N/A

5-6 shaft tombs; Multiple interments shaft tomb; multiple Interments shaft tomb; multiple Interments; one with cedar pole litter; Galt bifaces

N/A

N/A

N/A

N/A

N/A

N/A

N/A

50-60

100+

burials on mound platform 11 burials in mound

N/A N/A

N/A N/A

N/A N/A

28 9 15 N/A N/A N/A N/A

N/A 3 0 N/A N/A N/A N/A

14 4 12 N/A N/A N/A 15

37 35 35 47

25 25 21 30

8 9 13 12

41 28 29 39 46 37 38 47 31 190+ 30+ 34+ 40

29 20 20 28 29 31 22 22 24 170 27 34 20

8 8 8 11 12 6 14 15 7 17 N/A N/A 7

Burial in mounds Camp Joy N/A Peach Orchard Overlook N/A

Burials with large chipped Galt bifaces Galt Pleasure Point H. R. Taylor Spider Lilly French-Daily Mutt McGrede Sword

3 N/A 54 N/A N/A N/A N/A

extended supine burial position unknown extended supine extended supine extended supine extended supine extended supine; two Large bifaces

Extended supine burials with large quantities of grave goods

Copyright © 2012. Nebraska Paperback. All rights reserved.

Caldwell Caldwell Tuck Carpenter

H. R. Taylor

Joe Justiss P. S. Cash J. M. Riley

Harold Williams Spider Lilly B. J. Horton

1 4 1 19

2 11 12 59 4 3 4 11 15 N/A N/A N/A 3

extended supine extended supine extended supine extended supine; female*; also with Galt biface extended supine; male** extended supine; male extended supine; male extended supine; male extended supine extended supine extended supine extended supine extended supine extended supine extended supine extended supine extended supine

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Table 13-5. Notable burials from Titus phase cemeteries and kind of mortuary treatment (continued) Site*

Burial No. Burial Treatment

Total Total Total Spec. Arrow Vessels

Double extended supine burials with quantities of grave goods Tuck Carpenter

21 23

H. R. Taylor

45

Lost Indian

N/A

double extended supine; Male and indeterminate double extended supine; indeterminate and adolescent double extended supine; Probable male and Female double extended supine

30

11

17

36

22

11

72

23

26

27+

N/A

26

Copyright © 2012. Nebraska Paperback. All rights reserved.

Note: See Perttula (2005:table 11-7) for references concerning each of these burials. * Sex assignments based on an examination of human skeletal remains (Turner 1978, 1992). ** Sex assignments based only on the range and kind of grave goods in burials (Thurmond 1990).

this diversity in the absolute quantity and kinds of funerary offerings in Titus phase Caddo sites represent the existence of social differences among the Caddo communities that created and maintained these cemeteries, particularly differences between men and women. These burials are from Caddo peoples in different Titus phase political communities that must have had considerable power and authority, at least in their own community, and perhaps as also recognized in other Caddo communities outside the Big Cypress Creek basin. Such burials are rare, certainly comprising less than 2 percent of the Titus phase population at any one time, and probably less than that given the spotty record of complete burial documentation on many of the Titus phase cemeteries listed in table 13-3. Some 20 Titus phase sites have notable burials (fig. 13-8). Quite a few are large community cemeteries, including the Pleasure Point, Shelby/ Tracy, H. R. Taylor, Spider Lilly, Tuck Carpenter, and the Harold Williams sites (see fig. 13-2). Others are in the same general area as the various political community centers already discussed along the larger eastward and southern-flowing tributaries to Big Cypress Creek. These important burials are found concentrated along Big Cypress Creek and its tributaries in the Pineywoods, from near the headwaters of Brushy Creek to the lower reaches of Big Cypress Creek in the Lake O’ the Pines dam area (see fig. 13-8). Two of the sites with notable Titus phase Caddo burials are on tributaries to Little Cypress Creek, another is Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

405

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

13-8. Distribution of Titus phase cemeteries with burials of individuals of presumed high social rank. Note that numbers on map correspond to sites listed in table 13-3.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Copyright © 2012. Nebraska Paperback. All rights reserved.

on Hawkins Creek in the Sabine River basin, and one is on a northwardflowing tributary to White Oak Creek. Only 10 percent of these particular sites are in the Post Oak Savanna, with 90 percent occurring in the Pineywoods. This focused distribution across the cultural landscape of the burial places of notable Caddo peoples is a further reminder that this particular area had a regionally complex sociopolitical organization during Titus phase times (Perttula 2004:401; Story 1990:339–340). Seriation analysis of ceramic vessels and arrow point styles from Titus phase burials and cemetery sites (see Perttula 1992:appendix A; Turner 1978) suggest that most of the notable Caddo burials listed in table 13-5, and the large community cemeteries, date from the early sixteenth century until at least the mid-seventeenth century. Their abandonment may have been “a reflection of reduced social complexity and the scope of community integration, perhaps accompanied by a spatial coalescence and/or decrease in settlement density” (Perttula 1998:82). The Titus phase Caddo cemeteries found in the Big Cypress Creek basin represent archaeological evidence of the social, material, and religious symbolic beliefs about death, the soul, and the afterlife among a series of related communities and affiliated farmsteads. The adults and children buried in these cemeteries are believed to have been lineally related, perhaps in some form of corporate descent group or lineage (Goldstein 1980:7–8). In these situations, the placement of burials in cemeteries (areas used exclusively for the dead) are expected to reflect these lineal relationships, with burials with the same orientation and proximity in rows likely to represent distinctive groups of related individuals. The placement of burials one to another, and the kinds of funerary offerings placed with the dead, are intimately tied to each person’s life, their ties to others in the community, and their social memories within that community, as the deceased and their material surroundings “become significant in the orientation of persons, both deceased and alive, in relation to their past, present, and future” (Hallam and Hockey 2001:190). At a broader spatial scale, the distribution of Titus phase Caddo cemeteries across the Big Cypress Creek basin represent tangible evidence of the Caddo’s past in the landscape, as also do the abandoned farmsteads and compounds. The burials themselves, and the consecration of each cemetery as sacred ground, gave each place a long-lasting meaning and cultural significance in everyday life. Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

407

Conclusions

Copyright © 2012. Nebraska Paperback. All rights reserved.

The communities of Caddo peoples that lived in the Big Cypress Creek basin in the fifteenth to seventeenth centuries were linked by social, political, and religious beliefs embodied in the creation and use of large community cemeteries—several of which contain the graves of more than 80–100 individuals from several generations of use—and the construction of small earthen mounds at key locations within the basin. These mound centers appear to have been the ceremonial and political facilities of at least five political communities (see fig. 13-2) or politically integrated social groups of Caddo peoples. The development of these social institutions, unique to these agricultural Caddo societies in the northeast Texas Pineywoods, in times of climatic fluctuations created the context for different forms of social life and interaction that were in turn sustained for 10 or more generations among these Titus phase groups.3 These Big Cypress Creek basin Caddo groups had leaders who apparently promoted social solidarity and the incorporation of diverse groups and communities into a number of linked polities. Based on comparisons with other contemporaneous small-scale agricultural societies in the Mississippian world, “power and political decision making [was] diffused across many elements of society and individual wealth and status [was] de-emphasized. These kinds of polities are united by ideologies that stress the corporate solidarity of society” (King 2006:76). Conflict and warfare were not part of the social equation that led to the establishment of relatively permanent political communities, probably because of the expansive territory that was available for settlement, as well as the absence of competing polities (e.g., Dye 2008:12). Each of these polities had at its heart a newly created larger and community-centered Caddo mound and village settlement, places where the most important and life-giving ceremonies, rituals, and decisions were made by the social and political elite that guided and organized the changing Titus phase societies living along Big Cypress Creek and its many tributaries; each had their own distinctive social identity. Smaller farming households were dispersed for several miles around the mound centers and community cemeteries. Life for these Caddo was organized around the rhythm of planting and harvesting the cultivated plants, men hunting large game, the rituals and ceremonies of the seasons, and daily life in the households and village settlements. 408

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

At death, these Caddo peoples were laid to rest in sacred cemetery plots used by related families or in large community cemeteries set up and maintained for several generations near the seat of political authority in each of the political communities (see fig. 13-2). These cemeteries were usually established at some distance from the domestic compounds. Caddo children who died were kept close to the living, as they were buried beneath or near their households. The deceased men, women, and adolescents, as well as the social and political elite, were buried in ceremonies that lasted several days, and they were accompanied by various offerings placed in the graves that were meant to help them in their journey to the afterlife. The Caddo peoples who lived in the Big Cypress thrived for ca. 250 years, “well into the protohistoric period” (Early 2004:573). However, by the time Europeans began to regularly explore northeast Texas in the first quarter of the eighteenth century, these Caddo peoples and their communities had abandoned the region. That “shatter zone” (Ethridge 2009:2) seems to have taken place between ca. 1680 and 1720. While it is likely that some Titus phase peoples moved to live with the Red River Kadohadacho or among the Hasinai Caddo south of the Sabine River, current explanations for the demise of the Titus phase hinge on the introduction and the continued exposure of Caddo groups to European epidemic diseases (Thurmond 1990:233; Perttula 1992), which led to the abandonment of the Pineywoods and the disappearance of Titus phase Caddo peoples. Subsequent to the discontinuation of large community cemeteries by the late seventeenth century, most of the upper Sabine River and Cypress Creek basins were abandoned. The only post-1680 Caddo occupations that can be related to earlier use of the Pineywoods are to be found in the lower Sulphur and Sabine Rivers at known trade portages, along trail crossings of major streams, or in pockets of settlements in the Little Cypress (Perttula and Nelson 2007) and peripheral parts of the upper Sulphur River basin (Perttula et al. 2009). None of the Caddo communities that lived in the Big Cypress Creek/upper Sabine basins before ca. 1680 were ethnographically described. Notes 1. With calibrated radiocarbon dates on annually grown plant remains (i.e., seeds, nuts, maize), our best outcome relying on 2 sigma age ranges is to have a 100- to 200-year resolution or age range for any one time. Thus most of the

Fifteenth- to Seventeenth-Century Caddo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

409

Copyright © 2012. Nebraska Paperback. All rights reserved.

calibrated radiocarbon dates from a phase that may have lasted only 250–300 years will considerably overlap. Temporally constrained models (see Nicholls and Jones 2001) may be appropriately applied if there is stratigraphic evidence from particular sites that suggests samples are from a historical sequence rather than from a set of contemporaneous features. In an innovative analysis in the American Southwest that may be applicable to northeast Texas Caddo sites, Gregory and Diehl (2002:204–206) relied on extensive radiocarbon dating in combination with measured fluoride content of animal bone from the fill of pits and structures to develop detailed chronological arrangements of hundreds of features, linking the relative fluoride chronology to calibrated radiocarbon dates from fluoride-dated features. 2. Based on extensive ethnographic research, the absolute quantity of funerary offerings is apparently a less important reflection of social status of the deceased than are such characteristics as the internal organization of cemeteries, the overall amount of energy expended on a specific interment (i.e., grave construction), the disposition of the body, the kinds of burial treatments (i.e., in cemeteries, in mounds, in shaft tombs, etc.), the number of persons placed in a grave, and the kinds of funerary offerings found associated with the deceased (Carr 1995:table XXII). 3. I recognize the essentially unwieldy nature of phases in archaeological discussions of the Caddo or other southeastern U.S. archaeological cultures (see Dunnell [2008a:61–64] for a scathing consideration of the use of phases in modern archaeological discourse). I have tried, therefore, to resist the temptation to link the use of the term phase exclusively with the notion of “‘people’ or ‘cultural group’ as a more or less discrete entity in both time and space” (Dunnell 2008a:64).

410

Perttula

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

14. The Belcher Phase Sixteenth- and Seventeenth-Century Caddo Occupation of the Red River Valley in Northwest Louisiana and Southwest Arkansas david b. kelley

Introduction The Belcher phase represents the peak of Caddo cultural development in the Red River valley of Northwest Louisiana. The phase, or focus as it was originally termed in the McKern system, was first proposed by Alex D. Krieger (1946) based on data from Clarence Webb’s excavations at a mound site near the small town of that name in Caddo Parish, Louisiana (Webb 1959). Webb’s work at the Belcher site (16CD13) remains the basis for much of what we know about this phase, but research conducted over the past 50 years has significantly increased our understanding of a number of its aspects.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Chronology Chronometric dating of the Belcher phase has proven difficult in part because of variability in the radiocarbon calibration curve after 1650 (Goslar et al. 2005). At present, dating of the phase rests largely on seven radiocarbon dates and six thermoluminescence dates. Five of the radiocarbon dates are from the Belcher site (Thurmond 1990:table 59; Webb 1959:207). One of these dates is anomalously early, but the other dates cluster in the 1500s and 1600s when calibrated to the dendrochronology scale. The two remaining radiocarbon dates are from the McLelland site (16BO236) in southern Bossier Parish, Louisiana (Kelley 1997:table 45). When calibrated, both have multiple intercepts with the calibration curve, and the highest probabilities for the actual dates of the samples are in the late 1600s or very early 1700s. Another date from the McLelland site on mussel shell and two dates from the nearby Joe 411

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Clark site (16BO237) are anomalously early. Twelve radiocarbon dates were obtained on charred plant remain samples from the Cedar Grove site (3LA97) in Lafayette County, Arkansas, but all of them have standard deviations of over 300 years and are therefore useless with respect to their chronological precision (Wolfman 1984:table 17-1). The thermoluminescence dates mentioned above are also from the Cedar Grove site, but these are more accurate (Wolfman 1984:table 17-3). They come from shell-tempered pottery from two features at the site, and yielded average dates in the late 1500s and mid-1600s. Taken as a whole the chronometric dates on the Belcher phase suggest a range from about 1500 to 1700. Settlement Patterns

Copyright © 2012. Nebraska Paperback. All rights reserved.

The distribution of Belcher phase sites is primarily along the Red River valley between Texarkana, Arkansas, and Shreveport, Louisiana (fig. 141). Most of the known sites are located in the floodplain, in contrast to sites of the preceding Bossier phase when occupations in the uplands are also common. The Belcher phase settlement pattern appears to have consisted of small mound groups and farmsteads located along the natural levees of active and abandoned river channels. Webb’s (1959) Belcher site report remains the only published report on a Belcher phase mound excavation. The mounds there were constructed in stages as platforms for structures. During the Belcher phase occupation, there were two conjoined mounds, labeled A and B, and a third mound located north of them that Webb referred to as a “built-up plateau.” Throughout much of the site’s history two structures appear to have stood simultaneously on the conjoined mounds, suggesting that one may have served as a specialized religious structure, while the other was the residence of the caddi. After a period of use, each structure on the two mounds was ritually burned and immediately covered by a layer of earth. A series of burials, including single and multiple interments with a wide range of grave goods, were then dug through the fill within the wall lines of the buried structure. One of the multiple burials in each group contained an individual who was accompanied by a number of high status items, including marine shell cups, shell beads, and large quantities of pottery vessels, as well as the bodies of family members or other persons selected for sacrifice. After the burials were in place, the fill overlying the burned 412

Kelley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

14-1. The location of Belcher phase sites mentioned in the text.

structure was leveled, and a new structure was built on the resulting platform. Webb (1959:110) suggested that the death of the high status individual initiated the cycle of burning of the old temple or residence, burial of it and this individual, and construction of the new building. Other Belcher phase sites with mounds include the Battle (3LA1), Foster (3LA27), and Friday (3LA28) sites in Lafayette County, Arkansas, and the McClure (3MI29) site in Miller County, Arkansas, all of which The Belcher Phase

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

413

Copyright © 2012. Nebraska Paperback. All rights reserved.

were visited by Clarence B. Moore (1912). Unfortunately, the mounds at many of these are now gone, and we can only assume that they were contemporary with the burials that Moore excavated. That is not the case at the Battle site, where the great mound is still in relatively good condition. The University of Texas, with Alex Krieger as the principal investigator, conducted excavations in the mound in 1948 under the direction of Lynn Howard, then a graduate student at the University of Michigan (McKinnon 2010b). The work has never been written up, but Howard’s field notes suggest that the initial construction of the mound took place during the Haley phase (ca. A.D. 1200–1500) and that the latest stage was built during the Belcher phase (Howard 1948). While the mound sites are the most obvious remnants of Belcher phase culture and have received most of the attention from archaeologists, the majority of the Belcher phase population resided in farmsteads dispersed along the river valley. This pattern is dramatically illustrated in the map produced by the 1691 Terán expedition of a Caddo settlement located on the Red River in northeast Texas (Swanton 1942:Plate 1; see also Sabo, this volume). The entire settlement included over twenty farmsteads and extended several kilometers along the river valley. Each farmstead included one or two circular houses, a ramada or drying rack, and a covered storage platform, all enclosed within a cleared area surrounded by a line of trees or bushes. Archaeological data on Belcher phase farmsteads were virtually nonexistent until about 30 years ago. One of the first such sites to be investigated was the Cedar Grove site, excavated by the Arkansas Archeological Survey in 1980 after it had been exposed by construction of a revetment along the Red River (Schambach et al. 1982; Trubowitz 1984). This site, which was located about 5 km south of the Battle site, rested on a point bar ridge that had been buried by about 1 m of crevasse deposits from a 1927 flood. Further complicating the excavation was the fact that the site had also been the location of an African American church and cemetery prior to the flood and contained more than 100 nineteenth- and early twentieth-century burials. The Caddo features identified at Cedar Grove included one well-defined house pattern, portions of two other possible houses, several small to medium-size pits, 14 burials, and 2 dog burials. Trubowitz (1984:270) argued that two of the houses and seven of the human burials were associated with the Belcher phase occupation of 414

Kelley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

the site. The remaining features were associated with a later Chakanina phase occupation (ca. 1700–1730). In the early 1990s another construction project on the Red River led to the location and excavation of two small Belcher phase farmsteads in southern Bossier Parish, Louisiana. The sites in question, McLelland and Joe Clark (16BO237), were identified during a survey by Coastal Environments, Inc. of access roads for Lock and Dam No. 5 of the Red River Waterway (Hunter et al. 1992). They were located about 200 m apart on natural levee deposits associated with an abandoned channel of the river. The excavations at the McLelland site identified two houses, a ramada, 67 small and medium-size pits, and seven burials (Kelley 1997:fig. 15). The Joe Clark site contained a single house, 11 pits, and a burial (Kelley 1997:fig. 25).

Copyright © 2012. Nebraska Paperback. All rights reserved.

Architecture All of the Belcher phase houses and temples excavated to date are large circular structures; however, they exhibit some minor differences that apparently have temporal or functional significance (fig. 14-2). The most obvious of these is in the entrances. The earliest Belcher phase structures at the Belcher site, Houses 2, 6, and 7, had extended entranceways, while the latest, Houses 1 and 5, lacked them. There was also a size difference between the structures located on Mound B (Houses 1 and 2) and those on Mound A (House 7) and the low platform (Houses 5 and 6). The structures on Mound B ranged from 9.1 to 9.5 m in diameter, while the other structures varied from 11.3 to 12.2 m in diameter. This may be related to the possible functional distinction between residences and temples alluded to above. Burned daub was common in all of the structures at Belcher, leading Webb (1959:59) to argue that the walls were covered with wattle and daub and the roofs with grass thatch. The one well-defined house at Cedar Grove was 9.6 m in diameter, similar in size to the Mound B structures at the Belcher site, and like those buildings, it appears to have been built with walls covered with wattle and daub. It is unclear whether it possessed an extended entranceway because it was not possible to expose the entire outline of the structure during the excavations (Trubowitz 1984:fig. 9-14). The three houses found at the McLelland and Joe Clark sites were similar in size to the largest structures at Belcher, and all lacked extended entranceways. Structure 1, the better-preserved of the two structures at The Belcher Phase

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

415

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

14-2. Belcher phase structure patterns: a. house 6 at the Belcher site; b. structure 1 at the McLelland site.

Copyright © 2012. Nebraska Paperback. All rights reserved.

the McLelland site, was 11 m in diameter, as was the lone structure at the Joe Clark site. Structure 2 at McLelland appeared to be slightly larger, about 12 m in diameter, but it had been impacted by levee construction in the 1920s, and the wall line was difficult to follow in places. Burned daub was scarce in all of these structures, but burned dirt dauber nests were present, suggesting that these structures burned and were covered entirely with grass thatch rather than wattle and daub (Kelley 1997:31– 32). If that interpretation is correct, then at least some Belcher phase structures were built like the beehive-shaped houses described by European explorers (Swanton 1942:148–153).

Copyright © 2012. Nebraska Paperback. All rights reserved.

Material Culture Ceramics Belcher phase ceramics are probably the most distinctive and thoroughly documented aspect of the material culture. Webb’s (1959:117–161) analysis of the ceramics from the Belcher site remains the basis for much of our current understanding of the ceramic tradition. Based on the stratigraphy in the mounds, Webb identified an early Belcher phase component, which he termed Belcher III, that included Houses 2, 6, and 7, and a later Belcher phase component, labeled Belcher IV, that included Houses 1 and 5 (Webb 1959:150–153). The principal ceramic types associated with the two components were Belcher Engraved, Hodges Engraved, Glassell Engraved, Dunkin Incised, Foster Trailed-Incised, Cowhide Stamped, Belcher Ridged, Karnack Brushed-Incised, and Briarfield Plain (fig. 14-3). Belcher Engraved was by far the most numerous of the fine wares, followed by Hodges Engraved. The utilitarian wares were dominated by Belcher Ridged, Karnack Brushed-Incised, and in the latest occupation, Briarfield Plain. Present in much smaller frequencies were Dunkin Incised, Foster Trailed-Incised, and Cowhide Stamped. Much of the research on Belcher ceramics since Webb’s original study has focused on refining the typological units that he defined in order to identify smaller temporal or spatial units. Frank Schambach and John Miller’s 1984 analysis of the Cedar Grove ceramics is the best example of this. They used the class-pattern-design system pioneered by Schambach, J. Cynthia Weber, and Ann M. Early for describing relatively minor amounts of design variability in aboriginal ceramics as the basis for a type-variety classification that resulted in the identification of one new type and 30 new varieties. The ceramics from the burials at The Belcher Phase

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

417

Copyright © 2012. Nebraska Paperback. All rights reserved.

14-3. Belcher phase pottery from the Belcher site: a. Belcher Engraved carinated bowl; b. Belcher Engraved bottle; c. Glassell Engraved carinated bowl; d. Belcher Ridged jar; e. Foster Trailed-Incised jar.

Cedar Grove were then seriated, and the resulting groupings were used to identify two temporal components at the site, one associated with the late Belcher phase and the other with the historic Caddo Chakanina phase. The Belcher phase component at Cedar Grove included the fine wares Belcher Engraved, vars. Belcher and Ogden; Hodges Engraved, vars. Armour, 418

Kelley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Sentell, Candler, and Kelly’s Lake; smaller quantities of Avery Engraved, vars. Baker, Bradshaw, and Graves; and Glassell Engraved, vars. Akins and McGee. The coarse utility wares included Foster Trailed-Incised, vars. Dobson, Foster, and Red Lake; Belcher Ridged, var. Wilson’s Island; and Karnack Brushed-Incised, vars. Karnack and Fish Bayou. This assemblage differed in only a few respects from that at the Belcher site, and Schambach and Miller (1984:166–167) argued that the differences were probably related to time. These included the absence of Cowhide Stamped at Cedar Grove, the higher frequency of shell tempering there, and the replacement of Belcher Engraved, var. Soda Lake by var. Ogden. My analysis of the McLelland and Joe Clark ceramic assemblages also attempted to address culture historical questions, specifically concerning the contemporaneity of the two sites. Temporal markers among the ceramics indicated that the principal occupations at both began late in the Belcher phase. The most common fine ware types were Natchitoches Engraved, Hodges Engraved, Glassell Engraved, and Keno Trailed. The coarse utility wares were dominated by Belcher Ridged and Karnack Brushed-Incised (Kelley 1997:table 7). Present in lower frequencies were two other coarse ware types, one that I originally identified as Emory Punctated-Incised and the other as Ebarb Incised, a type defined by Gregory (1973:313–333) at the Spanish mission and presidio of Los Adaes in Northwest Louisiana. The material labeled Emory Punctated-Incised at the McLelland and Joe Clark sites I now consider to represent varieties of Foster Trailed-Incised. As I noted in 1997, the body sherds of the two types were unsortable, and only the rims could be used to identify them. After discussions with other Caddo archaeologists, I now think it makes more typological sense to treat the two varieties of Emory Punctated-Incised that I identified, vars. Horseshoe Bayou and Cane River, as varieties of Foster Trailed-Incised. Horseshoe Bayou has a short rim with rows of tool punctations and a body with concentric incised or trailed circles, as in early varieties of Foster Trailed-Incised. It was present at the Belcher site where Webb (1959:figs. 110g, 110i) classified it as Foster Trailed-Incised. The other variety, Cane River, has a similar rim and a body with incised interlocking scrolls, as in later varieties of Foster Trailed-Incised. Cane River appears at the Southern Compress site (16NA14) in Natchitoches, Louisiana, and at Los Adaes (16NA16), and apparently continues later than the Horseshoe Bayou variety of Foster Trailed-Incised. I also examined the vessel form composition of the McLelland sherd The Belcher Phase

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

419

Copyright © 2012. Nebraska Paperback. All rights reserved.

collection as illustrative of a Late Caddo Belcher phase domestic assemblage. This assemblage was dominated by jars (54.6 percent) and simple bowls (26.4 percent). Carinated bowls (16.5 percent) and bottles (0.9 percent) made up relatively small proportions of the assemblage. In this respect it differed from roughly contemporary mortuary assemblages, such as those from the Belcher and Cedar Grove sites, which had much higher percentages of carinated bowls (36.9 percent and 30.2 percent, respectively) and bottles (22.8 percent and 19.0 percent, respectively). The McLelland site ceramic assemblage also contained examples of a rare vessel form in the Red River valley, the seed jar: seven sherds representing at least four vessels were present (Kelley 1997:39–40). None were recovered from the Belcher or Cedar Grove sites. It is not known whether the McLelland vessels were locally made or trade items from the Middle Ouachita region of south-central Arkansas, where they were common. The ceramics from the McLelland and Joe Clark sites also produced evidence of trade with groups in the Lower Mississippi Valley. Cracker Road Incised and De Siard Incised, two types associated with the Koroa, an aboriginal group found in northeast Louisiana and southeast Arkansas, were recovered in small numbers at the McLelland site, and one sherd of Fatherland Incised, a Natchezan type, was found at the Joe Clark site (Kelley 1997:46–50). Despite the historical documentation for trade in salt, buffalo hides, and horses between the Caddo and tribes in the Lower Mississippi Valley, the archaeological evidence for this exchange at Belcher phase sites is presently limited. Jeffrey Girard (2006) has presented data on another Belcher phase ceramic assemblage, this one from the Beene Plantation site (16BO19), which is located on the east side of the Red River just north of Bossier City, Louisiana. This site was originally identified by Clarence Webb, but due to subsequent development in the area, much of it has been disturbed. Girard analyzed Webb’s collection from the site as well as smaller recent collections. Approximately 70 percent of the ceramic assemblage is shell-tempered, as compared with 53 percent at the McLelland site and 86 percent of the coarse utility wares at Cedar Grove. As at Cedar Grove, the most common coarse ware type is Foster TrailedIncised, followed by Belcher Ridged. The fine wares include the types Hodges Engraved, Natchitoches Engraved, Belcher Engraved, and Keno Trailed. 420

Kelley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Comparing the available Belcher phase domestic ceramic assemblages using type frequencies presents some interesting differences. Here I have been forced to exclude the Beene Plantation assemblage because Girard (2006) does not present counts for ceramic types, only decorative classes. Looking first at the coarse wares, the two Belcher site assemblages are very similar and the McLelland assemblage differs only slightly from them (table 14-1). Belcher Ridged is by far the most abundant type in these three contexts, accounting for 65 to 78 percent of the total coarse wares. Karnack Brushed-Incised is second in frequency with 17 to 21 percent of the total, and Foster Trailed-Incised is a distant third with 5 to 14 percent. The Cedar Grove assemblage differs markedly from the other three components. There Foster Trailed-Incised represents 60 percent of the total coarse wares, while Belcher Ridged accounts for 39 percent, and Karnack Brushed-Incised makes up the remaining 1 percent. What is the explanation of the changes in coarse ware types between these assemblages? Girard (2006:37–38) has made similar comparisons based on decorative classes rather than types and suggested that time is the principal factor. At the Cedar Grove and Beene Plantation sites the relative frequency of incising and trailing (most of it classifiable as Foster Trailed-Incised) increases further, while ridged decoration (Belcher Ridged) declines markedly. The occupation at McLelland postdates those at Belcher and may be slightly earlier than those at Cedar Grove and Beene Plantation, and the frequency of Foster Trailed-Incised seems to reflect this. The fine wares exhibit considerably more variation. In the earlier Belcher site assemblage, Hodges Engraved (50 percent) is the most common type, followed by Glassell Engraved (26 percent) and Belcher Engraved (24 percent). In the later Belcher assemblage Belcher Engraved (57 percent) is the most numerous type followed by Hodges Engraved (23 percent) and Glassell Engraved (19 percent). At McLelland the most abundant fine ware type is Keno Trailed (51 percent), which shows up only in small numbers in the later Belcher site assemblage, followed by Natchitoches Engraved (27 percent), Hodges Engraved (15 percent), and Glassell Engraved (7 percent). Keno Trailed (44 percent) is also the most common fine ware type at Cedar Grove, followed by Natchitoches Engraved (26 percent), Hodges Engraved (21 percent), and Belcher Engraved (10 percent). The question then becomes how to interpret this The Belcher Phase

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

421

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

— 558

131 146 281

2610 180 3359

569

23.5% 26.2% 50.4% 0.0% 0.0%

77.7% 5.4%

16.9%

5 474

270 90 109

2112 155 2868

601

57.0% 19.0% 23.0% 0.0% 1.1%

73.6% 5.4%

21.0%

Belcher IV

1 21 482 862 163 319

865 1921 1330

273

0.3% 6.6% 15.0% 27.0% 51.1%

65.0% 14.4%

20.5%

McLelland

80 1 1642 2022 346 793

325 501 834

8

10.1% 0.1% 20.7% 25.5% 43.6%

39.0% 60.1%

1.0%

Cedar Grove

Notes: 1. The count for Foster Trailed-Incised from the McLelland site includes material originally classified as Emory Punctated-Incised. 2. The counts for Hodges Engraved and Natchitoches Engraved from the Cedar Grove and McLelland sites include sherds identifiable to these two types plus sherds from a combined category, Hodges/Natchitoches. At both sites I have used the proportions of identifiable Hodges Engraved and Natchitoches Engraved sherds to distribute the sherds in the combined category between the two types.

Belcher Engraved Glassell Engraved Hodges Engraved Natchitoches Engraved Keno Trailed Total

Fine wares

Belcher Ridged Foster Trailed-Incised Total

Karnack Brushed-Incised

Coarse wares

Belcher III

Table 14-1. Frequency of ceramic types in four Belcher phase assemblages

Copyright © 2012. Nebraska Paperback. All rights reserved.

variability. One type, Glassell Engraved, exhibits consistent declines in relative frequency through time. Two other types, Keno Trailed and Natchitoches Engraved, make their appearance late in the Belcher phase and become marker types for historic period Caddo occupations along this part of the Red River valley. The two remaining types, Belcher Engraved and Hodges Engraved, appear to change pattern through time. Belcher Engraved increases in frequency from the earlier component to the later component at the Belcher site, drops precipitously at McLelland, and then increases again at Cedar Grove. Hodges Engraved declines consistently through the two Belcher components and the McLelland assemblage and then increases at Cedar Grove. The changes in the latter two types raise the possibility that some of the variability may be related not to time but to stylistic differences between villages. One other artifact found at Belcher phase sites should be mentioned: the ceramic pipe. The typical form in use at this time was the elbow pipe, but there is evidence that it underwent changes during this phase. Moore (1912:637) was the first to note that some of the elbow pipes from sites along the Red River exhibited a turned-up distal projection while on others the projection had become conventionalized. Radiocarbon dates are needed to determine when this change was occurring.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Stone Artifacts The Belcher phase stone tool industry consisted of a chipped stone technology geared toward the production of arrow points and a few other bifacial and unifacial tools and a limited ground stone technology directed mainly toward producing celts. The chipped stone industry relied mainly on locally available chert pebbles, often resorting to thermal alteration to improve the workability of the material. Arrow points were the most common tool category, and while several types were often represented in an assemblage, in most cases one type was more numerous than the others. At the Belcher site, Bassett points dominated the earlier Belcher III component. They were also the most common arrow point style in the Belcher IV component, but also present were several examples of straight-stemmed points that Webb (1959:162) classified as Alba points. Straight-stemmed points dominated the slightly later McLelland assemblage, although in this case Hunter (1997a:table 11) classified most of these as Bonham points.1 Also present at the McLelland site were two The Belcher Phase

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

423

Copyright © 2012. Nebraska Paperback. All rights reserved.

examples of Nodena points, a type that increased in popularity in the historic period. Aside from arrow points the most common tools from Belcher phase sites were thin bifacial perforators or drills. Thirteen of these tools were recovered from the McLelland site, 5 from Joe Clark, and 6 from Cedar Grove. The Belcher site excavations produced only two items that are similar, but this may be due to the different and primarily nondomestic contexts investigated there. Examination of examples of these tools from the McLelland and Joe Clark sites under low-power magnification indicates that some, but not all, exhibit polish on their distal ends. A possibly related tool category present at some Belcher phase sites is the thumbnail end scraper. Five such tools were recovered from the McLelland site and 1 from Joe Clark. Thumbnail end scrapers and bifacial perforators are both associated with hide working and may be an indication of the beginning of the deerskin trade in this region. Finally, some mention must be made of the presence of large, chipped stone blades that occur at a few sites, most notably Foster (Moore 1912:figs. 92, 93) and Belcher (Webb 1959:fig. 42q). The raw material for these ceremonial objects was imported, but whether they were obtained as finished items or preforms is presently unknown. In contrast to the chipped stone industry, the ground stone industry had to rely on imported raw materials due to the lack of suitable local alternatives. The only formal ground stone tools found at Belcher phase sites are celts, which usually occur in small numbers. In addition to these a variety of ground stone expediency tools, such as hammerstones and abraders, were used by the Belcher phase Caddo. Small numbers of ground stone ornaments have also been recovered from burials at Belcher phase sites. The best-known examples are limestone ear spools and zoomorphic pendants of limestone and slate from the Foster and Friday sites (Moore 1912:figs. 79, 90, 96, 97). Bone Artifacts A variety of bone tools are found on Belcher phase sites, but the majority of these are related to flint knapping. The most common are flaking tools made from deer ulnae and antler tines (Hunter 1997:figs. 58– 59; Kay 1984:fig. 13-17). These occur with sandstone abraders, tested pebbles, preforms, and finished arrow points in flint knappers’ tool kits placed in burials at the Cedar Grove (Kay 1984:192) and McLelland 424

Kelley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

(Hunter 1997:85–88) sites. Antler tine projectile points are also relatively numerous, occurring at the Belcher, Cedar Grove, and McLelland sites. Less common are tools such as a bone chisel and deer mandible cornshellers from Belcher (Webb 1959:fig. 128) and turtle carapace bowls or cups from McLelland (Hunter 1997:89–90). Bone ornaments have also been recovered from Belcher phase sites. Examples include ear spools, disks, and pins from Belcher (Webb 1959:fig. 128) and a canine tooth pendant from Cedar Grove (Kay 1984:fig. 13-15).

Copyright © 2012. Nebraska Paperback. All rights reserved.

Shell Artifacts Marine shell ornaments occur at a number of Belcher phase sites, usually as burial offerings. The most distinctive of these are the so-called lizard effigy pendants made from conch shells (fig. 14-4a-b). With a few exceptions that almost certainly represent trade, these have been found only at Belcher phase sites. Webb (1959:170–173) excavated 67 of the pendants at the Belcher site, and they have also been recovered at the Foster site (Moore 1912:fig. 89), Cedar Grove (Kay 1984:fig. 13-22), McLelland (Hunter 1997:91–92), and the Clements site (41CS25), which is located on a tributary of the Red River in northeast Texas (Perttula 1992:189; Perttula et al. 2010:fig. 34). The distribution of these items would suggest that they were locally made; to date, however, no evidence of shell working has been found at any of the excavated Belcher phase sites. Two other types of marine shell ornaments found at Belcher phase sites, conch shell cups and gorgets, are part of the widespread Southeastern Ceremonial Complex. Twenty-five cups have been reported from five sites: 8 from Belcher (Webb 1959:169–170), 13 from Foster (Moore 1912:fig. 86), 2 from Friday (Moore 1912:fig. 76), 1 from Cedar Grove (Kay 1984:figs. 13, 14), and 1 from Battle.2 Most of these are undecorated, the exceptions being 3 from the Belcher site and 1 from the Foster site. The most elaborate decoration is that on the cup from Burial 2 at Belcher, which exhibits a combined eagle-rattlesnake motif (Phillips and Brown 1978:fig. 220; Webb 1959:fig. 129). Shell gorgets are less common than the cups, only three examples being known at present. Webb (1959:fig. 135) found one in Burial 26 at the Belcher site that has two concentric circles of fenestrations (fig. 144c). Brain and Phillips (1996:39) classify this gorget as an example of the Krieger style, which belongs to the geometric genre. Moore (1912:figs. The Belcher Phase

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

425

Copyright © 2012. Nebraska Paperback. All rights reserved.

14-4. Marine shell ornaments from the Belcher site: a. necklace of lizard effigy pendants; b. lizard effigy pendant; c. gorget. Reproduced courtesy of the Society for American Archaeology, Washington dc.

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

94–95) recovered 2 engraved gorgets from Burial 10 at the Foster site. One is an example of the triskele genre (Brain and Phillips 1996:121), and the other falls in the cruciform genre (Brain and Phillips 1996:35). Marine shell beads were recovered in large numbers from the Belcher site (Webb 1959:174–175). They were also present at Foster (Moore 1912:596–597), Battle (Timothy K. Perttula, October 2008 personal communication), Cedar Grove (Kay 1984:194–197), and McLelland (Hunter 1997:92–94). Shell tools also occur at Belcher phase sites, but in contrast to the ornaments these are virtually all made from freshwater mussel shells. The most common type is a large perforated shell used as a hoe, which occurs at Belcher (Webb 1959:fig. 137), Cedar Grove (Kay 1984:197), and McLelland (Hunter 1997:94–95).

Copyright © 2012. Nebraska Paperback. All rights reserved.

Subsistence and Diet The areas that have seen the most significant increase in our understanding of Belcher phase lifeways in the last 20 years are subsistence and diet. This is the direct result of the fine screening and flotation of large volumes of soil from excavations that has led to the recovery of sizeable samples of carbonized plant remains and animal bones and the development of techniques for analyzing the isotopic content of human bone. Looking first at the analyzed floral and faunal assemblages, three sites, Cedar Grove, McLelland, and Joe Clark, are presently the source of virtually all of this information for the Belcher phase. At Cedar Grove, approximately 4,000 liters of soil were processed by flotation and 378 samples were analyzed, 19 of those from features. At McLelland, 390 liters from 55 features were floated, and at Joe Clark 71 liters were floated from 12 features. Maize (Zea mays) was the most abundant cultigen at all three sites, occurring in 80 percent of the features at the McLelland site, 67 percent of those at Joe Clark (Gardner 1997:114), and 60 percent of the features at Cedar Grove (King 1984:207). A variety of other cultigens were recovered, but in much smaller quantities, and their representation varied from site to site. Squash (Cucurbita pepo) was present in 7 percent of the flotation samples from features at the McLelland site and 2.5 percent of the post holes at Cedar Grove, but it was absent from the Joe Clark site. Common beans (Phaseolus vulgaris) also occurred in 7 percent of the flotation samples from features at McLelland, but they were The Belcher Phase

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

427

Copyright © 2012. Nebraska Paperback. All rights reserved.

absent from the Joe Clark and Cedar Grove sites. Other cultigens, including chenopod (Chenopodium sp.), sumpweed (Iva annua), and sunflower (Helianthus annuus), were recovered from fine screen samples from the middens at McLelland and Joe Clark, but not from the flotation samples. The chenopod and sumpweed were present in both domesticated and wild forms. The only other cultigen recovered from the Cedar Grove site was bottle gourd (Lagenaria siceraria). Nutshells were abundant at McLelland and Joe Clark, but less numerous at Cedar Grove. Thick-shelled hickories (Carya sp.) were the most common at all three sites, followed by acorns (Quercus sp.) and pecans (Carya illinoensis) (Gardner 1997:tables 32–33; King 1984:table 14-1). While preservation bias and use for fuel rather than food may partly explain the abundance of thick-shelled hickories, the high ubiquity figures for all of the nut taxa at the McLelland and Joe Clark sites suggest that they were an important food source, at least seasonally. The remains of wild fruits were less abundant in Belcher phase contexts, but some, particularly persimmon (Diospyros virginiana), were numerous enough to have made a significant contribution to the diet. Persimmon occurred in 49 percent of the flotation samples from features at McLelland, 17 percent of those from Joe Clark, and 16 percent of the feature samples from the Cedar Grove site. Other fruits were present in smaller frequencies, such as grapes (Vitis sp.), elderberry (Sambucus canadensis), cherry or plum (Prunus sp.), honey locust (Gleditsia triancanthos), and nightshade (Solanum sp.). In general, the floral remains are indicative of a mixed agricultural and foraging subsistence strategy. Large samples of faunal remains were recovered from the Cedar Grove and McLelland sites. The excavations at Cedar Grove yielded 6,626 specimens from a minimum of 226 individuals, while those at McLelland produced 18,941 specimens from a minimum of 92 individuals (Weinand and Reitz 1997:table 17). At both sites fishes contributed the greatest percentage of the minimum number of individuals, 45 percent at Cedar Grove and 36 percent at McLelland, but deer were argued to have provided most of the meat (Styles and Purdue 1984:225; Weinand and Reitz 1997:table 21). The other source of information on the diet of Belcher phase peoples is through analysis of the isotopic content of human bone from burials. Data are available from 16 burials from four sites: Belcher (n=5), Cedar Grove (n=3), McLelland (n=7), and Joe Clark (n=1). With the exception 428

Kelley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

of the burial from the Joe Clark site, all of the individuals exhibit 13C/12C ratios that indicate maize-dependent diets (Tiné and Tieszen 1997:table 43). There also appears to have been a trend toward increasing maize dependence through time. The average value from the Belcher site individuals,-15.88‰, suggests that about 45 percent of their diet came from maize, while the average values from the later Cedar Grove,-14.11‰, and McLelland,-14.01‰, sites suggest that about 57 percent of their diet came from maize (Schoeninger and Schurr 1998:127). Tribal Identity The tribal affiliation of Belcher phase people is a topic that has been examined by a number of previous researchers, but there are still some unanswered questions. The journals of Henri Joutel (1966), Father Anastasius Douay (Cox 1905), Henri de Tonti (Cox 1905), Jacques Gaignard (Bolton 1914:2:83–100), and the Freeman-Custis Expedition (Flores 1984) identify two Kadohadacho villages on the Red River in the late 1600s and 1700s. One of these villages was located on the north side of the river above the Great Bend at Fulton, Arkansas, and the other was below it (Bolton 1915; Swanton 1942:fig. 1; Williams 1964:fig. 1). The upriver village has not been identified, unless one accepts the argument of Sierzchula et al. (1995:61–64) that it is the Moore/Higginbotham site (3MI3/30). This is a multiple mound site that has produced Caddo ceramics that may date to this period of time, but apparently no associated European artifacts. Given the limited amount of archaeological research that has taken place at the Moore/Higginbotham site, the lack of reported European artifacts would not be particularly surprising. The problem is the site’s location south of the Red River. Mildred Wedel (1978) examined the documentary evidence for the location of this village and concluded that the journals of Joutel, Tonti, and La Harpe all place it north of the Red River. Furthermore, the 1806 Nicholas King map, which is based on the journals of the Freeman-Custis Expedition (Flores 1984), shows the village on the north side of a large meander loop of the river. Map overlays done by Wedel (1978:16) suggest that this meander loop had been cut off by 1838 when the first General Land Office survey of this township was conducted. The oxbow lake formed by this cutoff is shown on the plat map derived from that survey and is still visible today north of the river. The lower Kadohadacho village has received more attention from archaeologists. Relying partly on archaeological data and partly on the The Belcher Phase

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

429

journals of the Freeman-Custis Expedition, Webb (1945:79) suggested that the Foster site represented the nucleus of this village. However, Schambach has noted that the Foster site has produced neither historic period Caddo ceramics nor European artifacts (Schambach et al. (1982:93). He argues that the village was located several kilometers downstream from the Foster site, extending from the vicinity of the Friday site to the Cedar Grove site (Schambach and Miller 1984:168). Regardless of its precise location, the archaeological data from these sites make it clear that Belcher phase groups were the antecedents of the people of the lower Kadohadacho village. The descendants of southern Belcher phase groups in Louisiana are less well known. The journals of Tonti and Jean Baptiste Le Moyne, Sieur de Bienville (McWilliams 1981:146–156) place a Yatasi village in the vicinity of Shreveport in the late 1600s. Girard (2006:40) has argued that the Beene Plantation site represents a portion of this village. My own research on the McLelland and Joe Clark sites led to the suggestion that they were associated with the Nakasas, a group encountered by Bienville living near the Yatasi (Kelley 1997:139). This raises the question of whether the Kadohadacho and the Yatasi both developed from Belcher phase groups. It is to be hoped that future archaeological research will allow us to identify specific material culture markers for these two Caddo tribes. Notes

Copyright © 2012. Nebraska Paperback. All rights reserved.

1. The identification of Alba and Bonham points in this late context raises the question of historical continuity, since Alba points typically occur in Early Caddo assemblages and Bonham points in Middle Caddo ones. In neither case is there strong evidence of continuity that I am aware of. But if a new type name is proposed for the Belcher phase specimens, the problem becomes one of sortability. 2. The conch shell cup from the Battle site reportedly came from a burial at the north end of the large mound and is now in the Gilcrease Museum in Tulsa, Oklahoma (Timothy K. Perttula, personal communication October 2008).

430

Kelley

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

15. The Terán Map and Caddo Cosmology george sabo iii

Introduction The well-known Terán map of an Upper Nasoni village along the Red River, produced as a result of the 1691–1692 exploration of eastern Texas led by Domingo Terán de los Ríos, plays an important role in modern understandings of pre- and post-Columbian Caddo settlement patterns. The map reflects a dispersed community of individual farmsteads, a centrally located leader’s residence, and an associated temple mound; this depiction has been identified as an archetype of Caddo settlement patterning (Schambach 1982a; Trubowitz 1984; Perttula 1992). Here I want to explore an additional interpretation: that the map also reflects key elements of Caddo beliefs about their relations with the spirit world. In short, I suggest that the Terán map represents a cosmogram embedded in the layout of the depicted community.

Copyright © 2012. Nebraska Paperback. All rights reserved.

Historical Background The attempted settlement of the Texas Gulf Coast in 1684–1685 by the French explorer René-Robert Cavelier, Sieur de La Salle, prompted a series of Spanish expeditions to locate and arrest his group. Beginning with the Barroto/Romero expedition in 1686, several explorations scoured the coast east to west from Pensacola and west to east from New Spain until the destroyed colony’s remains were discovered in 1689 during the third expedition undertaken by Alonso De León, governor of the province of Coahuila in northern New Spain (Kessell 2002:135–141; Weber 1992:148–152). Secondhand information acquired by De León during the course of his explorations led to Spanish beliefs that the Tejas Indians, who occupied central and eastern Texas, were a settled, agricultural, and monotheistic society whose conversion surely would come more readily than that of their less sedentary neighbors. What is more, De León was persuaded that the Tejas (also known as the Hasinai Caddo) 431

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

already possessed elements of Christian belief and practice, bestowed by miraculous visits from the so-called Lady in Blue (see Colahan 1994; Hickerson 1990). Government and church officials alike viewed this information as direct evidence of “the workings of divine providence” in the events connected with the search for La Salle’s colony (Dunn 1917:111). Plans consequently were made for placing missions in east Texas, leading to several expeditions involving Franciscan priests who hoped to gather the region’s inhabitants into permanent agrarian settlements where the Indians could be “civilized” and converted to Catholicism (Castañeda 1936–1958:2:17–22). The first of these expeditions departed from Santiago de Monclova, the capital of Coahuila, in March 1690. De León led the expedition along with Father Damián Massanet, head of the Colegio de la Santa Cruz in Querétaro, which had been designated by the bishop of Guadalajara (Mexico) to take charge of the Texas missionary effort. The expedition succeeded in establishing the mission San Francisco de los Tejas along the Neches River in the heart of Hasinai country, despite the development of a fractious relationship between the two leaders. Three friars—Miguel de Fontcuberta, Francisco Casañas de Jesús María, and Antonio Bordoy—remained at the mission when De León and Massanet returned to Monclova in June (Chipman 1992:87–90). De León and Massanet submitted separate plans for the further settlement of the east Texas region and the establishment of additional missions. De León’s plan included provision for a string of military presidios to protect the mission communities. Massanet strongly opposed any secular intervention in his zealous quest to harvest Indian souls, and he worked vigorously to undermine De León’s plan as well as the governor’s character and fitness for leadership. In this he succeeded, and in 1691 the Conde de Galve, viceroy of Mexico, appointed Domingo Terán de los Ríos as the first governor of the province of Texas and charged him with leading the next expedition, again in the company of Massanet. This expedition had three main goals: to establish additional missions among the Indians, to explore the province and describe its major natural and cultural features, and to investigate rumors concerning the lingering presence of Frenchmen (Dunn 1917:130–131; Hatcher 1932:6– 7; Chipman 1992:91–94). The expedition departed Monclova in May 1691, but discord quickly 432

Sabo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

developed between Massanet and Terán. Their inability to work toward common goals, in combination with an unanticipated resistance to the Spanish presence on the part of the Hasinai Caddo villagers, led to a failed effort during which they established no new missions and performed only a limited reconnaissance of the territory. One event of importance to present concerns, however, was a trip made in November and December from the Hasinai settlements to Kadohadacho settlements along the Great Bend of the Red River. The Kadohadacho hospitably received the Spanish visitors, and Terán and his companions (including Massanet) spent a week reconnoitering the area to ascertain that no Frenchmen remained in the region. Terán included a descriptive account of the Kadohadacho community in his diary, and later a fascinating map, made by an anonymous cartographer, was also produced that depicted the community layout. Curiously, no mention of the map or of any map-making activity is included in the primary written sources concerning the expedition.

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Map The map, titled Mapa de la provincia donde habita la Nacion Casdudacho, is deposited in the Archivo General de Indias, Mapas y Planos, in Seville, Spain. Herbert E. Bolton first reproduced it in his 1915 Texas in the Middle Eighteenth Century: Studies in Spanish Colonial History and Administration, and it is described in The Lowery Collection: A Descriptive List of Maps of the Spanish Possessions within the Present Limits of the United States, 1502–1820 (Lowery 1912:186–187). A facsimile copy of the map (fig. 15-1) exists in the Eugene C. Barker Texas History Collection at the Center for American History, the University of Texas at Austin. It depicts an Upper Nasoni community located along the Great Bend of the Red River. The Nasonis were part of the Kadohadacho alliance, one of three Caddo “confederacies” identified in John R. Swanton’s monumental 1942 ethnohistory of the southern Caddo people. Terán’s diary, along with that of Massanet and other participants in the expedition, is preserved in the Archivo General y Publico de Mexico, Historia: Coleccion de Memorias de Nuevo España, vol. 27. The Terán and Massanet diaries are translated in Hatcher (1932), although Mildred Mott Wedel (1978:6) noted omissions and errors. These shortcomings do not affect the brief passages from the Hatcher translation that I quote below. The Terán and Massanet diaries both provide daily travel itineraries, The Terán Map and Caddo Cosmology

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

433

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

15-1. Facsimile of the Terán map, 1691. (Courtesy J. P. Bryan Map Collection, CT0108, Center for American History at the University of Texas at Austin.)

Copyright © 2012. Nebraska Paperback. All rights reserved.

Copyright © 2012. Nebraska Paperback. All rights reserved.

brief descriptions of the country traversed and its inhabitants, and names bequeathed to rivers, landscape features, and Spanish and Indian elements of the cultural landscape. The map contains three primary elements. First, along the top of the map, there is a banner bearing the title “Rio de San Andres de Galve,” the name Terán gave to the Red River. A large cartouche, bearing the Terán family crest, dominates the left edge of the map. A translation of the text written below the crest, provided on the Texas Beyond History website, reads as follows: “This new discovery of the Caddudacho Nation was . . . made by General Don Domingo Theran de los Rios, Lt. Gen. of all New Spain, in the year 1691. This visit was made in the said section and province on the 28 of November with only thirty men, in the said year.” The third map element, showing the layout of the community, dominates the center portion. The community depicted on the map consists of twenty-five building clusters or compounds, most of which are surrounded by tree rows that may be stylized representations of woodlots. About half of the trees are barren (consistent with the winter visit of Terán’s party); the others are shown partly or fully covered with leaves. Four compounds are located along the north bank of the Red River; the rest are distributed along the south bank. Two relict channels are also drawn below the south bank; these are named “Laguna de San Blas” and “Laguna de la Canoa.” While there are no clear indications of crop fields, most of the compounds appear to represent individual farmsteads with one or two dwellings, elevated storage platforms, and open platforms. The dwellings are shown with cylindrical walls and clearly distinguishable “beehive” roofs that appear to be thatched. The elevated storage platforms are also covered with thatched “beehive” roofs. The open platforms appear to have mat, brush, or bark coverings. Two building clusters vary from this pattern. The one located on the far left, or western, end of the community consists of a structure, much like the dwellings in form, but here placed on top of an earthen mound. In fact, this building appears to be partly buried within the mound. The word Templo appears directly above the roof peak. An open platform is depicted just off the right-hand side of the mound and just behind it from the viewer’s perspective. A second compound, located in the center of the community, is that of the caddi, or community leader. This compound includes two dwellThe Terán Map and Caddo Cosmology

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

435

Copyright © 2012. Nebraska Paperback. All rights reserved.

ings, one larger than the other. The word Cadi appears below the larger dwelling, the caddi’s residence where community officials gathered periodically for town meetings. The smaller of the two dwellings likely housed the community officials’ assistants. Two upright posts are shown in front of the caddi’s dwelling, and a Christian cross is shown to the right and behind this dwelling. A rectangular, gable-roofed (European style) structure is also shown below the caddi’s dwelling. This is a curious element of the map, for the Spaniards did not establish a mission on this visit nor is there any mention of other Spanish construction. On the other hand, Terán and his companions resided at the caddi’s compound during their weeklong visit, so perhaps this represents a shelter they built for themselves during their stay. The compound also features one open platform but no elevated storage facility, in keeping with the community practice of supplying the material needs of the highest ranking civil and religious leaders. Terán named the caddi’s residence “Santa Galia,” and those words are written in the lower part of this compound. It is also perhaps noteworthy that five dwellings shown directly alongside the two lagoons have no accompanying platforms. Finally, three canoes are shown on the map. Two appear in the main channel of the Red River; one is propelled by a long pole while the other is being paddled. The canoe shown in the Laguna de la Canoa is propelled by fore and aft poles. Finally, it is worth mentioning that no indications of cemeteries are included on the map. From archaeological investigations, we know that cemeteries were important components of Caddo villages in pre-Columbian and early historic times. Mildred Mott Wedel examined the Terán map along with related documentary sources as part of her investigation of the location of JeanBaptise Bénard de La Harpe’s 1719 trading post on the Red River. She concluded that the Upper Nasoni community visited by Terán was also visited in 1687 by Joutel, observed in 1690 by Henri de Tonty, and visited by La Harpe in 1719; thus, “This village, east of the temple mound, was probably the main concentration of Upper Nasoni from 1687 to 1720, as well as in prehistoric times” (Wedel 1978:14). Wedel also determined that François Grappé, a trader of French-Caddo descent who grew up along this stretch of the Red River, likely referred to the remains of this same community when, in the early 1800s, he told John Sibley, the U.S. Indian agent at Natchitoches, about the location of a site where the lo436

Sabo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

cal Indians “pay a devout and sacred homage” (Sibley 1805:729). Finally, Wedel associated the community depicted on Terán’s map with the Hatchel-Mitchell-Eli Moores and Tilson (now known as the Horace Cabe site, 41BW14) complex of archaeological sites located along a stretch of the Red River northwest of the modern city of Texarkana. The Hatchel Mound (41BW3), thought to be the very temple mound depicted on the map, was the subject of extensive excavations conducted by the University of Texas and the Works Progress Administration in 1938–1939 (Creel 1996; Perttula 2005). Archaeological Uses of the Terán Map Although the Terán map has long been of interest to historians (e.g., Bolton 1987; Griffith 1954) as well as archaeologists (e.g., Harrington 1920), very little analysis or interpretation was made until Frank F. Schambach used it to develop an archetypal model of pre- and post-Columbian Caddo settlement patterning:

Copyright © 2012. Nebraska Paperback. All rights reserved.

Although there are many [historical] references to Caddo “villages,” these were, in fact, dispersed communities composed mainly of small farmsteads, each with one or two houses (the latter possibly sheltering polygynous households), several opensided brush-covered shelters, and a storage platform with a beehive-shaped thatched roof. There is no need to speculate about this pattern. The Terán map of 1691–92 clearly shows such a community. (Schambach 1982a:7) Schambach compared the Terán map’s depiction of an entire community settlement pattern to photographs of a Caddo farmstead near Fort Sill, Oklahoma—Longhat’s camp—taken by Will Soule between 1868 and 1872 (Swanton 1942:Plate 14). Schambach (1982b:121) noted the remarkable similarity, despite the separation in time, in the depiction of farmsteads in these two sources, suggesting that “between these two documents we have historically based models of both a single Caddo farmstead and a complete Caddo settlement to take to the field to test, and to guide research.” Accordingly, Schambach listed a series of questions and hypotheses that dealt with expectations concerning the basic archaeological structure and material content of farmsteads and communities, differences between elite and non-elite residences, ceremonial activities, and burial patterns. Schambach’s model proved to be a useful interpretive device for evaluation of settlement patterning at pre- and post-Columbian Caddo The Terán Map and Caddo Cosmology

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

437

Copyright © 2012. Nebraska Paperback. All rights reserved.

archaeological sites (e.g., Perttula 1992:155–161). There have been two specific attempts to test the model via archaeological fieldwork. The first test, undertaken by Neal Trubowitz in his investigation of the Cedar Grove site on the Red River in southwest Arkansas, attempted to confirm or refute specific elements of the model. This proved difficult, since much of the data retrieved by the excavations turned out to be ambiguous in relation to expectations derived from the model (Trubowitz 1984:266–271). More recently, Schambach and Jami J. Lockhart used the model as a guide for site exploration during their investigation of the Tom Jones site, also in southwest Arkansas. Since excavations associated with this latter project were to be guided by prior remote sensing surveys of the site, geophysical surveys were conducted not only in areas where structural features were expected but also in areas where such features were not predicted. The unanticipated discovery of a series of buildings alongside the site’s large temple mound, contra expectations based on the Terán map, led to a thorough investigation of the temple mound precinct, including excavation of several off-mound buildings. Analysis of the retrieved data has not been completed, but it is perhaps noteworthy that ceramics—including very large cooking vessels—comprised the majority of artifacts associated with the off-mound buildings, along with considerable amounts of animal and plant food remains. These data suggest that ceremonial food preparation may have been a specialized function associated with these structures, which were also deliberately burned and buried following their use (Lockhart 2007). Recent studies of the 1938–1939 wpa excavations at the Hatchel site, which as noted above is believed to contain the temple mound depicted on the Terán map, also cast doubt on the idea that few archaeological remains will be found within the sacred temple mound precincts of dispersed Caddo settlements. These excavations located a large cemetery area immediately east of the Hatchel mound, and five village areas with hundreds of post holes (some resolving into circular house patterns), extensive midden deposits, and pit features, can be identified to the south and southeast of the mound; other village and cemetery areas are known to the northwest and west of the mound. Most of these remains represent occupation during the Late Caddo period Texarkana phase, although site occupation extends at least as far back in time as the Early Caddo Lost Prairie phase (Creel 1996; Perttula and Nelson 2003b; Perttula 2005). 438

Sabo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

More studies like these will surely add important new insights concerning Caddo settlement patterns, associated social and ceremonial activities, and methodological applications of the Terán-Soule model (including the chapters in this book by Walker and McKinnon [chap. 7] and Girard [chap. 9], that also provide critiques of the model). This is well and good, and archaeologists armed with remote sensing technologies and Geographical Information Systems can have a field day—so to speak—figuring out the intricacies and environmental relationships associated with Caddo settlement patterning. I have a slightly different interest, not more important than these other concerns but perhaps every bit as interesting: To what extent does the Terán map provide an illustration of a seventeenth-century Caddo cosmogram, or map of the cognized universe?

Copyright © 2012. Nebraska Paperback. All rights reserved.

The Terán Map as a Cosmogram To answer this question, we must first review some basic features of Caddo cosmology and social organization. Like many Southeastern Indian communities, a basic duality represented by the concepts of good and bad, or order and chaos, finds symbolic representation in a distinction between an Above World and a Below World. The Caddo origin story establishes this distinction: First Man and First Woman lead the Caddo people from their underground world through a passageway to the surface of earth, until Wolf disobeys instructions not to look back, resulting in the opening’s closure, which leaves behind in darkness those who had not yet emerged (Mooney 1896:1093–1094). This is not to suggest that a one-to-one correspondence equates an Above realm with goodness and orderliness while a Below realm reflects the converse; rather, the Above World–Below World distinction establishes a relationship of possibilities that oscillate between positive and adverse conditions. The in-between position of This World establishes the possibility of mediating or acting upon potential conditions. The goal is to maintain an order based on reciprocal or balanced relationships between This World and the spirit realm, and in this belief system reciprocity is not a given but something requiring active maintenance through work (Hudson 1976:120–183; Reilly 2004). In the social realm, these relationships are expressed by roles assigned to various categories of individuals. Several community leaders, headed by the caddi, work to maintain reciprocal relations within the The Terán Map and Caddo Cosmology

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

439

Copyright © 2012. Nebraska Paperback. All rights reserved.

human community, while members of that community work to provide basic material support for their leaders. Another category of leader, a “high priest” called the xinesi, works to maintain reciprocal relations between the human community and spirit beings populating the Above and Below World realms. Across these realms—”separated” by permeable boundaries subject to penetration via ritual—the various categories of beings were ordered by a hierarchical system of ranks. Ranks were based on relative levels of power to maintain reciprocal balances. Therefore, the earthly categories of community members, subordinate leaders (e.g., Tanmas and Canahas), and superordinate leaders (e.g., caddis and xinesis) were matched in the spirit realm by ancestral souls, subordinate spirit beings, and superordinate spirit beings. Since the caddi was the highest-ranking individual responsible for the maintenance of balance within the community, it is appropriate that his residence occupied a central location. From the temple mound, the xinesi’s counterpart in the spirit realm was Ayo-Caddi-Aymay, the supreme being or “Captain of the Sky,” whose symbolic representations included the sun and the eternal fire kept alive within each community’s temple (Sabo 1998; Wyckoff and Baugh 1980). Community members also played important roles in mediating relationships with the spirit realm. Adult men and women, for example, performed an annual cycle of agricultural and hunting ceremonies to maintain positive relationships between humans, food plants, and animals. Some of these, such as first fruits and harvest ceremonies, required the participation of specialists—medicine men or priests—but whether performed individually or with the assistance of adepts these rites served to maintain reciprocal relationships between Caddos and the cosmological forces that transcended their worldly existence. This is a highly simplified summary of a much more complex system of relationships that existed in a changing series of configurations that European Americans described (albeit imperfectly) from the time of first encounters in the sixteenth century. The core features of this system of relations can be identified within the modern Caddo community, but demonstration of that point is beyond the scope of this chapter. I concentrate here on a seventeenth-century instance of these relationships as depicted—or, better, as encoded—in the Terán map and the accompanying narrative accounts. To examine the Terán map from a perspective sensitized to these ele440

Sabo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

ments of Caddo belief, we can begin by considering the temple location at the western end of the settlement. Terán’s description of this part of the community is instructive: when he first approached the settlement, Terán wrote, “We crossed this level space, still going in the same direction [northeast], passed two lakes and came to a grove of trees. From this point we caught sight of one of the rancherias of the nation.” From their vantage point, the temple mound would be the first element to appear, and indeed Terán goes on to say that this first rancheria “was located on a hill, which commanded the whole country . . . the Indian guide informed us through Brother Antonio, who served as an interpreter, that this was the temple in which the Indians worshiped and made offerings to their gods.” Continuing: “We proceeded from this place and made camp at the home of an Indian whom they call Caddi, located about half a league, more or less, from the temple. . . . I named this place Santa Galia.” Then: “Through the caddi . . . I learned that the settlements extended down the stream along the riverbank” (Hatcher 1932:33). This was indeed a sizable community; Schambach (1982a:7) estimates that it extended along the river for a distance “no less than 4 km and possibly 9 km or more.” Terán’s account suggests that the temple complex functioned as a “gateway” through which visitors entered the community. Caddo leaders typically marched out in procession to receive visiting dignitaries, who then were carried to the temple where they were welcomed with ceremonies including speeches, food offerings, calumet smoking, and other cleansing rituals (e.g., Sabo 1995). In view of the previously mentioned comment by the anonymous Indian guide that the temple also served as a place “in which the Indians worshipped and made offerings to their gods,” we may also suggest that the temple represented another “gateway” connecting the human realm (This World) with the spirit realm. If the temple depicted on the Terán map is conceptualized as a gateway connecting the human and spirit realms, then the direction of the temple entryway and the spatial orientation of the rituals performed therein are of great interest. Among the Caddo, the western cardinal direction, associated with the setting sun, symbolizes death and the realm of ancestral souls whereas the eastern cardinal direction, associated with the rising sun, symbolizes life and the realm of the living. East-facing rituals accordingly would be associated with life-affirming events, such The Terán Map and Caddo Cosmology

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

441

Copyright © 2012. Nebraska Paperback. All rights reserved.

as agricultural ceremonies or community renewal celebrations. Westfacing rituals, on the other hand, would logically be associated with lifethreatening events, such as hunting, warfare, or funeral rites (see, e.g., Kay and Sabo 2006). Unfortunately, as no entryway is depicted, the map provides no help on this matter. This does not contradict our interpretation of the temple mound location as a boundary marker, although we would understand much more about the nature of this boundary if these directional details had been incorporated into the map. Archaeological evidence preserved at sites such as Hatchel may be helpful in this regard. According to Perttula (personal communication, 2008), archaeological evidence at Hatchel and other nearby sites indicates that entranceways were oriented in relation to local topographic and/or settlement features rather than by strict adherence to cardinal directions. In fact, this makes a great deal of sense, even with reference to a “native point of view.” Francis La Flesche, for example, described how Osage villages were laid out with reference to an east-west path or street that symbolized the surface of the earth and the path of the sun on its daily journey across the sky. Houses occupied by families belonging to Sky people clans were arranged on the north side of the street, in relation to the symbolic association of north with the Above World. Conversely, Earth people clans lived on the south side of the street. Each clan had a specific place allocated on their designated side of the village, and this spatial organization was replicated in seasonal hunting camps; however, in the latter case the directionality was “relative,” insofar as local topographic features dictated the actual orientation of the camp (Bailey 1995:42–44). The point here is that principles of directionality contributing to village organization may be applied in a relative manner as required by situational factors; orientation to exact directions is not always required. The key to identifying directional patterning, then, lies in detecting consistency in relative orientation across a village or regional cultural landscape. Assuming, then, that the temple location does mark an invisible boundary between the human and spirit realms, we should expect two consequences: (1) religious ceremonies performed at the temple should reflect a perception of such a gateway at the boundary of separate realms, and (2) it should also be possible to identify material or behavioral connections between the spirit world, the temple complex, and the community as a whole. 442

Sabo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

First, did rituals performed at the temple complex reflect the possibility of transcending a boundary between This World and the spirit world? The Terán expedition documents provide no information about welcoming ceremonies performed in the Kadohadacho temple or elsewhere in the community, so we must refer to contemporaneous accounts of other Caddo rituals. During the seventeenth century, the Caddo agricultural cycle was punctuated by a series of rituals that included a forecasting ceremony, a planting ceremony, and a harvest ceremony. These ceremonies were well described by the Franciscan priests who labored among the Hasinai (Bolton 1987; Griffith 1954; Hatcher 1927). In the forecasting ceremony, priests made a tobacco offering to the temple fire, then used an eagle feather “to show that the eagle whose feathers they are using has risen on high to consult with the captain who is there in regard to the weather for the year” (Morfi 1932:35). In this pantomime, the eagle feather is a material symbol serving to “witness” the indirect communication between the priests and the “captain” of the Above World, in anticipation of a revelation concerning weather conditions for the coming year. This is consistent with the belief in a cosmological boundary as well as belief in the priestly capacity to transcend that boundary through ritual action. Jackson (2003:68–74) provides a modern example of the “witness” concept in his analysis of modern Yuchi ceremonial practices. During the planting ceremony, the caddi collected offerings made by women, which he then burned in the temple where the resulting smoke arose from the fire and dissipated into the sky. The purpose of this ceremony was to gain Ayo-Caddi-Aymay’s blessing upon the agricultural work performed by the women of the community. Here, the smoke produced by offerings consumed in the fire mediates another indirect communication with the Above World for a benefit realized in This World (Morfi 1932:31). The same symbolism is reflected in the harvest ceremony, a lengthy and complicated affair that stretched over several days. In preparation for this ceremony, hunters departed from the community in all four cardinal directions to procure meat for communal feasts. People gathered in spatially separated ranks for the ceremony, during which women from every household presented food offerings to priests. These offerings were “sacrificed” in the sacred fire when the constellation Pleiades moved into a particular alignment with the temple (Hatcher 1927:171–173). In sum, all three rituals performed in or near the temple employed The Terán Map and Caddo Cosmology

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

443

Copyright © 2012. Nebraska Paperback. All rights reserved.

the sacred fire and deliberate spatial orientations to communicate across an invisible boundary with spirit beings in the Above World. The ceremonies maintained a system of reciprocity that conferred positive results within the human community. From this we may conclude that the temple mound complex indeed represented a gateway connecting human and spirit realms. What about our second expectation concerning the existence of tangible connections between the temple complex, the community as a whole, and the perceived realm of spirit beings? Here we anticipate three relationships: (1) between the Above World and the sacred fire; (2) between the sacred fire and the temple; and (3) between the temple and community households. What connection can we establish between the Above World and the sacred fire and, in turn, between the sacred fire and the temple? Fray Isidro Felix de Espinosa, who served as the head of the Franciscan missionary effort in eastern Texas at the beginning of the eighteenth century, tells us that the Hasinai “have especial superstitions in connection with the fire and they worship it. There is a house set apart for this purpose where there is always a fire. They have appointed an old man whose duty it is to keep it up always. He is their chenesi [Xinesi] or chief priest.” Moreover, in his description of the harvest ceremony, Espinosa tells us that “after the [final] song, they all await the rising of the sun. Certain young men and boys are sent out into the nearby woods as if calling or speaking to the sun for the purpose of hastening its coming. Just as it begins to rise they run about joyously and gaily and it seems as if they were giving thanks for their past crop or were beseeching the sun to aid them in the projects they are beginning” (Hatcher 1927:173). Espinosa’s statements reveal rather clearly the belief in a symbolic relationship between the sun and Ayo-Caddi-Aymay, whom the Caddos thank for past successes and beseech for assistance in future endeavors. Espinosa recognized the sacred status of the fire burning in the temple and tended by the xinesi. Given our previous observation that offerings to Ayo-Caddo-Aymay were burned in the sacred fire, we may identify a second symbolic relationship connecting the temple fire and Ayo-Caddi-Aymay. What about the connection between the temple and the larger community, represented by its assemblage of individual households? We turn again to Espinosa, who wrote: “All of the houses, or most of them, 444

Sabo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

are supplied with fire from the principal temple—not that it is carried every day, but they are supplied therefrom when the houses are built, and they keep it burning. If it goes out, they consider it a sign that all that family will die and they bring new fire from the fire temple with great ceremony” (Hatcher 1927:161). Here again, a symbolic relationship between the sacred temple fire and household fires is affirmed. The common use of the temple fire as a source for household fires is, in fact, a key feature that confers a sense of community on the separate households. It is important to reiterate here that Espinosa’s observations were made not among the Kadohadacho, whose Upper Nasoni village is depicted on the Terán map, but among the Hasinai groups, who lived farther to the south along the Neches and Angelina Rivers in east Texas. The Kadohadacho practiced the same ceremonies—as observed, for example, by Joutel in 1687 (Foster 1998). So it is permissible to suggest that these performances reflect the existence of a similar set of underlying beliefs. In summary, three sets of associated symbolic relationships have been identified: (1) between the sun and Ayo-Caddi-Aymay; (2) between Ayo-Caddi-Aymay and the sacred fire burning in the temple; and (3) between the temple fire and the fires burning in the community’s individual households. These relationships are expressed, to a significant degree, in terms of deliberate spatial orientations manifested at the household, community, and landscape levels that reflect a cosmological structure understood by the community as a whole. This finding supports two conclusions: first, that the temple located at the western terminus of the community illustrated in the Terán map served as a “gateway” not only to external regions of This World but to an invisible spirit realm as well; and second, that the Terán map itself represents a cosmogram that reveals the Caddo’s view of their place in the world and the relationships that define and confer meaning upon that place. This is very much a concentric cosmogram, with the caddi’s residence marking the hierarchical center of the human community and the temple complex denoting the center point connecting This World with the invisible spirit world. It is not necessary, of course, to insist that Terán’s anonymous mapmaker was aware of these relationships. Fortunately, the care with which the map was produced preserved a distinctive spatial configuration, the meaning of which can be ascertained via analysis of accompanying documentary information concerning the community so illustrated and The Terán Map and Caddo Cosmology

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

445

Copyright © 2012. Nebraska Paperback. All rights reserved.

the nature of ritual activities that occurred at specific places within that community’s cultural landscape. What is revealed by this analysis is a glimpse of an ancient Caddo worldview that persists today, although in a somewhat modified form. What is the value of this interpretation of the Terán map, beyond speculative inferences concerning seventeenth-century worldviews? Following Schambach’s lead of using the map to generate expectations that can be tested archaeologically, I suggest that the interpretations offered here provide another set of hypotheses for future investigation. The cosmogram that I see reflected in the arrangement of places on the map expresses a specific spatial order comprised of a “nesting” of two central places: the caddi’s compound representing the center of internal community social relationships and the xinesi’s temple complex representing a central point of articulation for the community’s external cosmological relationships. Individual farmsteads arranged along the river in relation to those central places create a hierarchically ordered community, wherein activities performed at specific locations are spatially oriented with reference to the central places. This much is ethnographically supported, but even so we lack key details, as reflected, for example, in the Terán map’s lack of indication of the temple’s orientation. What directional orientations have been discerned within excavated Caddo temple mounds, and what associations can be identified with ritual activities conducted within those temples, or at other locations, within the adjacent communities? Further, do orientation patterns reflect a tight alignment to cardinal or inter-cardinal directions, or are they “relational,” as illustrated in our example of Osage hunting camps? Many Caddo activities possessed a ritual component, which in turn imposed a spatial orientation relative to a cosmological order expressed at various levels of resolution, from individual structures or other activity locations, to community layouts, to the regional cultural landscape. The archaeological investigation of spatial ordering at comparable levels of resolution may reveal new insights concerning the use of cosmological principles in pre-contact times and the manner in which these principles persisted or changed over time. The validation of such ordering within and between sites and regions suggests a final conclusion: that Caddo (and other American Indian) settlement organization was influenced or structured by religious principles as much as by other, more generally 446

Sabo

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

cited factors including environmental constraint, social structure, and political economy. There is, in fact, considerable warrant for this perspective. In this very book, James Brown’s analysis (see chap. 5) of the contents of the so-called Great Mortuary in the largest cone of the Craig mound at the Spiro site provides an elegant demonstration of the efficacy of religious principles and ritual performances in structuring the disposition of artifacts and human remains at a specific activity location within a settlement. Similarly, in their chapter, Edwin Jackson, Susan Scott, and Frank Schambach (see chapter 3) provide compelling evidence that implicates religious concepts and ritual activities in accounting for the nature of the faunal assemblage and its distribution at the Crenshaw site, an Early Caddo ceremonial center along the Red River in southwest Arkansas. An archaeological example of the operation of religious ideas and rituals at a landscape level of spatial resolution comes from my own examination of rock art sites in the central Arkansas River Valley, where images at sites distributed north of the Arkansas River uniformly reflect spirit world subject matter, whereas images at sites located south of the river portray This World subject matter, thereby reflecting the operation of cosmological principles remarkably similar to those described earlier that structure the layout of Osage villages and camps (Sabo 2008). In sum, the cosmological features of the Terán map may well provide a single example of a much larger phenomenon involving the use of religious principles to structure community settlement patterns at varying scales of spatial resolution. As argued here, this phenomenon may apply not only to Caddo settlement patterns but to those of many other American Indian communities, both before and after the Columbian episode. This conclusion provides a wealth of new possibilities to investigate.

The Terán Map and Caddo Cosmology

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

447

Copyright © 2012. Nebraska Paperback. All rights reserved. The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

References Cited Albert, Lois E. 2000 The Norman Site: Descriptions. Caddoan Archeology 11(1–2):23–59. Albertson, Paul E., Maureen K. Corcoran, Whitney Autin, John Kruger, and Theresa Foster 1996 Geomorphic Investigation of the Great Bend Region, Red River. Technical Report GL-96. U.S. Army Corps of Engineers, Waterways Experiment Station, Vicksburg. Albertson, Paul E., and Joseph B. Dunbar 1993 Geomorphic Investigation of Shreveport to Daingerfield Navigation Project. Technical Report GL-93–31. U.S. Army Corps of Engineers, Waterways Experiment Station, Vicksburg. Ambrose, Stanley H. 1993 Isotopic Analysis of Paleodiets: Methodological and Interpretive Considerations. In Investigations of Ancient Human Tissue Chemical Analyses in Anthropology, edited by Mary K. Stanford, pp. 59–130. Gordon and Breach, Langhorne pa. Ambrose, Stanley H., Jane Buikstra, and Harold W. Krueger 2003 Status and Gender Differences in Diet at Mound 72, Cahokia, Revealed by Isotopic Analysis of Bone. Journal of Anthropological Archaeology 22:217–226. Ambrose, Stanley H., and Lynette Norr 1993 Experimental Evidence for the Relationship of Carbon Isotope Ratios of Whole Diet and Dietary Protein to Those of Bone Collagen and Carbonate. In Prehistoric Human Bone: Archaeology at the Molecular Level, edited by Joseph B. Lambert and Gisela Grupe, pp. 1–37. Springer ny. Anderson, David G. 1994 The Savannah River Chiefdoms: Political Change in the Late Prehistoric Southeast. University of Alabama Press, Tuscaloosa. 2001 Climate and Culture Change in Prehistoric and Early Historic Eastern North America. Archaeology of Eastern North America 29:143–186. Anderson, Gary C. 2005 The Conquest of Texas: Ethnic Cleansing in the Promised Land, 1820–1875. University of Oklahoma Press, Norman. Arnold, J. Barto, III 1975 Porosity and Refiring Tests on Ceramics from the George C. Davis Site, Texas. Bulletin of the Texas Archeological Society 46:231–241. Arnold, Jeanne E. 1996 Emergent Complexity: The Evolution of Intermediate Societies. Archaeological Series No. 9. International Monographs in Prehistory, Ann Arbor.

449

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Ashmore, Wendy, and Pamela L. Geller 2005 Social Dimensions of Mortuary Space. In Interacting with the Dead: Perspectives on Mortuary Archaeology for the New Millennium, edited by Gordon F. M. Rakita, Jane E. Buikstra, Lane A. Beck, and Sloan R. Williams, pp. 81–92. University Press of Florida, Gainesville. Autin, Whitney J. 1997 Geological Correlation of Surficial Deposits at Willow Chute, Lower Red River Valley, Bossier Parish, Louisiana. In Regional Archaeology Program, Management Unit 1: Eighth Annual Report, edited by Jeffrey S. Girard, pp. 90–100. Report on file at the Louisiana Division of Archaeology, Department of Culture, Recreation, and Tourism, Baton Rouge. Baerreis, David A. 1957 The Southern Cult and the Spiro Ceremonial Complex. Bulletin of the Oklahoma Anthropological Society 5:23–38. Bagur, Jacques D. 2001 A History of Navigation on Cypress Bayou and the Lakes. University of North Texas Press, Denton. Bailey, Garrick A. 1995 The Osage and the Invisible World: From the Works of Francis La Flesche. University of Oklahoma Press, Norman. Banks, Larry 1990 From Mountain Peaks to Alligator Stomachs: A Review of Lithic Resources in the TransMississippi South, the Southern Plains, and Adjacent Southwest. Memoir No. 4. Oklahoma Anthropological Society, Norman. Bareis, Charles John 1955 The Bracket Site, Ck-43, of Cherokee County, Oklahoma. Bulletin of the Oklahoma Anthropological Society 3:1–39. Barker, Alex W. 1992 Powhatan’s Pursestrings: On the Meaning of Surplus in a Seventeenth-Century Algonkian Chiefdom. In Lords of the Southeast: Social Inequality and the Native Elites of Southeastern North America, edited by Alex W. Barker and Timothy R. Pauketat, pp. 61–80. Archaeological Papers No. 3. American Anthropological Association, Washington dc. Barker, Alex W., Craig E. Skinner, M. Steven Shackley, Michael D. Glascock, and J. Daniel Rogers 2002 Mesoamerican Origin for an Obsidian Scraper from the Precolumbian Southeastern United States. American Antiquity 67(1):103–108. Barnes, Mark R., and Timothy K. Perttula 1999 Caddoan Ceremonial Sites of the Caddoan Cultural Area of Arkansas, Louisiana, Oklahoma, and Texas: Draft Caddo National Historic Landmark Nomination. Caddoan Archeology 10(1):5–29. Barr, Juliana 2007 Peace Came in the Form of a Woman: Indians and Spaniards in the Texas Borderlands. University of North Carolina Press, Chapel Hill. Barr, Thomas 1965 Three Archaeological Sites in the Pine Creek Reservoir, McCurtain County, Oklahoma. Archaeological Site Report No. 11. Oklahoma River Basin Survey, Norman.

450

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Baugh, Timothy G. 2007 Warfare among the Kirikir’is: Archeology, Ethnography, and Ethnohistory. Kansas Anthropologist 28:1–22. 2009 Protohistoric Kirikir’i-s Societies on the Central and Southern Plains. Oklahoma Anthropological Society Bulletin 57:15–59. Baugh, Timothy G., and Stephen M. Perkins (editors) 2008 Land of Our Ancestors: Studies in Protohistoric and Historic Wichita Cultures. Memoir 40. Plains Anthropologist 53(208):375–593. Bell, Robert E. 1947 Trade Materials at Spiro Mound as Indicated by Artifacts. American Antiquity 12:181–184. 1972 The Harlan Site, CK-6, a Prehistoric Mound Center in Cherokee County, Eastern Oklahoma. Memoirs No. 2. Oklahoma Anthropological Society, Norman. Bell, Robert E., and David A. Baerreis 1951 A Survey of Oklahoma Archaeology. Bulletin of the Texas Archeological and Paleontological Society 22:7–100. Bènard de La Harpe, Jean-Baptiste 1718 Journal du Voyage de la Louissianne fait par le Sr Benard de la Harpe et des DeCouvertes Quil a fait dans la Party de l’ouest de Cette Colonie. Bibliotheque Nationale, MSS. Francais, 8989:fols. 1–36. Bender, Barbara 1993 Landscape, Politics, and Perspective. Berg, Oxford. Binford, Lewis R. 1978 Nunamiut Ethnoarchaeology. Academic Press, New York ny. Blitz, John H. 2010 New Perspectives in Mississippian Archaeology. Journal of Archaeological Research 18(1):1–39. Bogan, Arthur E. 1980 A Comparison of Late Prehistoric Dallas and Overkill Cherokee Subsistence Strategies in the Little Tennessee River Valley. Ph.D. dissertation, Department of Anthropology, University of Tennessee, Knoxville. Bolton, Herbert E. 1915 Texas in the Middle Eighteenth Century: Studies in Spanish Colonial History and Administration. University of California Press, Berkeley. 1987 The Hasinais: Southern Caddoans as Seen by the Earliest Europeans. Edited by Russell M. Magnaghi. University of Oklahoma Press, Norman. Bolton, Herbert E. (editor) 1914 Athanase de Mézières and the Louisiana-Texas Frontier, 1768–1780. 2 vols. Clark, Cleveland oh. Bond, Clell 1970 Spinach Patch, 3FR1, and the River Bank Site, 3FR23, West-Central Arkansas. Report submitted to the National Park Service, Southeast Region, and the University of Arkansas Museum, Contract No. 14–10–7:911–14. On file, Arkansas Archeological Survey, Fayetteville. Bowman, Isaiah 1911 Forest Physiography: Physiography of the United States and Principles of Soils in Relation to Forestry. Wiley, London.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

451

Copyright © 2012. Nebraska Paperback. All rights reserved.

Bradley, Raymond S., K. R. Briffa, J. Cole, M. K. Hughes, and T. J. Osborn 2003 The Climate of the Last Millennium. In Paleoclimate, Global Change, and the Future, edited by Keith D. Alverson, Raymond S. Bradley, and Thomas F. Pedersen, pp. 105–141. Springer, Berlin. Bradley, Richard 2000 An Archaeology of Natural Places. Routledge, London. Brain, Jeffrey P., and Philip Phillips 1996 Shell Gorgets: Styles of the Late Prehistoric and Protohistoric Southeast. Peabody Museum Press, Cambridge ma. Brandt, Roel, Bert J. Groenewoudt, and Kenneth L. Kvamme 1992 An Experiment in Archaeological Site Location: Modeling in the Netherlands Using gis Techniques. World Archaeology 24(2):268–282. Brock, A., and M. Reynolds 2007 New Radiocarbon Dates from 3CL593. Field Notes (Newsletter of the Arkansas Archeological Society) 339:10–11. Bronk Ramsey, Christopher 2001 Development of the Radiocarbon Program OxCal. Radiocarbon 43(2A):355–363. 2009 Bayesian Analysis of Radiocarbon Dates. Radiocarbon 51:337–360. Brooks, Robert L. 2006 From Mounds to Monasteries: A Look at Spiro and Other Centers through the Use of Metaphor. Caddo Archeology Journal 15:41–58. Brown, Cecil H. 2006 Glottochronology and the Chronology of Maize in the Americas. In Histories of Maize: Multidisciplinary Approaches to the Prehistory, Linguistics, Biogeography, Domestication, and Evolution of Maize, edited by John Staller, Robert Tykot, and Bruce Benz, pp. 647–664. Academic Press, Burlington ma. Brown, Greg 1992 Dream Cafe, track 6. Red House Records, St. Paul mn. Brown, James A. 1966 Spiro Studies. Vol. 1. Description of the Mound Group. First part of the Second Annual Report of Caddoan Archaeology, Spiro Focus Research. University of Oklahoma Research Institute, Norman. 1971 The Dimensions of Status in the Burials at Spiro. In “Approaches to the Social Dimensions of Mortuary Practices,” edited by James A. Brown. Memoirs of the Society for American Archaeology 25:92–122. 1975 Spiro Art and Its Mortuary Contexts. In Death and the Afterlife in Pre-Columbian America, edited by Elizabeth P. Benson, pp. 1–32. Dumbarton Oaks Research Library and Collections, Washington dc. 1981 The Search for Rank in Prehistoric Burials. In The Archaeology of Death, edited by Robert Chapman, Ian Kinnes, and Klavs Randsborg, pp. 25–37. Cambridge University Press, Cambridge. 1983 Spiroan Exchange Connections Revealed by Sources of Imported Raw Materials. In Southeastern Natives and Their Pasts: A Collection of Papers Honoring Dr. Robert E. Bell, edited by Don G. Wyckoff and Jack L. Hofman, pp. 129–162. Studies in Oklahoma’s Past No. 11. Oklahoma Archeological Survey, Norman. 1984a Arkansas Valley Caddoan: The Spiro Phase. In Prehistory of Oklahoma, edited by Robert E. Bell, pp. 241–263. Academic Press, Orlando.

452

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

1984b Prehistoric Southern Ozark Marginality: A Myth Exposed. Special Publication No. 6. Missouri Archaeological Society, Columbia. 1990 Archaeology Confronts History at the Natchez Temple. Southeastern Archaeology 9:1–10. 1991 The Falcon and the Serpent: Life in the Southeastern United States at the Time of Columbus. In Circa 1492: Art in the Age of Exploration, edited by Jay A. Levenson, pp. 529–534, 582–589. National Gallery of Art, Washington dc and Yale University Press, New Haven ct. 1996 The Spiro Ceremonial Center: The Archaeology of Arkansas Valley Caddoan Culture in Eastern Oklahoma. 2 vols. Memoirs No. 29. Museum of Anthropology, University of Michigan, Ann Arbor. 2003a The Cahokia Mound 72 Sub1 Burials as Collective Representation. In “A Deep-Time Perspective: Studies in Symbols, Meaning, and the Archaeological Record,” edited by John D. Richards and Melvin L. Fowler, pp. 81–97. Wisconsin Archeologist 84 (nos. 1 and 2). 2003b Collective Burial Practices across the Agricultural Transition in the Eastern Woodlands. In Stones and Bones: Formal Disposal of the Dead in Atlantic Europe during the Mesolithic-Neolithic Interface, 6000–3000 bc. Archaeological Conference in Honour of the Late Professor Michael J. O’Kelly. Proceedings of the Stones and Bones Conference in Sligo, Ireland, May 1–5, 2002, edited by Goran Burenhult and Susanne Westergaard, pp. 203–219. bar International Series 1201. Oxford, England. 2004 Exchange and Interaction to ad 1500. In Handbook for North American Indians, vol. 14, Southeast, edited by Raymond D. Fogelson, pp. 677–685. Smithsonian Institution, Washington dc. 2006 Where’s the Power in Mound Building? An Eastern Woodland Perspective. In Leadership and Polity in Mississippian Society, edited by Brian M. Butler and Paul D. Welch, pp. 197–213. Occasional Paper No. 33. Center for Archaeological Investigations, Southern Illinois University, Carbondale. 2007 On the Identity of the Birdman within Mississippian Period Art and Iconography. In Ancient Objects and Sacred Realms: Interpretations of Mississippian Iconography, edited by F. Kent Reilly III and James F. Garber, pp. 56–106. University of Texas Press, Austin. 2010 Cosmological Layouts of Secondary Burials as Political Instruments. In Mississippian Mortuary Practices: Beyond Hierarchy and the Representationist Perspective, edited by Lynne P. Sullivan and Robert C. Mainfort Jr., pp. 30–53. University Press of Florida, Gainesville. Brown, James A., Robert E. Bell, and Don G. Wyckoff 1978 Caddoan Settlement Patterns in the Arkansas River Drainage. In Mississippian Settlement Patterns, edited by Bruce D. Smith, pp. 169–200. Academic Press, New York. Brown, James A., and David H. Dye 2007 Severed Heads and Sacred Scalplocks; Mississippian Iconographic Trophies. In The Taking and Display of Human Body Parts as Trophies by Amerindians, edited by Richard J. Chacon and David H. Dye, pp. 278–298. Springer, New York. Brown, James A., Richard A. Kerber, and Howard D. Winters 1990 Trade and the Evolution of Exchange Relations at the Beginning of the

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

453

Copyright © 2012. Nebraska Paperback. All rights reserved.

Mississippian Period before ad 1200. In Mississippian Emergence: The Evolution of Ranked Agricultural Societies in Eastern North America, edited by Bruce D. Smith, pp. 251–280. Smithsonian Institution Press, Washington dc. Brown, James A., and J. Daniel Rogers 1999 ams Dates on Artifacts of the Southeastern Ceremonial Complex at Spiro. Southeastern Archaeology 18:134–141. Bruseth, James E. 1987 Late Holocene Environmental Change and Human Adaptive Strategies in Northeast Texas. Ph.D. dissertation, Department of Anthropology, Southern Methodist University, Dallas tx. Bruseth, James E., and Timothy K. Perttula 1981 Prehistoric Settlement Patterns at Lake Fork Reservoir. Texas Antiquities Permit Series, Report No. 2. Texas Antiquities Committee, Austin, and Southern Methodist University, Dallas tx. Bruseth, James E., and Bill Pierson 2004 Magnetometer Survey at the George C. Davis Site (41CE19). Current Archeology in Texas 6(1):7–9. Bruseth, James E., William T. Pierson, and Robert M. Johnson 2007 Large-Area Remote Sensing Coverage of Archaeological Sites with a Dual-Track, Cart-Mounted Cesium Magnetometer. Journal of Field Archaeology 32:1–15. Bruseth, James E., Diane E. Wilson, and Timothy K. Perttula 1995 The Sanders Site: A Spiroan Entrepot in Texas? Plains Anthropologist 40(153):223–236. Buikstra, Jane E., and George Milner 1991 Isotopic and Archaeological Interpretations of Diet in the Central Mississippi Valley. Journal of Archaeological Science 18:319–329. Burnett, Barbara A. 1990 The Bioarchaeological Synthesis of the Eastern Portion of the Gulf Coastal Plain. In The Archeology and Bioarchaeology of the Gulf Coastal Plan: Volume 2, by Dee Ann Story, Janice A. Guy, Barbara A. Burnett, Martha D. Freeman, Jerome C. Rose, D. Gentry Steele, Ben W. Olive, and Karl J. Reinhard, pp. 385–418. 2 Vols. Research Series No. 38. Arkansas Archeological Survey, Fayetteville. 1999 The Bioarchaeological Synthesis of the Eastern Portion of the Gulf Coastal Plain. In Bioarcheology of the South Central United States, edited by Jerome C. Rose, pp. 102–132. Research Series No. 55. Arkansas Archeological Survey, Fayetteville. 2010 Intertribal Warfare between the Trans-Mississippi South and the Southern Plains. Paper presented at the 52nd Caddo Conference and 17th East Texas Archeological Conference, Tyler, Texas. Burrough, Peter A. 1986 Principles of Geographic Information Systems for Land Resource Management. Oxford University Press, New York ny. Burson, Elizabeth, and Maynard B. Cliff 2000 Cultural Resources Survey of Six Proposed Timber-Cutting Tracts at Wright Patman Lake, Bowie County, and Lake O’ the Pines, Marion and Upshur Counties, Texas. Miscellaneous Reports of Investigations No. 201. Geo-Marine, Plano tx.

454

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Burton, Susan S. 1970 The Hugo Dam Site, Ch-112, Choctaw County, Southeast Oklahoma. Archaeological Site Report No. 16. Oklahoma River Basin Survey, Norman. Butler, Barbara H. 1969 Analysis of the Human Skeletal Remains. In Investigations at the Sam Kaufman site, Red River County, Texas, edited by S. Alan Skinner, R. King Harris, and Keith M. Anderson, pp. 115–136. Contributions in Anthropology No. 5. Southern Methodist University, Dallas. Butler, Brian M., and Paul D. Welch (editors) 2006 Leadership and Polity in Mississippian Society. Occasional Paper No. 33. Center for Archaeological Investigations, Southern Illinois University, Carbondale. Byrd, Kathleen M. 1980 Zooarchaeological Analysis of the Hanna Site: An Alto Focus Occupation in Louisiana. Louisiana Archaeology 5:235–265. Caldwell, Warren W. 1958 Smithsonian Institution River Basin Surveys: Field Forms and Notes. On file, Arkansas Archeological Survey, Fayetteville. Cannon, Aubrey 2002 Spatial Narratives of Death, Memory, and Transcendence. In The Space and Place of Death, edited by Helaine Silverman and David B. Small, pp. 191–199. Archeological Papers No. 11. American Anthropological Association, Washington dc. Carr, Christopher 1995 Mortuary Practices: Their Social, Philosophical-Religious, Circumstantial, and Physical Determinants. Journal of Archaeological Method and Theory 2(2):105–200. Carr, Soraya 1984 Subsistence and Ceremony: Faunal Utilization in a Late Preclassic Community at Cerros, Belize. In Prehistoric Lowland Maya Environment and Subsistence Economy, edited by Mary Pohl, pp. 115–132. Papers of the Peabody Museum, vol. 77. Peabody Museum of Archaeology and Ethnology, Harvard University, Cambridge. 1996 Precolumbian Mayan Exploitation and Management of Deer Populations. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use, edited by S. L. Fedick, pp. 251–261. University of Utah Press, Salt Lake City. Carter, Cecile E. 1995 Caddo Indians: Where We Come From. University of Oklahoma Press, Norman. Casselberry, Samuel E. 1974 Further Refinement of Formulae for Determining Population from Floor Space. World Archaeology 6(1):117–124. Cast, Robert, Timothy K. Perttula, Bobby Gonzalez, and Bo Nelson 2006 Documentation of Caddo Ceramic Vessels from 41WD60, Wood County, Texas. Historic Preservation Program, Caddo Nation of Oklahoma, Binger ok. Castañeda, Carlos E. 1936– 1958 Our Catholic Heritage in Texas. 6 vols. Von Boeckmann-Jones, Austin.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

455

Copyright © 2012. Nebraska Paperback. All rights reserved.

Charles, Douglas K. 2005 The Archaeology of Death as Anthropology. In Interacting with the Dead: Perspectives on Mortuary Archaeology for the New Millennium, edited by Gordon F. M. Rakita, Jane E. Buikstra, Lane A. Beck, and Sloan R. Williams, pp. 15–24. University Press of Florida, Gainesville. Chesson, Meredith S. 2001 Embodied Memories of Place and People: Death and Society in an Early Urban Community. In Social Memory, Identity, and Death: Anthropological Perspectives on Mortuary Rituals, edited by Meredith S. Chesson, pp. 100–113. Archeological Papers No. 10. American Anthropological Association, Washington dc. Chipman, Donald E. 1992 Spanish Texas, 1519–1821. University of Texas Press, Austin. Clements, Forrest M. 1945 Historical Sketches of the Spiro Mound. Contributions from the Museum of American Indian 14:48–68. Heye Foundation, New York ny. Cobb, Charles R. 2003 Mississippian Chiefdoms: How Complex? Annual Review of Anthropology 32:63–84. Cobb, Charles R., and Bretton Giles 2009 War Is Shell: The Ideology and Embodiment of Mississippian Conflict. In Warfare in Cultural Context: Practice, Agency, and the Archaeology of Violence, edited by Axel E. Nielsen and William H. Walker, pp. 84–108. University of Arizona Press, Tucson. Cohen, Mark N., and George J. Armelagos (editors) 1984 Paleopathology at the Origins of Agriculture. Academic Press, New York ny. Colahan, Clark 1994 The Visions of Sor Maria de Agreda: Writing Knowledge and Power. University of Arizona Press, Tucson. Colby, Gail R. 1997 Analysis of Dental Abscess Formation in a Texas Hunter-Gatherer and a Texas Agricultural Population. Bulletin of the Texas Archeological Society 68:179–190. Conner, Michael D. 1998 National Register of Historic Places Registration Form for 23MD46, Pineville Site. On file, Center for Archaeological Research, Southwest Missouri State University, Springfield. Cook, Sherburne F. 1972 Prehistoric Demography. In Current Topics in Anthropology, vol. 3, module 16. Addison-Wesley Modular Publications, Reading ma. Cox, Isaac J. (editor) 1905 The Journeys of Rene Robert Cavelier, Sieur de La Salle. Trail Makers, New York ny. Cranford, David 2007 The Norman (34WG2) and Harlan (34CK6) Sites: Political Dynamics of Closely Spaced Mississippian Polities in Eastern Oklahoma. Master’s thesis, Department of Anthropology, University of Oklahoma, Norman. Creel, Darrell G. 1996 Hatchel-Mitchell Site. In The New Handbook of Texas, vol. 3, edited by Ron Tyler, pp. 504–505. Texas State Historical Association, Austin.

456

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Creel, Darrell G., Dale Hudler, Samuel M. Wilson, T. Clay Schultz, and Chester P. Walker 2005 A Magnetometer Survey of Caddoan Mounds State Historic Site. Technical Report 51. Texas Archeological Research Laboratory, University of Texas at Austin. 2008 Geophysical Survey of the Mound B Area at the George C. Davis Site. Bulletin of the Texas Archaeological Society 79:177–190. Davis, E. Mott 1958 The Whelan Site, a Late Caddoan Component in the Ferrell’s Bridge Reservoir, Northeastern Texas. Report submitted to the National Park Service by University of Texas, Division of Research in Anthropology. 1970 Archeological and Historical Assessment of the Red River in Texas. In Archeological and Historical Resources of the Red River Basin, edited by Hester A. Davis, pp. 27–69. Research Series No. 1. Arkansas Archeological Survey, Fayetteville. Davis, E. Mott, and Jules R. Gipson 1960 The Dalton Site: A Late Caddoan Mound Site in the Ferrell’s Bridge Reservoir Area, Northeastern Texas. Report submitted to National Park Service, Contract No. 14–10–333–242, by Division of Research in Anthropology, University of Texas at Austin. Davis, Hester A. 1962 The Destruction of Mound C at the Crenshaw Site. Plains Anthropologist 7(16):88. 1967 The Puzzle of Point Remove. Field Notes: Monthly Newsletter of the Arkansas Archeological Society 33:2–7. Davis, Hester A., and E. Mott Davis 2009 An Account of the Birth and Growth of Caddo Archeology, as Seen by Review of 50 Caddo Conferences, 1946–2008. Caddo Archeology Journal 19:3–72. Delcourt, Paul A., and Hazel R. Delcourt 1981 Vegetation Maps for Eastern North America: 40,000 Yr. B.P. to the Present. In Geobotany II, edited by R. C. Romans, pp. 123–165. Plenum Press, New York ny. DeNiro, Michael J., and Margaret J. Schoeninger 1983 Stable Carbon and Nitrogen Isotope Ratios of Bone Collagen: Variation within Individuals, between Sexes, and within Populations Raised on Monotonous Diets. Journal of Archaeological Science 10:199–203. Dering, J. Phil 2004 Archaeobotanical Evidence for Agriculture and Wild Plant Use at 41RK214. In The Oak Hill Village Site (41RK214), Rusk County, Texas, by Robert Rogers and Timothy K. Perttula, pp. 329–336. Document No. 030083. pbs&j, Austin. Derrick, Sharon M. 1997 A Study in Frustration: Analysis of Human Remains Removed from the Coker Mound Site (41CS1). Journal of Northeast Texas Archaeology 9:21–29. Derrick, Sharon M., Gail R. Colby, and D. Gentry Steele 2008 Analysis of Human Remains from the Arnold Roitsch Site (41RR16). In Collected Papers from Past Texas Archeological Society Field Schools, edited by Timothy K. Perttula, pp. 413–432. Special Publication No. 5. Texas Archeological Society, San Antonio. Derrick, Sharon M., and Diane Wilson 2001 The Effects of Epidemic Disease on Caddo Demographic Structure. Bulletin of the Texas Archeological Society 72:91–103.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

457

Copyright © 2012. Nebraska Paperback. All rights reserved.

Dickinson, Samuel D. 1936 Ceramic Relationships of the Pre-Caddo Pottery from the Crenshaw Site. Bulletin of the Texas Archeological and Paleontological Society 8:56–70. Dietler, Michael, and Brian Hayden (editors) 2001 Feasts: Archaeological and Ethnographic Perspectives. Smithsonian Institution Press, Washington dc. Diggs, George M., Jr., Barney L. Lipscomb, Monique D. Reed, and Robert J. O’Kennon 2006 Illustrated Flora of East Texas, Volume One: Introduction, Pteridophytes, Gymnosperms, and Monocotyledons. Botanical Research Institute of Texas, Fort Worth. Dorsey, George A. 1905 Traditions of the Caddo. Publication 41. Carnegie Institution of Washington, Washington dc. Douglas, Mary 1971 Deciphering a Meal. In Myth, Symbol, and Culture, edited by Clifford Geertz, pp. 71–82. W. W. Norton, New York. 1982 In the Active Voice. Routledge and Kegan Paul, London. Duffield, Lathel, F. 1969 Vertebrate Faunal Remains from the School Land I and School Land II Sites. Bulletin of the Oklahoma Anthropological Society 18:47–65. 1973 The Oklahoma Craig Mound: Another Look at an Old Problem. Bulletin of the Oklahoma Anthropological Society 22:1–10. Dunn, William E. 1917 Spanish and French Rivalry in the Gulf Region of the United States, 1678–1702: The Beginnings of Texas and Pensacola. Studies in History No. 1. University of Texas, Austin. Dunnell, Robert C. 1982 Science, Social Science, and Common Sense: The Agonizing Dilemma of Modern Archaeology. Journal of Anthropological Research 38:1–25. 2008a Archaeological Things: Languages of Observation. In Time’s River: Archaeological Syntheses from the Lower Mississippi River Valley, edited by Janet Rafferty and Evan Peacock, pp. 45–68. University of Alabama Press, Tuscaloosa. 2008b Archaeology in the Lower Mississippi Valley. In Time’s River: Archaeological Syntheses from the Lower Mississippi River Valley, edited by Janet Rafferty and Evan Peacock, pp. 16–44. University of Alabama Press, Tuscaloosa. Durham, James H. 1976 Additional Information on Burial Associations, Crenshaw Site, Mound C, Miller County, Arkansas. Bulletin of the Oklahoma Anthropological Society 24:155–160. Durham, James H., and Michael K. Davis 1975 Report on Burials Found at Crenshaw Mound “C,” Miller County, Arkansas. Bulletin of the Oklahoma Anthropological Society 23:1–90. Durham, James H., and Glen L. Kizzia 1964 A Deep Burial in Crenshaw Mound “C,” Miller County, Arkansas. Central States Archaeological Journal 11(2):44–67. Dye, David H. 2006 The Transformation of Mississippian Warfare: Four Case Studies from the Mid-South. In The Archaeology of Warfare: Prehistories of Raiding and Conquest, edited

458

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

by Elizabeth N. Arkush and Mark W. Allen, pp. 101–147. University Press of Florida, Gainesville. 2008 War Paths, Peace Paths: An Archaeology of Cooperation and Conflict in Native Eastern North America. AltaMira Press, Lanham md. Dye, David H., and Adam King 2007 Desecrating the Sacred Ancestor Temples: Chiefly Conflict and Violence in the American Southeast. In North American Indigenous Warfare and Ritual Violence, edited by Richard J. Chacon and Ruben G. Mendoza, pp. 160–181. University of Arizona Press, Tucson. Early, Ann M. 1988 Standridge: Caddoan Settlement in a Mountain Environment. Research Series No. 29. Arkansas Archeological Survey, Fayetteville. 1993 Ceramics. In Caddoan Saltmakers in the Ouachita Valley: The Hardman Site, edited by Ann M. Early, pp. 63–119. Research Series No. 43. Arkansas Archeological Survey, Fayetteville. 2000 The Caddos of the Trans-Mississippi South. In Indians of the Greater Southeast: Historical Archaeology and Ethnohistory, edited by Bonnie G. McEwan, pp. 122–141. University Press of Florida, Gainesville. 2002 Arkansas Prehistory and History in Review: The Social Hill Phase. Field Notes (Newsletter of the Arkansas Archeological Society) 306:10–13. 2004 Prehistory of the Western Interior After 500 bc. In Handbook of North American Indians, Volume 14, Southeast edited by Raymond D. Fogelson, pp. 560–573. Smithsonian Institution, Washington dc. Early, Ann M., editor 1993 Caddoan Saltmakers in the Ouachita Valley: The Hardman Site. Research Series No. 43. Arkansas Archeological Survey, Fayetteville. 2000 Forest Farmsteads: A Millennium of Human Occupation at Winding Stair in the Ouachita Mountains. Research Series 57. Arkansas Archeological Survey, Fayetteville. Elder, William H. 1965 Primeval Deer Hunting Pressure Revealed by Remains from American Indian Middens. Journal of Wildlife Management 29:336–370. Eliade, Mircea 1958 Patterns in Comparative Religion. Translated by Rosemary Sheed. Sheed and Ward, London. 1959 The Sacred and the Profane: The Nature of Religion. Translated by Willard R. Trask. Harcourt Brace Jovanovich, New York. Elson, Katherine M., Christopher Smith, and Timothy K. Perttula 2004 Additional Maize Studies. In The Oak Hill Village Site (41RK214), Rusk County, Texas, by Robert Rogers and Timothy K. Perttula, pp. 337–344. Document No. 030083. pbs&j, Austin. Emerson, Thomas E. 1989 Water, Serpents, and the Underworld: An Exploration into Cahokian Symbolism. In Southern Ceremonial Complex, Artifacts and Analysis: The Cottonlandia Conference, edited by Patricia E. Galloway, pp. 45–92. University of Nebraska Press, Lincoln.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

459

Copyright © 2012. Nebraska Paperback. All rights reserved.

Emerson, Thomas E. and Jeffrey S. Girard 2004 Dating Gahagan and Its Implications for Understanding Cahokia-Caddo Interactions. Southeastern Archaeology 23(1):57–64. Emerson, Thomas E., and Randall E. Hughes 2000 Figurines, Flint Clay Sourcing, the Ozark Highlands, and Cahokian Acquisition. American Antiquity 65(1):79–101. Emerson, Thomas E., Randall E. Hughes, Mary R. Hynes, and Sarah R. Wisseman 2002 Implications of Sourcing Cahokia-Style Flint Clay Figurines in the American Bottom and the Upper Mississippi River Valley. Midcontinental Journal of Archaeology 27(1):309–338. 2003 The Sourcing and Interpretation of Cahokia-Style Figurines in the TransMississippi South and Southeast. American Antiquity 68:287–313. Emery, Kitty F. 2003a Animals from the Mayan Underworld: Reconstructing Elite Maya Ritual at the Cueva de los Quetzales. In Behavior behind Bones: The Zooarchaeology of Religion, Ritual, Status, and Identity, edited by A. Ernynck, pp. 101–113. Oxbow Books, Oxford. 2003b The Noble Beast: Status and Differential Access to Animals in the Maya World. World Archaeology 34:498–515. Etchieson, Meeks n.d. Archeological Investigation of the U.S. Forest Service–Weyerhaeuser Land Exchange, McCurtain County, Oklahoma and Garland, Pike, and Yell Counties, Arkansas. Draft on file, Ouachita National Forest, Hot Springs ar. Ethridge, Robbie 2009 Introduction: Mapping the Mississippian Shatter Zone. In Mapping the Mississippian Shatter Zone: The Colonial Indian Slave Trade and Regional Instability in the American South, edited by Robbie Ethridge and Sheri M. Shuck-Hall, pp. 1–62. University of Nebraska Press, Lincoln. Fagan, Brian M. 1995 Ancient North America: The Archaeology of a Continent. 2nd ed. Thames and Hudson, New York. Farley, James A., W. Fredrick Limp, and Jami J. Lockhart 1990 The Archaeologist’s Workbench: Integrating gis, Remote Sensing, EDA, and Database Management. In Interpreting Space: gis and Archaeology, edited by Kathleen M. S. Allen, Stanton W. Green, and Ezra B. W. Zubrow, pp. 141–164. Taylor and Francis, London. Feathers, James K. 2008 Absolute Dating in the Mississippi Delta. In Time’s River: Archaeological Syntheses from the Lower Mississippi River Valley, edited by Janet Rafferty and Evan Peacock, pp. 168–181. University of Alabama Press, Tuscaloosa. 2009 Problems of Ceramic Chronology in the Southeast: Does Shell-Tempered Pottery Appear Earlier Than We Think? American Antiquity 74(1):113–142. Fernandez-Armesto, Felipe 2009 1492: The Year the World Began. HarperOne, New York. Fieldhouse, Paul 1986 Food and Nutrition: Customs and Class. Croom Helm, London.

460

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Fields, Ross C. 2008 The Pine Tree Mound Site and the Entrada of the Hernando de Soto Expedition of 1542. Journal of Northeast Texas Archaeology 28:1–8. Fields, Ross C., Marie Blake, and Karl W. Kibler 1997 Synthesis of the Prehistoric and Historic Archaeology of Cooper Lake, Delta and Hopkins Counties, Texas. Reports of Investigations No. 104. Prewitt and Associates, Inc., Austin. Findlow, Frank J., and Jonathan E. Ericson (editors) 1980 Catchment Analysis: Essays on Prehistoric Resource Space. Anthropology UCLA 10 (1&2). Finklestein, J. Joe 1940 The Norman Site Excavations near Wagoner, Oklahoma. Oklahoma Prehistorian 3(3):2–15. Reprinted in Caddoan Archeology 11(1–2):6–22. Fishbeck, Helmut J., J. Daniel Rogers, Stuart R. Ryan, and Fern E. Swenson 1989 Sourcing Ceramics in the Spiro Region: A Preliminary Study Using ProtonInduced X-Ray Emission (Pixie) Analysis. Midcontinental Journal of Archaeology 14(1):3–17. Flores, Dan L. 1984 Jefferson and Southwestern Exploration: The Freeman and Custis Accounts of the Red River Expedition of 1806. University of Oklahoma Press, Norman. 2002 Southern Counterpart to Lewis and Clark: The Freeman and Custis Expedition of 1806. University of Oklahoma Press, Norman. Flynn, Peggy 1976 A Study of Red-Filmed Pottery from the Clement Site (Mc-8), McCurtain County, Oklahoma. Bulletin of the Oklahoma Anthropological Society 25:127–134. Foreman, Grant 1939 Fort Davis. Chronicles of Oklahoma 17:147–151. Foster, William C. (editor) 1998 The La Salle Expedition to Texas: The Journal of Henri Joutel, 1684–1687. Translated by Johanna S. Warren. Texas State Historical Association, Austin. Fowke, Gerard 1928 Archaeological Investigations-II. In Forty-fourth Annual Report of the Bureau of American Ethnology, pp. 399–540. U.S. Government Printing Office, Washington dc. Fraley, Chris, and Adrian E. Raftery 1998 How Many Clusters? Which Clustering Method? Answers via Model-Based Cluster Analysis. Computer Journal 41(8):578–588. Franciscus, Robert G. 2008 Preliminary Report on the Skeletal Analysis of Burials from the Washington Square Mound Site (41NA49) in Nacogdoches County, Texas. In Documentation of Associated and Unassociated Funerary Objects in the Stephen F. Austin State University Collections, Nacogdoches, Texas, by Timothy K. Perttula, Mark Walters, Bo Nelson, Bobby Gonzalez, and Robert Cast, pp. 81–92. Historic Preservation Program, Caddo Nation of Oklahoma, Binger ok. Friedman, Jonathan, and Michael J. Rowlands 1978 Notes towards an Epigenetic Model of the Evolution of Civilization. In The

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

461

Copyright © 2012. Nebraska Paperback. All rights reserved.

Evolution of Political Systems, edited by Jonathan Friedman and Michael J. Rowlands, pp. 201–276. University of Pittsburgh Press, Pittsburgh. Frink, Douglas S. 1992 The Chemical Variability of Carbonized Organic Matter through Time. Archaeology of Eastern North America 20:67–79. Frink, Douglas S., and Timothy K. Perttula 2002 ocr Data from East Texas and Implications for Long-Term Weather Patterns. In Archeological Investigations at the Proposed Lake Naconiche, Nacogdoches County, Texas, edited by Timothy K. Perttula, pp. 323–335. 2 vols. Report of Investigations No. 42. Archeological and Environmental Consultants, Austin. Fritz, Gayle J. 1979 Mounds in the Ozarks of Northwest Arkansas. Paper presented at the Arkansas Archeological Society Annual Meetings, Arkadelphia. 1986 Mounds in Northwest Arkansas: A More Positive Approach. In Contributions to Ozark Prehistory, edited by George Sabo III, pp. 49–54. Research Series No. 27. Arkansas Archeological Survey, Fayetteville. Gabaccia, Donna 2002 What Do We Eat? In Food in the USA, edited by Carole M. Counihan, pp. 35–40. Routledge, New York. Galan, Victor 1998 Excavations at 41TT653, the Ear Spool Site. CRM News & Views 10(2):21–25. Archeology Division, Texas Historical Commission, Austin. Gardner, Paul S. 1997 Plant Remains. In Two Caddoan Farmsteads in the Red River Valley: The Archeology of the McLelland and Joe Clark Sites, edited by David B. Kelley, pp. 109–120. Research Series 51. Arkansas Archeological Survey, Fayetteville. Gettys, Marshall 1975 Preliminary Report on Archaeological Investigations at Lukfata Reservoir, Southeast Oklahoma. General Survey Report No. 14. Oklahoma River Basin Survey, Norman. Gibson, Jon 2005 Bossier Tribes, Caddo in North Louisiana’s Pineywoods. Caddoan Archeology Journal 14:93–118. Gifford, Diane P. 1977 Observations of Modern Human Settlements as an Aid to Archaeological Interpretation. Ph.D. dissertation, Department of Anthropology, University of California, Berkeley. 1980 Ethnoarchaeological Contributions to the Taphonomy of Human Sites. In Fossils in the Making, edited by Anna K. Behrensmeyer and Andrew P. Hill, pp. 93–106. University of Chicago Press, Chicago. Gifford, Diane, and Anna Behrensmeyer 1977 Observed Formation and Burial of a Recent Human Occupation Site in Kenya. Quaternary Research 8:245–266. Gillespie, Susan D. 2001 Personhood, Agency, and Mortuary Ritual: A Case Study from the Ancient Maya. Journal of Anthropological Archaeology 20:73–112.

462

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Gill-King, Harrell 1999 Human Skeletal Remains: Pathological and Dietary/Nutritional Aspects. In The Hurricane Hill Site (41HP106): The Archaeology of a Late Archaic/Early Ceramic and Early-Middle Caddoan Settlement in Northeast Texas, edited by Timothy K. Perttula, pp. 329–343. Special Publication No. 4. Friends of Northeast Texas Archaeology, Pittsburg and Austin. Girard, Jeffrey S. 1995 An Early Ceramic Period Pit Feature at the Swan Lake Site, Bossier Parish, Louisiana. Caddoan Archeology Newsletter 5(4):6–11. 2000 Regional Archaeology Program, Management Unit 1: Eleventh Annual Report. Report on file at the Louisiana Division of Archaeology, Department of Culture, Recreation, and Tourism, Baton Rouge. 2002 Phelps Lake and Jim Burt: Two Middle Woodland Period Mounds in Northwestern Louisiana. Caddoan Archeology 12(4):5–17. 2006 Regional Archaeology Program, Management Unit 1: Seventeenth Annual Report. Report on file at the Louisiana Division of Archaeology, Department of Culture, Recreation, and Tourism, Baton Rouge. 2007a Regional Archaeology Program, Management Unit 1: Eighteenth Annual Report. Report on file at the Louisiana Division of Archaeology, Department of Culture, Recreation, and Tourism, Baton Rouge. 2007b Byram Ferry (16BO17): A Middle to Late Caddo Period Mound Site in the Red River Floodplain, Northwest Louisiana. Caddo Archeology Journal 16:9–25. Goddard, Ives 1996 The Classification of the Native Languages of North America. In Languages, edited by Ives Goddard, pp. 290–324. Volume 17, Handbook of North American Indians, Smithsonian Institution, Washington dc. Goldstein, Lynne 1980 Mississippian Mortuary Practices. Scientific Papers No. 4. Northwestern University Archaeological Program, Kampsville il. 2000 Mississippian Ritual as Viewed Through the Practice of Secondary Disposal of the Dead. In Mound, Modoc, and Mesoamerica: Papers in Honor of Melvin L. Fowler, edited by Steven R. Ahler, pp. 193–205. Scientific Papers XXVIII. Illinois State Museum, Springfield il. 2006 Mortuary Analysis and Bioarchaeology. In Bioarchaeology: The Contextual Analysis of Human Remains, edited by Jane E. Buikstra and Lane A. Beck, pp. 375–387. Elsevier and Academic Press, Burlington ma. Gonzalez, Bobby, Robert L. Cast, Timothy K. Perttula, and Bo Nelson 2005 A Rediscovering of Caddo Heritage: The W. T. Scott Collection at the American Museum of Natural History and Other Caddo Collections from Arkansas and Louisiana. Historic Preservation Program, Caddo Nation of Oklahoma, Binger. Goody, Jack 1982 Cooking, Cuisine, and Class. Cambridge University Press, Cambridge. Goslar, Tomasz, W. O. van der Knaap, Sheila Hicks, Maja Andriþ Justyna Czernik, Ewa Goslar, Satu Räsänen, and Heidi Hyötylä 2005 Radiocarbon Dating of Modern Peat Profiles: Pre- and Post-Bomb 14C Variations in the Construction of Age-Depth Models. Radiocarbon 47(1):115–134.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

463

Copyright © 2012. Nebraska Paperback. All rights reserved.

Grealy, Michael, and Lawrence B. Conyers 2008 EM31 Geophysical Survey Methods, Results, and Recommendations. In Integrated Cultural Resources Investigations for the Bowie County Levee Realignment Project, Bowie County, Texas and Little River County, Arkansas, by Scott A. Sundermeyer, John T. Penman, and Timothy K. Perttula, pp. 91–107. Miscellaneous Reports, Report of Investigations No. 29. LopezGarcia Group, Dallas. Greenlee, Diana M. 2006 Dietary Variation and Prehistoric Maize Farming in the Middle Ohio Valley. In Histories of Maize, edited by John E. Staller, Robert H. Tykot, and Bruce F. Benz, pp. 215–233. Academic Press, Burlington ma. Gregory, David A., and Michael W. Diehl 2002 Duration, Continuity, and Intensity of Occupation at a Late Cienega Phase Settlement in the Santa Cruz River Floodplain. In Traditions, Transitions, and Technologies: Themes in Southwestern Archaeology, edited by Sarah H. Schlanger, pp. 200–223. University Press of Colorado, Boulder. Gregory, Hiram F. 1973 Eighteenth-Century Caddoan Archaeology: A Study in Models and Interpretation. Ph.D. dissertation, Department of Anthropology, Southern Methodist University, Dallas. 1983 Los Adaes: The Archaeology of an Ethnic Enclave. Geoscience and Man 23:53–57. 2009 The Caddo and the Caddo Conference. Caddo Archeology Journal 19:1–2. Gregory, Hiram F., George Avery, Aubra L. Lee, and Jay C. Blaine 2004 Presidio Los Adaes: Spanish, French, and Caddoan Interaction on the Northern Frontier. Historical Archaeology 38(3):65–77. Griffith, William Joyce 1954 The Hasinai Indians of East Texas as Seen by Europeans, 1687–1772. Philological and Documentary Studies Vol. 2, No. 3. Middle American Research Institute, Tulane University, New Orleans. Guccione, Margaret J., and Philip Hays 2008 Geomorphology, Sedimentology, and Vegetation History along the Red River Floodplain, Bowie County, Texas. In Integrated Cultural Resources Investigations for the Bowie County Levee Realignment Project, Bowie County, Texas and Little River County, Arkansas, by Scott A. Sundermeyer, John T. Penman, and Timothy K. Perttula, pp. 109–166. Miscellaneous Reports, Report of Investigations No. 29. LopezGarcia Group, Dallas. Gumerman, George C. (editor) 1971 The Distribution of Prehistoric Population Aggregates. Papers in Anthropology No. 1. Prescott College, Prescott az. Gumerman, George, IV 1997 Food and Complex Societies. Journal of Archaeological Method and Theory 4(2):105–140. Halfmoon, Stacey 2004 Caddo Art: A Personal Perspective. In Hero, Hawk, and Open Hand: American Indian Art of the Ancient Midwest and South, edited by Richard E. Townsend, pp. 207–219. The Art Institute of Chicago and Yale University Press, New Haven.

464

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Hall, Stephen A. 1990 Channel Trenching and Climatic Change in the Southern U.S. Great Plains. Geology 18:342–345. Hallam, E. and J. Hockey 2001 Death, Memory & Material Culture. Berg Publishers, Oxford. Hally, David J. 1993 The Territorial Size of Mississippian Chiefdoms. In Archaeology of Eastern North America: Papers in Honor of Stephen Williams, edited by James B. Stoltman, pp. 143–168. Archeological Report No. 25. Mississippi Department of Archives and History, Jackson. 2008 King: The Social Archaeology of a Late Mississippian Town in Northwestern Georgia. University of Alabama Press, Tuscaloosa. Hamilton, Henry W. 1952 The Spiro Mound. Missouri Archaeologist 14:17–106. Hammond, Robert and Patrick S. McCullagh 1982 Quantitative Techniques in Geography: An Introduction. Clarendon Press, Oxford. Hand, David, Heikki Mannila, and Padhraic Smyth 2001 Principles of Data Mining. mit Press, Cambridge ma. Harmon, Anna M., and Jerome C. Rose 1988 The Role of Dental Microwear Analysis in the Reconstruction of Prehistoric Diet. In Diet and Subsistence: Current Archaeological Perspectives, edited by Brenda V. Kennedy and Genevieve M. Le Moine, pp. 267–272. University of Calgary Archaeology Association, Calgary. 1989 Bioarcheology of the Louisiana and Arkansas Study Area. In Archeology and Bioarcheology of the Lower Mississippi Valley and Trans-Mississippi South in Arkansas and Louisiana, by Marvin D. Jeter, Jerome C. Rose, G. Ishmael Williams Jr., and Anna M. Harmon, pp. 323–354. Research Series No. 37. Arkansas Archeological Survey, Fayetteville. Harrington, M. R. 1920 Certain Caddo Sites in Arkansas. Indian Notes and Monographs, Miscellaneous Series no. 10. Museum of the American Indian, Heye Foundation, New York. Harris, R. K. 1953 The Sam Kaufman Site, Red River County, Texas. Bulletin of the Texas Archeological Society 24:43–68. Hatcher, Mattie Austin (editor) 1927 Descriptions of the Tejas or Asinai Indians, 1691–1722. Southwestern Historical Quarterly 31:150–180. 1932 The Expedition of Don Domingo Terán de Los Rios into Texas. Preliminary Studies of the Texas Catholic Historical Society 2:1–67. 1999 The Expedition of Don Domingo Teran de los Rios into Texas. In Wilderness Mission: Preliminary Studies of the Texas Catholic Historical Society II, edited by Jesus F. de la Teja, pp. 1–66. Studies in Southwestern Catholic History 2. Texas Catholic Historical Society, Austin. Hayden, Brian 2001 Fabulous Feasts: A Prolegomenon to the Importance of Feasting. In Feasts: Archaeological and Ethnographic Perspectives, edited by Michael Dietler and Brian Hayden, pp. 23–64. Smithsonian Institution Press, Washington dc.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

465

Copyright © 2012. Nebraska Paperback. All rights reserved.

2003 Were luxury foods the first domesticates? Ethnoarchaeological perspectives from Southeast Asia. World Archaeology 34:458–469. Heinrich, Paul V. 1993 Natural Settings. In Cultural Resources Survey at Red River Below Denison Dam, Levee Rehabilitation/Restoration Item 1, Miller County, Arkansas. R. Christopher Goodwin and Associates, New Orleans. Helms, Mary 1992 Political Lords and Political Ideology in Southeastern Chiefdoms: Comments and Observations. In Lords of the Southeast: Social Inequality and the Native Elites of Southeastern North America, edited by Alex W. Barker and Timothy R. Pauketat, pp. 185–194. Archaeology Papers No. 3. American Anthropological Association, Washington dc. Hickerson, Nancy P. 1990 The Visits of the “Lady in Blue:” An Episode in the History of the Southern Plains, 1629. Journal of Anthropological Research 46:67–90. Hilliard, Jerry E., and John Riggs 2000 amasda Site Encoding Manual, Version 3.0. Technical Paper No. 1. Arkansas Archeological Survey, Fayetteville. Hilliard, Jerry E., George Sabo III, and Deborah Sabo 2004 Rock Art in the Cultural Landscape. In Rock Art in Arkansas, edited by George Sabo III and Deborah Sabo, pp. 44–58. Popular Series No. 5. Arkansas Archeological Survey, Fayetteville. Hillson, Simon W. 1979 Diet and Dental Disease. World Archaeology 11:147–162. Hinkle, Katherine A. 1988 A Cultural Resources Survey of the Malvern to Benton Arkla Line A, Hot Spring and Saline Counties, Arkansas. Project No. 676. Arkansas Archeological Survey, Sponsored Research Program, Fayetteville. Hirsch, Eric, and Michael O’Hanlon (editors) 1993 The Anthropology of Landscape: Perspectives on Place and Space. Oxford University Press, Oxford. Hoagland, Bruce 2006 Soils. In Historical Atlas of Oklahoma, edited by Charles R. Goins and Danney Goble, pp. 16–17, University of Oklahoma Press, Norman. Hodder, Ian, and Clive Orton 1976 Spatial Archaeology. Cambridge University Press, Cambridge. Hodges, Thomas L., and Mrs. Thomas L. Hodges 1945 Suggestions for the Identification of Certain Mid-Ouachita Pottery as Cahinnio Caddo. Bulletin of the Texas Archeological and Paleontological Society 16:98–116. House, John H., with contributions by Gayle J. Fritz and Katherine Murray 1997 Time, People and Material Culture at the Kuykendall Brake Archaeological Site, Pulaski County, Arkansas. Paper presented at the 54th Southeastern Archaeological Conference, November 6, 1997, Baton Rouge, Louisiana. Web published at http://www.uark.edu/campus-resources/archinfo/kuykend.html. House, John H., and Mary V. Farmer 2001 Phase II Testing of Seven Archeological Sites on the Pine Bluff Arsenal, Jefferson County, Arkansas. Report submitted to the U.S. Department of the Army, Pine Bluff

466

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Arsenal, by the University of Arkansas at Pine Bluff Department of Social and Behavioral Sciences. Howard, Lynn E. 1940 Preliminary Report of Cherokee County, Oklahoma Archaeology. Oklahoma Prehistorian 3(1):2–9. 1948 Daily Log of Excavations at the Battle Mound, Lafayette County, Arkansas. Manuscript on file, Arkansas Archeological Survey, Magnolia. Hudson, Charles M. 1976 The Southeastern Indians. University of Tennessee Press, Knoxville. 1993 Reconstructing the de Soto Expedition Route West of the Mississippi River: Summary and Contents. In The Expedition of Hernando de Soto West of the Mississippi, 1541–1543, edited by Gloria A. Young and Michael P. Hoffman, pp. 143–154. University of Arkansas Press, Fayetteville. 1997 Knights of Spain, Warriors of the Sun. University of Georgia Press, Athens. Hunter, Donald G. 1997a Lithic Analysis. In Two Caddoan Farmsteads in the Red River Valley: The Archeology of the McLelland and Joe Clark Sites, edited by David B. Kelley, pp. 61–84. Research Series 51. Arkansas Archeological Survey, Fayetteville. 1997b Miscellaneous Artifacts. In Two Caddoan Farmsteads in the Red River Valley: The Archeology of the McLelland and Joe Clark Sites, edited by David B. Kelley, pp. 85–96. Research Series 51. Arkansas Archeological Survey, Fayetteville. Hunter, Donald G., David B. Kelley, Allen R. Saltus, and Lettie L. Markins 1992 Archaeological Investigations at Lock and Damn No. 5, Red River Waterway, Louisiana. Coastal Environments, Baton Rouge. Hunter, Donald G., Charles E. Pearson, Josetta LeBoeuf, and David B. Kelley 2002 Cultural Resources Evaluations of the T. C. Bearden (16NA494), Ramblin Wreck (16NA495), and Charles Webb (16RR86) Archaeological Sites, Located Along Red River in Natchitoches and Red River Parishes, Louisiana. Coastal Environments, Baton Rouge. Jackson, A. T. 1931 Excavation of a Burial Site, P. S. Cash Farm, Camp County, Texas. Manuscript on file, Texas Archeological Research Laboratory, University of Texas at Austin. 2004 Excavation of an Earth Mound, Bowie County, Texas. Caddoan Archeology Journal 13 (3–4):57–64. Jackson, A. T., Marcus S. Goldstein, and Alex D. Krieger 2000 The 1931 Excavations at the Sanders Site, Lamar County, Texas. Archival Series No. 2. Texas Archeological Research Laboratory, University of Texas at Austin. Jackson, H. Edwin 2005 “Darkening the Skies in Their Flight”: A Zooarchaeological Accounting of Passenger Pigeons in the Prehistoric Southeast. In Engaged Archaeology: Research Essays on North American Archaeology, Ethnobotany, and Museology, edited by Michelle Hegmon and B. Sunday Eiselt, pp. 174–199. Anthropological Papers 94. University of Michigan, Museum of Anthropology, Ann Arbor. Jackson, H. Edwin, and Susan L. Scott 1995a The Faunal Record of the Southeastern Elite: The Implications of Economy, Social Relations, and Ideology. Southeastern Archaeology 14:103–119. 1995b Mississippian Homestead and Village Subsistence Organization: Contrasts in Large-Mammal Remains from Two Sites in the Tombigbee Valley. In

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

467

Copyright © 2012. Nebraska Paperback. All rights reserved.

Mississippian Communities and Households, edited by J. Daniel Rogers and Bruce D. Smith, pp. 181–200. University of Alabama Press, Tuscaloosa. 2003 Patterns of Elite Faunal Utilization at Moundville. American Antiquity 68(4):552–572. Jackson, Jason B. 2003 Yuchi Ceremonial Life: Performance, Meaning, and Tradition in a Contemporary American Indian Community. University of Nebraska Press, Lincoln. Jelks, Edward B., and Curtis D. Tunnell 1959 The Harroun Site, A Fulton Aspect Component of the Caddoan Area, Upshur County, Texas. Archaeology Series No. 2. Department of Anthropology, University of Texas at Austin. Jeter, Marvin D. 1990 Edward Palmer’s Arkansaw Mounds. University of Arkansas Press, Fayetteville. Jeter, Marvin D., and Ann M. Early 1999 Prehistory of the Saline River Drainage Basin, Central to Southeast Arkansas. In Arkansas Archaeology: Essays in Honor of Dan and Phyllis Morse, edited by Robert C. Mainfort Jr. and Marvin D. Jeter, pp. 31–63. University of Arkansas Press, Fayetteville. Jeter, Marvin D., and G. Ishmael Williams Jr. 1989 Late Prehistoric Cultures, A. D. 1000–1500. In Archeology and Bioarcheology of the Lower Mississippi Valley and Trans-Mississippi South in Arkansas and Louisiana, by Marvin D. Jeter, Jerome C. Rose, G. Ishmael Williams Jr., and Anne M. Harmon, pp. 171–220. Research Series No. 37. Arkansas Archeological Survey, Fayetteville. Johnson, Kenneth 2006a Topography and Principal Landforms. In Historical Atlas of Oklahoma, edited by Charles R. Goins and Danney Goble, pp. 6–7, University of Oklahoma Press, Norman. 2006b Geologic Formations. In Historical Atlas of Oklahoma, edited by Charles R. Goins and Danney Goble, pp. 8–9, University of Oklahoma Press, Norman. Jones, Lindsay 2000a The Hermeneutics of Sacred Architecture: Experience, Interpretation, Comparison. Vol. 1: Monumental Occasions, Reflections on the Eventfulness of Religious Architecture. Harvard University Press, Cambridge. 2000b The Hermeneutics of Sacred Architecture: Experience, Interpretation, Comparison. Vol. 2: Hermeneutical Calisthenics, A Morphology of Ritual-Architectural Priorities. Harvard University Press, Cambridge. Joutel, Henri 1966 The Last Voyage Performed by Monsieur de la Sale, To the Gulph of Mexico, to Find Out the Mouth of the Mississippi River. Reprinted, Readex Microprint. Translated by Sieur de Mitchel. Originally published 1714, London. Kaufman, Leonard, and Peter J. Rousseeuw 1990 Finding Groups in Data: An Introduction to Cluster Analysis. Wiley-Interscience, New York. Kay, Marvin 1984 Late Caddo Subtractive Technology in the Red River Basin. In Cedar Grove: An Interdisciplinary Investigation of a Late Caddo Farmstead in the Red River Valley,

468

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

edited by Neal L. Trubowitz, pp. 174–206. Research Series No. 23. Arkansas Archeological Survey, Fayetteville. 1986 Caddoan Mound Construction Chronologies of the Western Ozark Highlands, Arkansas. In Contributions to Ozark Prehistory, edited by George Sabo III, pp. 77–79. Research Series No. 27. Arkansas Archeological Survey, Fayetteville. Kay, Marvin, and George Sabo III 2006 Mortuary Ritual and Winter Solstice Imagery of the Harlan-Style Charnel House. Southeastern Archaeology 25(1):29–47. Kay, Marvin, George Sabo III, and Ralph Merletti 1989 Late Prehistoric Settlement Patterning: A View from Three Caddoan CivicCeremonial Centers in Northwest Arkansas. In Contributions to Spiro Archaeology: Mound Excavations and Regional Perspectives, edited by J. Daniel Rogers, Don G. Wyckoff, and Dennis A. Peterson, pp. 129–157. Studies in Oklahoma’s Past No. 16. Oklahoma Archeological Survey, Norman. Kelley, David B., and Carey L. Coxe 1998 Cultural Resources Survey of Levee Rehabilitation/Restoration Areas Along the Red River Between Fulton, Arkansas and the Louisiana State Line: Items 4,5, and 9. Report prepared for the U.S. Army Corps of Engineers, Vicksburg District. Coastal Environments, Baton Rouge. Kelley, David B. (editor) 1997 Two Caddoan Farmsteads in the Red River Valley: The Archeology of the McLelland and Joe Clark Sites. Research Series No. 51. Arkansas Archeological Survey, Fayetteville. Kelly, Lucretia 2001 A Case of Ritual Feasting at the Cahokia Site. In Feasts: Archaeological and Ethnographic Perspectives, edited by Michael Dietler and Brian Hayden, pp. 334–367. Smithsonian Institution Press, Washington dc. Kent, Susan 1989 Cross-cultural Perceptions of Farmers as Hunters and the Value of Meat. In Farmers as Hunters: The Implications of Sedentism, edited by Susan Kent, pp. 1–17. Cambridge University Press, Cambridge. Kessell, John L. 2002 Spain in the Southwest: A Narrative History of Colonial New Mexico, Arizona, Texas, and California. University of Oklahoma Press, Norman. Kidder, T. R. 1986 The Jordan Site (16MO1 [22-I-1]). In Final Report on the Archaeological Test Excavations in the Central Boeuf Basin, Louisiana, 1985, by T. R. Kidder, with contributions by D. M. Ring and Roger T. Saucier, pp. 248–378. Lower Mississippi Survey Bulletin No. 10. Peabody Museum of Archaeology and Ethnology, Harvard University, Cambridge ma. 1992 Excavations at the Jordan Site (16MO1), Morehouse Parish, Louisiana. Southeastern Archaeology 11(2):109–131. 1993 The Glendora Phase: Protohistoric-Early Historic Culture Dynamics on the Lower Ouachita River. In Archaeology of Eastern North America: Papers in Honor of Stephen Williams, edited by James B. Stoltman, pp. 231–260. Archaeological Report No. 25. Mississippi Department of Archives and History, Jackson. 1998 Rethinking Caddoan–Lower Mississippi Valley Interaction. In The Native History of the Caddo: Their Place in Southeastern Archeology and Ethnohistory, edited by

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

469

Copyright © 2012. Nebraska Paperback. All rights reserved.

Timothy K. Perttula and James E. Bruseth, pp. 129–143. Studies in Archeology 30. Texas Archeological Research Laboratory, University of Texas at Austin. 2004 Prehistory of the Lower Mississippi Valley After 800 B.C. In Handbook of North American Indians, Volume 14, Southeast, edited by Raymond D. Fogelson, pp. 545–559. Smithsonian Institution, Washington dc. 2006 Contemplating Plaquemine Culture. In Plaquemine Archaeology, edited by Mark A. Rees and Patrick C. Livingood, pp. 196–205. University of Alabama Press, Tuscaloosa. Kidder, T. R., L. Roe, and T. M. Schilling 2010 Early Woodland Settlement and Mound Building in the Upper Tensas Basin, Northeast Louisiana. Southeastern Archaeology 29(1):121–145. King, Adam 2006 Leadership Strategies and the Nature of Mississippian Chiefdoms in Northern Georgia. In Leadership and Polity in Mississippian Society, edited by Brian M. Butler and Paul D. Welch, pp. 73–90. Occasional Paper No. 33. Center for Archaeological Investigations, Southern Illinois University, Carbondale. King, Adam (editor) 2007 Southeastern Ceremonial Complex: Chronology, Content, Context. University of Alabama Press, Tuscaloosa. King, Francis B. 1984 Presettlement Vegetation of the Cedar Grove Site. In Cedar Grove: An Interdisciplinary Investigation of a Late Caddo Farmstead in the Red River Valley, edited by Neal L. Trubowitz, pp. 207–210. Research Series No. 23. Arkansas Archeological Survey, Fayetteville. King, Mary Elizabeth, and Joan S. Gardner 1981 The Analysis of Textiles from Spiro Mound, Oklahoma. In The Research Potential of Anthropological Museum Collections, edited by Anne-Marie E. Cantwell, James B. Griffin, and Nan A. Rothschild, pp. 123–139. Annals of the New York Academy of Science 376. Albany. King, Nicholas 1806 Map of the Red River in Louisiana; From the Spanish Camp where the exploring party of the U.S. was met by the Spanish Troops to where it enters the Mississippi. Map on file, Library of Congress. Kintigh, Keith W., Donna M. Glowacki, and Deborah L. Huntley 2004 Long-term Settlement History and the Emergence of Towns in the Zuni Area. American Antiquity 69(3):432–456. Kitchens, Hershel and Dot Kitchens 1968 Field notes from Battle Mound (3LA1). On file, Arkansas Archeological Survey, Fayetteville. Knapp, A. Bernard, and Wendy Ashmore 1999 Archaeological Landscapes: Constructed, Conceptualized, Ideational. In Archaeology of Landscapes: Contemporary Perspectives, edited by Wendy Ashmore and A. Bernard Knapp, pp. 1–32, Blackwell, Oxford. Knight, Vernon J. 1986 The Institutional Organization of Mississippian Religion. American Antiquity 51:675–689.

470

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

2001 Feasting and the Emergence of Platform Mound Ceremonialism in Eastern North America. In Feasts: Archaeological and Ethnographic Perspectives on Food, Politics, and Power, edited by Michael Dietler and Brian Hayden, pp. 311–333. Smithsonian Institution Press, Washington dc. 2010 Mound Excavations at Moundville: Architecture, Elites, and Social Order. University of Alabama Press, Tuscaloosa. Kozuch, Laura 2002 Olivella Beads from Spiro and the Plains. American Antiquity 67:697–709. Kraft, Kenneth C. 1997 The Harkey Mound Site: Analysis: Analysis of the 1975 Test Excavations at Harkey #3 (34MC206). Archeological Resource Report No. 16. Oklahoma Conservation Commission, Oklahoma City. Krieger, Alex D. 1944 Archaeological Horizons in the Caddo Area. In El Norte de Mexico y el Sur de los Estados Unidos, pp. 154–156. Sociedad Mexicana de Antropologica, Mexico, D.F. 1945 An Inquiry into Supposed Mexican Influence on a Prehistoric Cult in the Southern United States. American Anthropologist 47:483–515. 1946 Culture Complexes and Classification in Northern Texas, with Extensions of Puebloan Datings to the Mississippi Valley. Publication No. 4640. University of Texas, Austin. 1947 The First Conference on the Caddoan Archaeological Area. American Antiquity 12:198–207. 1949 Viking Fund Grant for Battle Mound Excavations, Lafayette County, Arkansas: A Report of Activities and Expenditures as of January 31, 1949. Manuscript on file, Arkansas Archeological Survey, Fayetteville. Kuttruff, Jenna T. 1993 Mississippian Period Status Differentiation through Textile Analysis: A Caddoan Example. American Antiquity 58:125–145. Kvamme, Kenneth L. 1996 Randomization Methods for Statistical Inference in Raster gis Contexts. In The Colloquia of the XIII International Congress of Prehistoric and Protohistoric Sciences, edited by Amilcare Bietti, Alberto Cazzella, Ian Johnson, and Albertus Voorrips, 107–111, abaco, Forli, Italy. 1999 Recent Directions and Developments in Geographical Information Systems. Journal of Archaeological Research 7(2):153–201. 2003 Geophysical Surveys as Landscape Archaeology. American Antiquity 68:435–457. 2006 There and Back Again: Revisiting Archaeological Locational Modeling. In gis and Archaeological Site Location Modeling, edited by Mark W. Mehrer and Konnie Wescott, pp. 3–38. Taylor and Francis, London. 2008 Remote Sensing Approaches to Archaeological Reasoning: Pattern Recognition and Physical Principles. In Archaeological Concepts for the Study of the Cultural Past, edited by Alan P. Sullivan III, pp. 65–84. University of Utah Press, Salt Lake City. Lafferty, Robert H., III, Ann Early, Michael C. Sierzchula, Mary C. Hill, Gina S. Powell, Neal H. Lopinot, Linda S. Cummings, Susan L. Scott, Scott K. Nash, and Timothy K. Perttula 2000 Data Recovery at the Helm Site, 3HS449, Hot Spring County, Arkansas. mcra Report 2000–1. Mid-Continental Research Associates, Lowell ar.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

471

Copyright © 2012. Nebraska Paperback. All rights reserved.

Lankford, George E. 2004 World on a String: Some Cosmological Components of the Southeastern Ceremonial Complex. In Hero, Hawk, and Open Hand: American Indian Art of the Ancient Midwest and South, edited by Richard E. Townsend, pp. 207–219. The Art Institute of Chicago and Yale University Press, New Haven. 2007a The “Path of Souls”: Some Death Imagery in the Southeastern Ceremonial Complex. In Ancient Objects and Sacred Realms: Interpretations of Mississippian Iconography, edited by F. Kent Reilly III and James F. Garber, pp. 174–212. University of Texas Press, Austin. 2007b Reachable Stars: Patterns in the Ethnoastronomy of Eastern North America. University of Alabama Press, Tuscaloosa. 2008 Looking for Lost Lore: Studies in Folklore, Ethnology, and Iconography. University of Alabama Press, Tuscaloosa. Larsen, Clark S., R. Shavit, and M. C. Griffin 1991 Dental Caries Evidence for Dietary Change: an Archaeological Context. In Advances in Dental Anthropology, edited by Marc A. Kelley and Clark S. Larsen, pp. 179–202. Wiley-Liss, New York. La Vere, David 2004 The Texas Indians. Texas a&m University Press, College Station. Lawton, Sherman P. 1960 Archaeological Survey of the Hugo Reservoir. Manuscript on file, Sam Noble Oklahoma Museum of Natural History, Norman. Lee, Christine 1999 Origins and Interactions of the Caddo: A Study in Dental and Cranial Nonmetric Traits. Master’s thesis. Department of Anthropology, Arizona State University, Tempe. Lemley, Harry J. 1936 Discoveries Indicating a Pre-Caddo Culture on the Red River in Arkansas. Bulletin of the Texas Archeological and Paleontological Society 8:25–55. 1938 Notes on Certain Explorations on the Battle Place, Red River Bottoms, Lafayette Co. Ark. 1938. MS original in possession of Thomas Gilcrease Museum, Tulsa. Levy, Thomas E., Thomas Higham, Christopher Bronk Ramsey, Neil G. Smith, Erez Ben-Yosef, Mark Robinson, Stefan Munger, Kyle Knabb, Jurgen P. Schulze, Mohammad Najjar, and Lisa Tauxe 2008 High-precision radiocarbon dating and historical biblical archaeology in southern Jordan. Proceedings of the National Academy of Sciences 105(43):16460–16465. Lewis, Kenneth E. 1972 Archaeological Investigations at Ch-37. Miscellaneous Site Reports. Oklahoma River Basin Survey, Norman. Lewis, R. Barry, Charles Stout, and Cameron Wesson 1998 The Design of Mississippian Towns. In Mississippian Towns and Sacred Spaces, edited by R. Barry Lewis and Charles Stout, pp. 1–22, University of Alabama Press, Tuscaloosa. Lintz, Christopher, Floyd Largent, Timothy K. Perttula, Vincent Dongarra, Marsha Prior, and Marie Huhnke

472

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

2007 National Register Evaluative Testing at Prehistoric Sites 16BO450, 16BO458, and 16BO473, Barksdale Air Force Base, Bossier Parish, Louisiana. Report of Investigations, No. 37. United State Air Force Air Combat Command Series. Geo-Marine, Plano tx. Lipo, Carl P., Mark E. Madsen, Robert C. Dunnell, and Tim Hunt 1997 Population Structure, Cultural Transmission, and Frequency Seriation. Journal of Anthropological Archaeology 6:301–333. Livingood, Patrick 2008 Recent Discussions in Late Prehistoric Southern Archaeology. Native South 1:1–26. Lock, Gary 2003 Using Computers in Archaeology: Towards Virtual Pasts. Routledge, London. Lockhart, Jami J. 2007 Prehistoric Caddo of Arkansas: A Multiscalar Examination of Past Cultural Landscapes. Ph.D. dissertation, Department of Environmental Dynamics, University of Arkansas, Fayetteville. Loveland, Carol J. 1986 Human Skeletal Remains from the Clark and Holdeman Sites, Red River County, Texas. Bulletin of the Texas Archeological Society 57:165–182. Lowery, Woodbury 1912 The Lowery Collection: A Descriptive List of Maps of the Spanish Possessions within the Present Limits of the United States, 1502–1820. Edited with notes by Philip Lee Phillips. U.S. Government Printing Office, Washington dc. Lyman, R. Lee, Michael J. O’Brien, and Robert C. Dunnell 1997 The Rise and Fall of Culture History. Plenum Press, New York. Lyman, R. Lee, Steve Wolverton, and Michael J. O’Brien 1998 Seriation, Superposition, and Interdigitation: A History of Americanist Graphic Depictions of Culture Change. American Antiquity 63(2):239–261. Lynott, Mark J., Thomas W. Boutton, James E. Price, and Dwight E. Nelson 1986 Stable Carbon Isotopic Evidence for Maize Agriculture in Southeast Missouri and Northeast Arkansas. American Antiquity 51(1):51–65. Lyon, Edwin A. 1996 A New Deal for Southeastern Archaeology. University of Alabama Press, Tuscaloosa. Marceaux, P. Shawn, and Timothy K. Perttula 2011 Negotiating Borders: The Southern Caddo and Their Relationships with Colonial Governments in East Texas. In Indians and the Market Economy, 1775– 1850, edited by Lance Greene and Mark Plane, 80–97. University of Alabama Press, Tuscaloosa. McCall, Edith 1983 The Attack on the Great Raft. American Heritage of Invention and Technology 3(3):10–16. McKern, Will C. 1939 The Midwestern Taxonomic Method as an Aid to Archaeological Study. American Antiquity 4:301–313. McKinnon, Duncan P. 2008 An Archaeogeophysical Analysis of Central Caddo Settlement Patterning at

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

473

Copyright © 2012. Nebraska Paperback. All rights reserved.

Battle Mound (3LA1). Master’s thesis, Department of Anthropology, University of Arkansas, Fayetteville. 2009 Exploring Settlement Patterning at a Premier Caddo Mound Site in the Red River Great Bend Region. Southeastern Archaeology 28(2):248–258. 2010a Continuing the Research: Archaeogeophysical Investigations at the Battle Mound Site (3LA1) in Lafayette County, Arkansas. Southeastern Archaeology 29(2): 250–260. 2010b Summer 1948: A Summary of Excavations at Battle Mound (3LA1), a Premier Caddo Mound Site in the Great Bend Region of the Red River. Arkansas Archeologist 49:1–16. McWilliams, Robert G. 1981 Iberville’s Gulf Journals. University of Alabama Press, Tuscaloosa. Metcalfe, D., and Kevin T. Jones 1988 A Reconsideration of Body Part Indices. American Antiquity 53(3):486–504. Miewald, Christina E. 2002 The Nutritional Impact of European Contact on the Omaha: A Continuing Legacy. In Food in the USA, edited by Carole Counihan, pp. 109–122. Routledge, New York. Milner, George R. 2004 Old Mounds, Ancient Hunter-Gatherers, and Modern Archaeologists. In Signs of Power: The Rise of Cultural Complexity in the Southeast, edited by Jon L. Gibson and Philip J. Carr, pp. 300–315. University of Alabama Press, Tuscaloosa. Milner, George R., David G. Anderson, and Marvin T. Smith 2001 The Distribution of Eastern Woodlands Peoples at the Prehistoric and Historic Interface. In Societies in Eclipse: Archaeology of the Eastern Woodlands, A.D. 1400– 1700, edited by David S. Brose, C. Wesley Cowan, and Robert C. Mainfort Jr., pp. 9–18. Smithsonian Institution Press, Washington dc. Mitchell, Mark D., and Laura L. Scheiber 2010 Crossing Divides: Archaeology as Long-Term History. In Across a Great Divide: Continuity and Change in Native North American Societies, 1400–1900, edited by Laura L. Scheiber and Mark D. Mitchell, pp. 1–22. University of Arizona Press, Tucson. Mitchell, Sharon A. 2000 Data Salvage Project: Greasy Creek Study. Manuscript on file, Texas Archeological Research Laboratory, University of Texas at Austin. Mooney, James 1896 The Ghost-Dance Religion, and the Sioux Outbreak of 1890. Fourteenth Annual Report of the Bureau of American Ethnology for 1892–1893, Part 2. U.S. Government Printing Office, Washington dc. Moore, Alexander (editor) 1988 Nairne’s Muskogean Journals: The 1708 Expedition to the Mississippi by Captain Thomas Nairne. University Press of Mississippi, Jackson. Moore, Clarence B. 1909 Antiquities of the Ouachita Valley. Journal of the Academy of Natural Sciences of Philadelphia 14 (Part 4):7–170. 1912 Some Aboriginal Sites on Red River. Journal of the Academy of Natural Sciences of Philadelphia 14 (Part 1):481–644.

474

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Moore, Michael C., E. Breitburg, Kevin E. Smith, and Mary Beth Trubitt 2006 One Hundred Years of Archaeology at Gordontown: A Fortified Mississippian Town in Middle Tennessee. Southeastern Archaeology 25(1):89–109. Morehart, Christopher T., and Noah Butler 2010 Ritual Exchange and the Fourth Obligation: Ancient Maya Food Offering and the Flexible Materiality of Ritual. Journal of the Royal Anthropological Institute (n.s.) 16:588–608. Morfi, Fray Juan Agustín de 1932 Excerpts from the Memorias for the History of the Province of Texas. Edited by Frederick M. Chabot. Privately published. Morrow, Juliet E. 2004 The Sacred Spiro Landscape, Cahokia Connections, and Flat Top Mounds. Central States Archaeological Journal 51:112–114. Muller, Jon 1987 Salt, Chert, and Shell: Mississippian Exchange and Economy. In Specialization, Exchange, and Complex Societies, edited by Elizabeth M. Brumfiel and Timothy K. Earle, pp. 10–21. Cambridge University Press, Cambridge. Muto, Guy R. 1978 The Habiukut of Eastern Oklahoma. Parris Mound, Part I, Phase I: An Archaeological Report. Series in Anthropology No. 3. Oklahoma Historical Society, Oklahoma City. Muto, Guy R., Molly S. Mayo, and Kay Zahrai (editors) 1980 From Hunting and Gathering to Business and Banditry: Holocene Adaptations in Lee’s Creek Valley, an Anthropological Assessment. Draft report on file at Oklahoma Archeological Survey, Norman. Neal, William L. 1988a Archeological Survey of Clearcut Areas along Little River, McCurtain and Pushmataha Counties, Oklahoma. Archeological Resources Survey Report 32. Oklahoma Archeological Survey, Norman. 1988b Selective Intensive Survey of Uncut Timberlands in Southeast Oklahoma. Manuscript on file, Oklahoma Archeological Survey, Norman. 1989 Historic and Prehistoric Resources of the Kiamichi River Valley, Oklahoma. Manuscript on file, Oklahoma Archeological Survey, Norman. 1990 Historic and Prehistoric Resources of the Kiamichi River Valley, Oklahoma II. Manuscript on file, Oklahoma Archeological Survey, Norman. Neal, William L., Aubra Lee, and Marjorie Duncan 1991 Archeological Survey for Pre-Allotment Choctaw Farmsteads in Southeastern Oklahoma. Manuscript on file, Oklahoma Archeological Survey, Norman. Neighbours, Kenneth E. 1973 Indian Exodus: Texas Indian Affairs, 1835–1859. Nortex Offset Publications, Quannah. 1975 Robert Simpson Neighbors and the Texas Frontier, 1836–1859. Texian Press, Waco. Nelson, Bo, and Timothy K. Perttula 2003a Archeological Survey along the Lake Bob Sandlin Shoreline, Camp, Franklin, and Titus Counties, Texas. Report of Investigations No. 46. Archeological & Environmental Consultants, Austin.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

475

Copyright © 2012. Nebraska Paperback. All rights reserved.

2003b Archaeological Investigations at the Underwood Site (41CP230), a Titus Phase Settlement along Big Cypress Creek in Camp County, Texas. Journal of Northeast Texas Archaeology 17:1–61. 2006 Archaeological Investigations at the Polk Estates Site (41CP245), Camp County, Texas. Journal of Northeast Texas Archaeology 24:1–83. Newell, H. Perry, and Alex D. Krieger 1949 The George C. Davis Site, Cherokee County, Texas. Memoirs No. 5. Society for American Archaeology, Menasha wi. Newkumet, Vynola B., and Howard L. Meredith 1988 Hasinai: A Traditional History of the Caddo Confederacy. Texas a&m University Press, College Station. Nicholls, G., and M. Jones 2001 Radiocarbon Dating with Temporal Order Constraints. Journal of Royal Statistical Society 50(4):503–521. Nichols, Peter, Larry Banks, Martha D. Freeman, Mark Parsons, Burt Rader, and David Shanabrook 1997 Test Excavations at Proposed Lake Gilmer, Upshur County, Texas. Horizon Environmental Services, Austin tx. Nieman, Fraser D. 1995 Stylistic Variation in Evolutionary Perspective: Inferences from Decorative Diversity and Interassemblage Distance in Illinois Woodland Ceramic Assemblages. American Antiquity 60:7–36. Orr, Kenneth G. 1941 The Eufala Mound: Contributions to the Spiro Focus. Oklahoma Prehistorian 4(1):2–15. 1942 The Eufala Mound, Oklahoma: Contributions to the Spiro Focus. Ph.D. dissertation, Department of Anthropology, University of Chicago, Chicago. 1946 The Archaeological Situation at Spiro, Oklahoma: A Preliminary Report. American Antiquity 11:228–255. Osburn, Tiffany, James Bruseth, and William Pierson 2008 Magnetometer Investigations at the George C. Davis Site, a Prehistoric Caddo Village. Bulletin of the Texas Archeological Society 79:191–200. Palmer, Edward 1917 Arkansas Mounds. Publications of the Arkansas Historical Association 4:390–448. Parker Pearson, Michael 1999 The Archaeology of Death and Burial. Sutton Publishing, Stroud, Gloucestershire, Great Britain. Parsons, Elsie C. 1939 Pueblo Indian Religion. 2 vols. University of Nebraska Press, Lincoln. Parsons, Mark L. 1998 41UR133: A Late Caddo Hamlet at Lake Gilmer. crm News & Views 10(1):16–19. Archeology Division, Texas Historical Commission, Austin. 2010 Mitigation Phase Archeological Investigations at Lake Gilmer, Upshur County, Texas. Draft report on file, Archeology Division, Texas Historical Commission, Austin.

476

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Parsons, Mark L., James E. Bruseth, Jacques Bagur, S. Eileen Goldborer, and Claude McCrocklin 2002 Finding Sha’chahdinnih (Timber Hill): The Last Village of the Kadohadacho in the Caddo Homeland. Archeological Reports Series No. 3. Texas Historical Commission, Austin. Pauketat, Timothy R. 2001 Ancient Cahokia and the Mississippians. Cambridge University Press, Cambridge. 2003 The Forgotten History of the Mississippians. In North American Archaeology, edited by Timothy R. Pauketat and Diana DiPaolo Loren, pp. 187–211. Blackwell Publishing, Malden ma. 2007 Chiefdoms and Other Archaeological Delusions. AltaMira Press, Lanham md. Pauketat, Timothy R., Lucretia Kelly, Gayle J. Fritz, Neal H. Lopinot, Scott Elias, and Eve Hargrove 2002 The Residues of Feasting and Public Ritual at Early Cahokia. American Antiquity 67(2):257–280. Payne, Claudine 2002 Architectural Reflections of Power and Authority in Mississippian Towns. In The Dynamics of Power, edited by Maria O’Donovan, pp. 188–213. Occasional Paper No. 30. Center for Archaeological Investigations, Southern Illinois University at Carbondale. Payne, Claudine, and John F. Scarry 1998 Town Structure at the Edge of the Mississippian World. In Mississippian Towns and Sacred Spaces: Searching for an Architectural Grammar, edited by R. Barry Lewis and Charles Stout, pp. 22–48. University of Alabama Press, Tuscaloosa. Pearson, Charles E. 1986 Dating the Course of the Lower Red River in Louisiana: the Archaeological Evidence. Geoarchaeology 1:39–42. Pearson, Charles E., and Donald G. Hunter 1993 Geoarchaeology of the Red River Valley. In Quaternary Geology and Geoarchaeology of the Lower Red River Valley, edited by W. Autin and J. Snead, pp. 25–43. Friends of the Pleistocene, South Central Cell, 11th Annual Field Conference Guidebook. Pearson, F. J., Jr., E. Mott Davis, Murray A. Tamers, and Robert W. Johnston 1965 University of Texas Radiocarbon Dates III. Radiocarbon 7;296–314. Pearson, F. J., Jr., E. Mott Davis, and Murray A. Tamers 1966 University of Texas Radiocarbon Dates IV. Radiocarbon 8:453–466. Perino, Gregory 1983 Archaeological Research at the Bob Williams Site (41RR16), Red River County, Texas. Museum of the Red River, Idabel ok. 1994 The Dan Holdeman Site (41RR11), Red River County, Texas. Journal of Northeast Texas Archaeology 6:3–65. Perttula, Timothy K. 1989 A Study of Mound Sites in the Sabine River Basin, Northeast Texas and Northwest Louisiana. Final report submitted to the Texas Historical Commission by the University of North Texas, Institute of Applied Sciences, Denton. 1992 “The Caddo Nation”: Archaeological and Ethnohistorical Perspectives. University of Texas Press, Austin.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

477

Copyright © 2012. Nebraska Paperback. All rights reserved.

1993 Historic Context: The Development of Agriculture in Northeast Texas before A.D. 1600. In Archeology in the Eastern Planning Region, Texas: A Planning Document, edited by Nancy A. Kenmotsu and Timothy K. Perttula, pp. 121–146. Cultural Resource Management Report 3. Department of Antiquities Protection, Texas Historical Commission, Austin. 1994 Caddoan Mound Sites in the Sabine River Basin of Northeast Texas. Caddoan Archeology Newsletter 4(4):4–19. 1995 The Archeology of the Pineywoods and Post Oak Savanna of Northeast Texas. Bulletin of the Texas Archeological Society 66:331–359. 1996 Caddoan Area Archaeology Since 1990. Journal of Archaeological Research 4(4):295–348. 1997 A Compendium of Radiocarbon and Oxidizable Carbon Ratio Dates from Archaeological Sites in East Texas, with a Discussion of the Age and Dating of Select Components and Phases. Radiocarbon 39(3):305–341. 1998 Late Caddoan Societies in the Northeast Texas Pineywoods. In The Native History of the Caddo: Their Place in Southeastern Archeology and Ethnohistory, edited by Timothy K. Perttula and James E. Bruseth, pp. 69–90. Studies in Archeology 30. Texas Archeological Research Laboratory, University of Texas at Austin. 2000 Functional and Stylistic Analyses of Ceramic Vessels from Mortuary Features at a 15th and 16th Century Caddo Site in Northeast Texas. Midcontinental Journal of Archaeology 25(1):101–151. 2002 Archaeological Evidence for the Long-Distance Exchange of Caddo Indian Ceramics in the Southern Plains, Midwest, and Southeastern United States. In Geochemical Evidence for Long-Distance Exchange, edited by Michael D. Glascock, pp. 89–107. Bergin & Garvey, Westport ct. 2004a Decorated Ceramics. In The Oak Hill Village Site (41RK214), Rusk County, Texas, by Robert Rogers and Timothy K. Perttula, pp. 209–271. Document No. 030083. pbs&j, Austin. 2004b The Prehistoric and Caddoan Archeology of the Northeastern Texas Pineywoods. In The Prehistory of Texas, edited by Timothy K. Perttula, pp. 370–407. Texas a&m University Press, College Station. 2005 1938–1939 WPA Excavations at the Hatchel Site (41BW3) on the Red River in Bowie County, Texas. Southeastern Archaeology 24(2):180–198. 2006a A Study of the Buddy Jones Collection from Northeast Texas Caddo Sites. Special Publication No. 6. Friends of Northeast Texas Archaeology, Pittsburg and Austin. 2006b The Structure and Growth of a Titus Phase Community Cemetery in Titus County, Texas. Journal of Northeast Texas Archaeology 25:1–18. 2008 Caddo Agriculture on the Western Frontier of the Eastern Woodlands. Plains Anthropologist 53(205):79–105. 2009 Extended Entranceway Structures in the Caddo Archaeological Area. Southeastern Archaeology 28(1):27–42. 2010 Papers on Geophysical Investigations of Woodland and Caddo Sites in the Caddo Area of the Southeastern U.S. Southeastern Archaeology 29(2):233–235. Perttula, Timothy K. (editor) 2005 Archeological Investigations at the Pilgrim’s Pride Site (41CP304), a Titus Phase

478

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Community in the Big Cypress Creek Basin, Camp County, Texas. 2 Vols. Report of Investigations No. 30. Archeological & Environmental Consultants, llc, Austin. 2008 The Archeology of the Roitsch Site (41RR16), an Early to Historic Caddo Period Village on the Red River in Northeast Texas. In Collected Papers from Past Texas Archeological Society Summer Field Schools, edited by Timothy K. Perttula, pp. 313–628. Special Publication No. 5. Texas Archeological Society, San Antonio. Perttula, Timothy K., and James E. Bruseth 1983 Early Caddoan Subsistence Strategies, Sabine River Basin, East Texas. Plains Anthropologist 28(99):9–21. Perttula, Timothy K., Cathy Crane, and James E. Bruseth 1982 A Consideration of Caddoan Subsistence. Southeastern Archaeology 1(2):89–102. Perttula, Timothy K., Elsbeth Dowd, Lee Green, George Morgan, Bo Nelson, LeeAnna Schniebs, Beau Schriever, Jesse Todd, and Mark Walters 2009 The Archaeology of the 16th and 17th Century Caddo in the Post Oak Savannah of Northeast Texas: The Tuinier Farm (41HP237), R. A. Watkins (41HP238), and Anglin (41HP240) Sites in the Stouts Creek Basin, Hopkins County, Texas. Journal of Northeast Texas Archaeology 30:1–132. Perttula, Timothy K., Ann M. Early, Lois E. Albert, and Jeffrey Girard (compilers) 2006 Caddo Bibliography: Archaeology and Bioarchaeology, Ethnohistory and Ethnography, and History. Technical Paper 10. Arkansas Archeological Survey, Fayetteville. Perttula, Timothy K., and Jeffrey R. Ferguson 2010 The Chemical Variation in Prehistoric and Early Historic Caddo Ceramics in Eastern Texas. In Studies on the Instrumental Neutron Activation Analysis of Woodland Period and Caddo Tradition Ceramics from Eastern Texas, compiled by Timothy K. Perttula, pp. 3–1 to 3–62. Special Publication No. 17. Friends of Northeast Texas Archaeology, Austin and Pittsburgh. Perttula, Timothy K., Marlin F. Hawley, and Fred W. Scott 2001 Caddo Trade Ceramics. Southeastern Archaeology 20(2):154–172. Perttula, Timothy K., Sergio A. Iruegas, and Hector Neff 2003 Caddoan Pottery in Central Texas: Geochemical Analyses of Ceramics from Fort Hood and Vicinity. Research Report No. 51, Archeological Resource Management Series. U.S. Army, Fort Hood. Perttula, Timothy K., and Bo Nelson 1998 Titus Phase Mortuary Practices in the Northeast Texas Pineywoods and Post Oak Savanna. In Analysis of the Titus Phase Mortuary Assemblage at the Mockingbird or “Kahbakayammaahin” Site (41TT550), by Timothy K. Perttula, Melinda Tate, Hector Neff, James W. Cogswell, Michael D. Glascock, Elizabeth Skokan, Susan Mulholland, Robert Rogers, and Bo Nelson, pp. 328–401. Document No. 970849. Espey, Huston & Associates, Austin. 2001 Archeological Investigations at the Camp Joy Mound (41UR144): A Titus Phase Earthen Mound Site at Lake O’ the Pines, Upshur County, Texas. Report of Investigations No. 44. Archeological & Environmental Consultants, Austin. 2002 Archeological Survey of Lake Bob Sandlin State Park, Titus County, Texas. Report of Investigations No. 48. Archeological and Environmental Consultants, Austin. 2003a The Nawi haia ina Site (41RK170): Archeological Investigations in the City of Henderson’s

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

479

Copyright © 2012. Nebraska Paperback. All rights reserved.

Southside Wastewater Treatment Plant, Rusk County, Texas. Report of Investigations No. 51. Archeological & Environmental Consultants, llc,Austin. 2003b Archeological Investigations of Village Areas at the Hatchel Site (41BW3), Bowie County, Texas. Report of Investigations No. 58. Archeological & Environmental Consultants, llc, Austin. 2003c Temporal and Spatial Patterns in the Prehistoric Settlement of the Lake Bob Sandlin Area, Big Cypress Creek Basin, Northeastern Texas. Caddoan Archeology Journal 13(2):28–35. 2004a Woodland and Caddo Archeology at the Broadway or Kanduts’ah Kuhnihdahahdisa’ Site (41SM273) on the City of Tyler-Lake Palestine wtp Project, Smith County, Texas. Report of Investigations No. 50. Archeological & Environmental Consultants, llc, Austin. 2004b Archeological Investigations at Five Prehistoric Sites at Broken Bow Reservoir, McCurtain County, Oklahoma. Report of Investigations No. 49. Archeological & Environmental Consultants, llc, Austin. 2007 The Gum Creek Cluster: Protohistoric Caddo Sites in the Little Cypress Creek Basin, ca. 1670–1720. Journal of Northeast Texas Archaeology 26:128–135. Perttula, Timothy K., Bo Nelson, Robert L. Cast, and Bobby Gonzalez 2010 The Clements Site (41CS25): A Late 17th to Early 18th-Century Nasoni Caddo Settlement and Cemetery. Anthropological Papers No. 92. American Museum of Natural History, New York. Perttula, Timothy K., Bo Nelson, J. Phil Dering, LeeAnna A. Schniebs, Robert L. Turner Jr., Mark Walters, and Diane E. Wilson 2004 Archeological Investigations at the Shelby Site (41CP71) on Greasy Creek, Camp County, Texas. Special Publication No. 5. Friends of Northeast Texas Archaeology, Austin and Pittsburg. Perttula, Timothy K., Daniel J. Prikryl, Bo Nelson, and Robert Cast 1998 Archeological Survey of the Broken Bow Lake Shoreline and the Relocation and Evaluation of Nine Previously Recorded Sites, McCurtain County, Oklahoma. Report of Investigations No. 7. Archeological and Environmental Consultants, Austin. Perttula, Timothy K., Bo Nelson, Mark Walters, Robert Cast, and Bobby Gonzalez 2009 Documentation of Caddo Funerary Objects in the Gilcrease Museum Collections. Historic Preservation Program, Caddo Nation of Oklahoma, Binger, Oklahoma. Perttula, Timothy K., and Robert Rogers 2007 The Evolution of a Caddo Community in Northeastern Texas: The Oak Hill Village Site (41RK214), Rusk County, Texas. American Antiquity 72(1):71–94. Perttula, Timothy K., and David L. Sherman 2009 Data Recovery Investigations at the Ear Spool Site (41TT653), Titus County, Texas. Document No. 070205. pbs&j, Austin. Perttula, Timothy K., Bob D. Skiles, and Bonnie C. Yates 1993 The Goldsmith Site (41WD208): Investigations of the Titus phase in the Upper Sabine River Basin, Northeast Texas. Bulletin of the Texas Archeological Society 61:139–191. Perttula, Timothy K., Melinda Tate, Hector Neff, James W. Cogswell, Michael D. Glascock, Elizabeth Skokan, Susan Mulholland, Robert Rogers, and Bo Nelson 1998 Analysis of the Titus Phase Mortuary Assemblage at the Mockingbird or

480

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

“Kahbakayammaahin” Site (41TT550). Document No. 970849. Espey, Huston & Associates, Austin. Perttula, Timothy K., and Chester P. Walker 2008 The History of Archaeological Investigations and Geophysical Survey at the Jamestown Mound Site (41SM54), an Archaeological Conservancy Preserve in Smith County, Texas. Archeological & Environmental Consultants, llc, and Archaeo-Geophysical Associates, Austin. Perttula, Timothy K., Chester P. Walker, and T. Clay Schultz 2008 A Revolution in Caddo Archaeology: The Remote Sensing and Archaeological View from the Hill Farm Site (41BW169) in Bowie County, Texas. Southeastern Archaeology 27(1):93–10 Phillips, Philip 1970 Archaeological Survey in the Lower Yazoo Basin, Mississippi, 1949–1955. Papers of the Peabody Museum of Archaeology and Ethnology Vol. 60. Harvard University, Cambridge, Massachusetts. Phillips, Philip, and James A. Brown 1978 Pre-Columbian Shell Engravings from the Craig Mound at Spiro, Oklahoma. Paperback Edition, Part 1. Peabody Museum Press, Harvard University, Cambridge. 1984 Pre-Columbian Shell Engravings from Craig Mound at Spiro, Oklahoma. Part 2. Peabody Museum Press, Harvard University, Cambridge. Pohl, Mary 1983 Maya Ritual Faunas: Vertebrate Remains form Burials, Caches, Caves and Cenotes in the Maya Lowlands. In Civilization in the Ancient Americas, edited by A. Kolata, pp. 55–103. University of New Mexico Press, Albuquerque. Polhemus, Richard R. 1987 The Toqua Site: A Late Mississippian Dallas Phase Town. Report of Investigations 41, Department of Anthropology, University of Tennessee, and Publications in Anthropology 44, Tennessee Valley Authority, Knoxville. Potter, James M. 2000 Ritual, Power, and Social Differentiation in Small-Scale Societies. In Hierarchies in Action: Cui Bono?, edited by Michael W. Diehl, pp. 295–316. Occasional Paper No. 27. Center for Archaeological Investigations, Southern Illinois University at Carbondale. Powell, John Wesley 1891 Indian Linguistic Families of America North of Mexico. In 7th Annual Report of the Bureau of [American] Ethnology for 1885–1886, pp. 1–142. Bureau of Ethnology, Washington dc. Powell, Mary L. 1977 Prehistoric Ritual Skulls at the Crenshaw Site, (3MI16), Southwest Arkansas. Bulletin of the Texas Archeological Society 48:111–118. 1985 The Analysis of Dental Wear and Caries for Dietary Reconstruction. In The Analysis of Prehistoric Diets, edited by R. I. Gilbert Jr. and James H. Mielke, pp. 307–338. Academic Press, Orlando. Prentice, Guy 1986 An Analysis of the Symbolism Expressed by the Birger Figurine. American Antiquity 51:239–266.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

481

Copyright © 2012. Nebraska Paperback. All rights reserved.

2000 Ancient Indian Architecture of the Lower Mississippi Delta Region. 2nd edition. Southeast Archeological Center, National Park Service, Tallahassee fl. Prewitt, Terry J. 1969 Stylistic Variation in a Caddoan Ceramic Type. Bulletin of the Oklahoma Anthropological Society 17:59–74. Prewitt, Terry J., Foster Kirby, and Dorothy Gaston 1982 Archaeological Investigations of the Waterfall-Gilford Creeks Watershed, McCurtain County, Oklahoma. Laboratory of Archaeology, University of Tulsa, Tulsa. Purrington, Burton L. 1971 The Prehistory of Delaware County, Oklahoma: Cultural Continuity and Change on the Western Ozark Periphery. Ph.D. dissertation, Department of Anthropology, University of Wisconsin, Madison. Pursell, Corin C. 2004 Geographic Distribution and Symbolism of Colored Mound Architecture in the Mississippian Southeast. Paper presented at the combined 50th Midwest Archaeological Conference and 61st Southeastern Archaeological Conference, St. Louis. Ray, Jack H. 1996 Preservation Measures and Limited Test Excavations at the Pineville Site. Missouri Archaeological Society Quarterly 13(3):12–18. Ray, Jack H., and Neal H. Lopinot 2008 Late Prehistoric Culture History for Southwest Missouri. The Missouri Archaeologist 69:57–102. Reber, Eleanora A. 2006 A Hard Row to Hoe: Changing Maize Use in the American Bottom and Surrounding Areas. In Histories of Maize, edited by John Staller, Robert Tykot, and Bruce Benz, pp. 236–248. Academic Press, Burlington ma. Rees, Mark A., and Patrick C. Livingood (editors) 2007 Plaquemine Archaeology. University of Alabama Press, Tuscaloosa. Reilly, F. Kent, III 2004 People of Earth, People of Sky: Visualizing the Sacred in Native North American Art of the Mississippian Period. In Hero, Hawk, and Open Hand: American Indian Art of the Ancient Midwest and South, edited by Richard F. Townsend, pp. 125–138. The Art Institute of Chicago and Yale University Press, New Haven. Reilly, F. Kent, III and James F. Garber (editors) 2007 Ancient Objects and Sacred Realms: Interpretations of Mississippian Iconography. University of Texas Press, Austin. Reimer, Paula J., Mike G. L. Baillie, Edouard Bard, Alex Bayliss, J. Warren Beck, Chanda J. H. Bertrand, Paul G. Blackwell, Caitlin E. Buck, George S. Burr, Kirsten B. Cutler, Paul E. Damon, R. Lawrence Edwards, Richard G. Fairbanks, Michael Friedrich, Thomas P. Guilderson, Alan G. Hogg, Konrad A. Hughen, Bernd Kromer, F. Gerry McCormac, Sturt W. Manning, Christopher Bronk Ramsey, Ron W. Reimer, Sabine Remmele, John R. Southon, Minze Stuiver, Sahra Talamo, F. W. Taylor, Johannes van der Plicht, and Constanze E. Weyhenmeyer 2004 IntCal04 Terrestrial Radiocarbon Age Calibration, 26–0 ka BP. Radiocarbon 46:1029–1058.

482

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Reimer, Paula J., M. G. I. Baillie, E. Bard, A. Bayliss, J. W. Beck, P. G. Blackwell, C. Bronk Ramsey, C. E. Buck, G. S. Burr, R. L. Edwards, M. Friedrich, P. M. Grootes, T. P. Guilderson, I. Hajdas, T. J. Heaton, A. G. Hogg, K. A. Hughen, Kaiser K. F., B. Kromer, F. G. McCormac, S. W. Manning, R. W. Reimer, D. A. Richards, J. R. Southon, S. Talamo, C. S. M. Turney, J. van der Plicht, and C. E. Weyhenmeyer 2009 IntCal09 and Marine09 Radiocarbon Age Calibration Curves, 0-50,000 Years Cal BP. Radiocarbon 51:1111–1150. Renfrew, Colin 2001 Production and Consumption in a Sacred Economy: The Material Correlates of High Devotional Expression at Chaco Canyon. American Antiquity 66(1):14–25. Reynolds, Matthew D., with contributions by Jeffrey Gaskin, Ankita Kumar, Elayne Pope, Jerome C. Rose, Robert J. Scott, Mary Beth Trubitt, and James C. Tyler 2007 Preliminary Report on Salvage Excavations at 3CL593. Limited distribution report submitted to Arkansas Historic Preservation Program, Little Rock. Arkansas Archeological Survey, Henderson State University Research Station, Arkadelphia. Riley, Carroll L. 2005 Becoming Aztlan: Mesoamerican Influence in the Greater Southwest, A.D. 1200–1500. University of Utah Press, Salt Lake City. Robinson, William S. 1951 A Method for Chronologically Ordering Archaeological Deposits. American Antiquity 16(4):295–300. Rogers, J. Daniel 1978 Federally Sponsored Archaeological Work in Oklahoma Before World War II. Manuscript on file, Oklahoma Archeological Survey, Norman. 1983 Social Ranking and Change in the Harlan and Spiro Phases of Eastern Oklahoma. In Southeastern Natives and Their Pasts: Papers Honoring Dr. Robert E. Bell, edited by Don G. Wyckoff and Jack L. Hofman, pp. 17–28. Studies in Oklahoma’s Past No. 11. Oklahoma Archeological Survey, Norman. 1991 Patterns of Change on the Western Margins of the Southeast, A.D. 600–900. In Stability, Transformation, and Variation: The Late Woodland Southeast, edited by Michael S. Nassaney and Charles R. Cobb, pp. 221–248. Plenum Press, New York. 1995 Dispersed Communities and Integrated Households: A Perspective from Spiro and the Arkansas Basin. In Mississippian Communities and Households, edited by J. Daniel Rogers and Bruce D. Smith, pp. 81–98. University of Alabama Press, Tuscaloosa. 1996 Markers of Social Integration: The Development of Centralized Authority in the Spiro Region. In Political Structure and Change in the Prehistoric Southeastern United States, edited by John F. Scarry, pp. 53–68. University Press of Florida, Gainesville. 2006 Chronology and the Demise of Chiefdoms: Eastern Oklahoma in the Sixteenth and Seventeenth Centuries. Southeastern Archaeology 25(1):20–28. Rogers, J. Daniel, and George Sabo III 2004 Caddo. In Handbook of North American Indians, Vol. 14, Southeast, edited by Raymond D. Fogelson, pp. 616–631. Smithsonian Institution, Washington dc.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

483

Copyright © 2012. Nebraska Paperback. All rights reserved.

Rogers, Robert 2004 The Caddoan Village. In The Oak Hill Village Site (41RK214), Rusk County, Texas, by Robert Rogers and Timothy K. Perttula, pp. 49–96. Document No. 030083. pbs&j, Austin. Rogers, Robert, Maynard B. Cliff, Timothy K. Perttula, Gary Rutenberg, Sally Victor, Phil Dering, and Mary Malainey 2003 Excavations at the Alex Justiss Site, 41TT13, Titus County, Texas. Archeological Studies Program Report No. 36. Environmental Affairs Division, Texas Department of Transportation, Austin. Rogers, Robert, Eugene Foster, and Kathryn Reese-Taylor 1994 National Register Testing at Eight Archaeological Sites within the Oak Hill 2,280Acre Study Area, Rusk County, Texas. Document No. 930169. Espey, Huston & Associates, Austin. Rogers, Robert, and Timothy K. Perttula 2004 The Oak Hill Village Site (41RK214), Rusk County, Texas. Document No. 030083. pbs&j, Austin. Rohrbaugh, Charles L. 1973 Hugo Reservoir III: A Report on Early Formative Cultural Manifestations in Hugo Reservoir and in the Caddoan Area. Archaeological Site Report No. 24. Oklahoma River Basin Survey, Norman. 1982 Spiro and Fort Coffee Phases: Changing Cultural Complexes of the Caddoan Area. Ph.D. dissertation, Department of Anthropology, University of Wisconsin, Madison. 1984 Arkansas Valley Caddoan: Fort Coffee and Neosho. In Prehistory of Oklahoma, edited by Robert E. Bell, pp. 265–285. Academic Press, Orlando. Rohrbaugh, Charles L., Robert J. Burton, Susan S. Burton, and Laura J. Rosewitz 1971 Hugo Reservoir I: The Description of the Archaeological Sites Excavated During the 1970 Field Season Including Ch-1, Ch-43, Ch-70, Ch-75, Pu-82, and Ch-90. Archaeological Site Report No. 22. Oklahoma River Basin Survey, Norman. Rolingson, Martha A., and Frank F. Schambach 1981 The Shallow Lake Site (3UN9/52) and its Place in Regional Prehistory. Research Series No. 12. Arkansas Archeological Survey, Fayetteville. Rose, Jerome C., Barbara A. Burnett, Michael S. Nassaney, and F. M. Blaeuer 1984 Paleopathology and the Origins of Maize Agriculture in the Lower Mississippi Valley and Caddoan Culture Areas. In Paleopathology at the Origins of Agriculture, edited by Mark N. Cohen and George J. Armelagos, pp. 393–425. Academic Press, New York. Rose, Jerome C., Michael P. Hoffman, Barbara A. Burnett, Anna M. Harmon and James E. Barnes 1998 Skeletal Biology of the Prehistoric Caddo. In The Native History of the Caddo: Their Place in Southeastern Archaeology and Ethnohistory, edited by Timothy K. Perttula and James E. Bruseth, pp. 113–128. Studies in Archeology 30. Texas Archeological Research Laboratory, University of Texas at Austin. Rose, Jerome C., and Murray K. Marks 1985 Bioarcheology of the Alexander Site. In The Alexander Site, Conway County, Arkansas, edited by E. Thomas Hemmings and John H. House, pp. 79–98. Research Series No. 24. Arkansas Archeological Survey, Fayetteville.

484

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Rousseeuw, Peter J. 1987 Silhouettes: A Graphical Aid to the Interpretation and Validation of Cluster Analysis. Journal of Computational and Applied Mathematics 20:53–65. Sabo, George, III 1982 The Huntsville Site (3MA22): A Caddoan Civic Ceremonial Center in the Arkansas Ozarks. Paper presented at the 47th Annual Meeting of the Society for American Archaeology, Minneapolis. 1986 Preliminary Excavations at the Huntsville Site. In Contributions to Ozark Prehistory, edited by George Sabo III, pp. 55–76. Research Series No. 27. Arkansas Archeological Survey, Fayetteville. 1995 Images and Encounters: European Contact and the Caddo Indians. Historical Reflections 21(2):217–242. 1998 The Structure of Caddo Leadership in the Colonial Era. In The Native History of the Caddo: Their Place in Southeastern Archeology and Ethnohistory, edited by Timothy K. Perttula and James E. Bruseth, pp. 159–174. Studies in Archeology 30, Texas Archeological Research Laboratory, University of Texas at Austin. 2005 Dancing into the Past: Colonial Legacies in Modern Caddo Indian Ceremony. In A Whole Country in Commotion: The Louisiana Purchase and the American Southwest, edited by Patrick G. Williams, S. Charles Bolton, and Jeannie M. Whayne, pp. 149–168. University of Arkansas Press, Fayetteville. 2008 Rock Art and the Study of Ancient Religions in Southeastern North America. In Religion, Archaeology, and the Material World, edited by Lars Fogelin, pp. 279–296. Occasional Paper No. 36. Center for Archaeological Investigations, Southern Illinois University, Carbondale. Sahlins, Marshall 1972 Stone Age Economics. Aldine, Chicago. Samuelsen, John R. 2009 Archaeogeophysical Investigations of Early Settlement Patterns at the Crenshaw Site (3MI6). Master’s Thesis, Department of Anthropology, University of Arkansas, Fayetteville. Sanderson, Alan R. 2008 Ritual Economy Among the Nahua of Northern Veracruz, Mexico. In Dimensions of Ritual Economy, edited by E. Christian Wells and Patricia A. McAnany. Research in Economic Anthropology 27:93–119. Saucier, Roger T., and John I. Snead (compilers) 1989 Quaternary Geology of the Lower Mississippi Valley. In Quaternary Non-glacial Geology: Conterminous United States, edited by Roger B. Morrison, plate 6. Geological Society of America, Boulder. Scarry, John F. 1999 How Great Were the Southeastern Polities. In Great Towns and Regional Polities in the Prehistoric American Southwest and Southeast, edited by Jill E. Neitzel, pp. 57–74. University of New Mexico Press, Albuquerque. Schambach, Frank F. 1970 Pre-Caddoan Cultures in the Trans-Mississippi South: A Beginning Sequence. Ph.D. dissertation, Department of Anthropology, Harvard University, Cambridge.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

485

Copyright © 2012. Nebraska Paperback. All rights reserved.

1972 Exploratory Excavations in the Midden Areas at the Crenshaw Site, Miller County, Arkansas. Interim Report submitted to the National Science Foundation on research conducted under grant GS-2684. 1977 Deer Ceremonialism in Early Caddo Culture: Evidence from the Crenshaw Site in Southwest Arkansas. Manuscript on file, Arkansas Archeological Survey, Fayetteville. 1982a The Archeology of the Great Bend Region in Arkansas. In Contributions to the Archeology of the Great Bend Region, edited by Frank F. Schambach and Frank Rackerby, pp. 1–11. Research Series No. 22. Arkansas Archeological Survey, Fayetteville. 1982b A Research Design for Continued Investigations of the Caddo V Component at the Cedar Grove Site. In Contributions to the Archeology of the Great Bend Region, edited by Frank F. Schambach and Frank Rackerby, pp. 118–122. Research Series No. 22. Arkansas Archeological Survey, Fayetteville. 1982c An Outline of Fourche Maline Culture in Southwest Arkansas. In Arkansas Archeology in Review, edited by Neal L. Trubowitz and Marvin D. Jeter, pp. 132–197. Research Series No. 15. Arkansas Archeological Survey, Fayetteville. 1989 The End of the Trail: The Route of Hernando de Soto’s Army Through Southwest Arkansas and East Texas. Arkansas Archeologist 27/28:9–34. 1990 The “Northern Caddoan Area” Was Not Caddoan. Caddoan Archeology 1(4):2–6. 1995 A Probable Spiroan Entrepot in the Red River Valley in Northeast Texas. Caddoan Archeology Newsletter 6(1):10–25. 1996 Mounds, Embankments, and Ceremonialism in the Trans-Mississippi South. In Mounds, Embankments, and Ceremonialism in the Midsouth, edited by Robert C. Mainfort and Richard Walling, pp. 36–43. Research Series No. 46. Arkansas Archeological Survey, Fayetteville. 1997 The Development of the Burial Mound Tradition in the Caddo Area. Journal of Northeast Texas Archaeology 9:53–72. 1998 Pre-Caddoan Cultures in the Trans-Mississippi South. Research Series 53. Arkansas Archeological Survey, Fayetteville. 1999 Spiro and the Tunica: A New Interpretation of the Role of the Tunica in the Culture History of the Southeast and the Southern Plains, A.D. 1100–1750. In Arkansas Archaeology: Essays in Honor of Dan and Phyllis Morse, edited by Robert C. Mainfort Jr. and Phyllis Morse, pp. 169–224. University of Arkansas Press, Fayetteville. 2000a Introduction: The Significance of the Sanders Site in the Culture History of the Mississippi Period Southeast and the Southern Plains. In The 1931 Excavations at the Sanders Site, Lamar County, Texas, by A. T. Jackson, Marcus S. Goldstein, and Alex D. Krieger, pp. 1–7. Archival Series No. 2. Texas Archeological Research Laboratory, University of Texas at Austin. 2000b Spiroan Traders, the Sanders Site, and the Plains Interaction Sphere: A Reply to Bruseth, Wilson, and Perttula. Plains Anthropologist 45(171):7–33. 2001 Fourche Maline and its Neighbors: Observations on an Important Woodland Period Culture of the Trans-Mississippi South. The Arkansas Archeologist 40:21–50. 2002 Fourche Maline: A Woodland Period Culture of the Trans-Mississippi South. In The Woodland Southeast, edited by David G. Anderson and Robert C. Mainfort Jr., pp. 90–112. University of Alabama Press, Tuscaloosa.

486

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

2003 Some Highlights of the June 7–22, 2003 Fieldwork at the Grandview Prairie Wildlife Management Area. Field Notes (Newsletter of the Arkansas Archeological Society) 315:3–8. Schambach, Frank F., and Ann M. Early 1982 Southwest Arkansas. In A State Plan for the Conservation of Archeological Resources in Arkansas, edited by Hester A. Davis. Research Series No. 21. Arkansas Archeological Survey, Fayetteville. Schambach, Frank F., and John E. Miller 1984 A Description and Analysis of the Ceramics. In Cedar Grove: An Interdisciplinary Investigation of a Late Caddo Farmstead in the Red River Valley, edited by Neal L. Trubowitz, pp. 109–170. Research Series No. 23. Arkansas Archeological Survey, Fayetteville. Schambach, Frank F., Neal L. Trubowitz, Frank Rackerby, E. Thomas Hemmings, W. Frederick Limp, and John E. Miller III 1982 Test Excavations at the Cedar Grove Site (3LA97): A Late Caddo Farmstead in the Great Bend Region, Southwest Arkansas. In Contributions to the Archeology of the Great Bend Region, edited by Frank F. Schambach and Frank Rackerby, pp. 90–129. Research Series No. 23. Arkansas Archeological Survey, Fayetteville. Schiffer, Michael B. 1978 Methodological Issues in Ethnoarchaeology. In Explorations in Ethnoarchaeology, edited by Richard A. Gould, pp. 229–247. University of New Mexico Press, Albuquerque. Schoeninger, Margaret J., and Michael J. DeNiro 1983 Stable Nitrogen Isotope Ratios of Bone Collagen Reflect Marine and Terrestrial Components of Prehistoric Human Diet. Science 220:1381–1383. Schoeninger, Margaret, and Mark Schurr 1998 Human Subsistence at Moundville: The Stable Isotope Data. In Archaeology of the Moundville Kingdom, edited by Vernon J. Knight and Vincas P. Steponaitis, pp. 120–132. Smithsonian Institution Press, Washington dc. Schultz, T. Clay 2010 Architectural Variability in the Caddo Area of Eastern Texas. Ph.D. dissertation, Department of Anthropology, University of Texas at Austin. Schurr, Mark R. 1992 Isotopic and Mortuary Variability in a Middle Mississippian Population. American Antiquity 57(2):300–320. Scott, E. C. 1979 Principal Axis Analysis of Dental Attrition Data. American Journal of Physical Anthropology 51:203–211. Scott, Susan L. 1983 Analysis, Synthesis and Interpretation of Faunal Remains from the Lubbub Creek Archaeological Locality. In Prehistoric Agricultural Communities in West Central Alabama, edited by Christopher S. Peebles, pp. 272–390. U.S. Army Corps of Engineers, Mobile District, Mobile. 2005 Vertebrate Fauna. In Lake Providence: A Terminal Coles Creek Culture Mound Center, East Carroll Parish, Louisiana, edited by Richard A. Weinstein, pp. 411–429. Report prepared for the U. S. Army Corps of Engineers, Vicksburg District. Coastal Environments, Baton Rouge.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

487

Copyright © 2012. Nebraska Paperback. All rights reserved.

Scott, Susan L., and H. Edwin Jackson 1998 Analysis of the Faunal Remains from the Crenshaw Site (3MI6), Southwestern Arkansas. The Arkansas Archeologist 37:1–37. Shennan, Stephen 1988 Quantifying Archaeology. Academic Press, San Diego. Sibley, John 1805 Letter to General Henry Dearborn, April 5, Natchitoches. In American State Papers: Indian Affairs, 1:729. U.S. Government Printing Office, Washington dc. Sierzchula, Michael C., Margaret J. Guccione, Robert H. Lafferty III, and M. Tracy Oates 1995 Archeological Investigations in the Great Bend Region, Miller County, Arkansas, Levee Items 2 and 3. Mid-Continental Research Associates. Submitted to the U.S. Army Corps of Engineers, Little Rock District. Silverman, Helaine 2002 Introduction: The Space and Place of Death. In The Space and Place of Death, edited by Helaine Silverman and David B. Small, pp. 1–11. Archeological Papers No. 11. American Anthropological Association, Washington dc. Skinner, S. Alan, R. King Harris, and Keith M. Anderson (editors) 1969 Archaeological Investigations at the Sam Kaufman Site, Red River County, Texas. Contributions in Anthropology No. 5. Southern Methodist University, Dallas. Smith, B. Holly 1984 Patterns of Molar Wear in Hunter-Gatherers and Agriculturalists. American Journal of Physical Anthropology 63:39–56. Smith, Bruce D. 1978b Prehistoric Patterns of Human Behavior: A Case Study in the Mississippi Valley. Academic Press, New York. Smith, Bruce D. (editor) 1978a Mississippian Settlement Patterns. Academic Press, New York. Smith, F. Todd 1996 The Caddos, the Wichitas, and the United States, 1846–1901. Texas a&m University Press, College Station. 2005 From Dominance to Disappearance: The Indians of Texas and the Near Southwest, 1786–1859. University of Nebraska Press, Lincoln. Smith, Karen Y., and Fraser D. Nieman 2007 Frequency Seriation, Correspondence Analysis, and Woodland Period Ceramic Assemblage Variation in the Deep South. Southeastern Archaeology 26(1):47–72. Spielmann, Katherine A. 2007 Ritual and Political Economies. In Mesoamerican Ritual Economy, edited by E. Christian Wells and Karla L. Davis-Salazar, pp. 287–300. University of Colorado Press, Boulder. Spock, Carolyn 1977 An Analysis of the Architectural and Related Features at the George C. Davis Site. Master’s thesis, Department of Anthropology, University of Texas at Austin. Stahle, David W., and Malcolm K. Cleaveland 1995 Texas Paleoclimatic Data from Daily to Millennial Time Scales. In The Changing

488

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Climate of Texas: Predictability and Implications for the Future, edited by James Norwine, James R. Giardino, Gerald R. North, and Juan B. Valdes, pp. 49–69. GeoBooks, College of Geosciences and Maritime Studies, Texas a&m University, College Station. Story, Dee Ann 1972 A Preliminary Report of the 1968, 1969 and 1970 Excavations at the George C. Davis Site, Cherokee County, Texas. Report of Field Research Conducted under National Science Foundations Grants (GS-2573 and 3200) and Interagency Contracts between the University of Texas at Austin and the Texas Building Commission and Texas Historical Survey Committee. Texas Archeological Research Laboratory, University of Texas at Austin. 1990 Cultural History of the Native Americans. In The Archeology and Bioarcheology of the Gulf Coastal Plain, by Dee Ann Story, Janice A. Guy, Barbara A. Burnett, Martha D. Freeman, Jerome C. Rose, D. Gentry Steele, Ben W. Olive, and Karl J. Reinhard, pp. 163–366. 2 Vols. Research Series No. 38. Arkansas Archeological Survey, Fayetteville. 1997 Archeological Investigations at the George C. Davis Site, Cherokee County, Texas. Bulletin of the Texas Archeological Society 68:1–113. 1998 The George C. Davis Site Mounds, Structures, and Burials: Glimpses into Early Caddoan Symbolism and Ideology. In The Native History of the Caddo: Their Place in Southeastern Archaeology and Ethnohistory, edited by Timothy K. Perttula and James E. Bruseth, pp. 9–43. Studies in Archeology 30. Texas Archeological Research Laboratory, University of Texas at Austin. 2000 Introduction. In The George C. Davis Site, Cherokee County, Texas, by H. Perry Newell and Alex D. Krieger, pp. 1–31. 2nd ed. Society for American Archaeology, Washington dc. Story, Dee Ann (editor) 1981 Archaeological Investigations at the George C. Davis Site, Cherokee County, Texas: Summers of 1979 and 1980. Occasional Papers 1. Texas Archeological Research Laboratory, University of Texas at Austin. Story, Dee Ann, Janice A. Guy, Barbara A. Burnett, Martha D. Freeman, Jerome C. Rose, D. Gentry Steele, Ben W. Olive, and Karl J. Reinhard 1990 The Archeology and Bioarchaeology of the Gulf Coastal Plain. 2 Vols. Research Series No. 38. Arkansas Archeological Survey, Fayetteville. Struever, Stuart 1968 Flotation Techniques for the Recovery of Small-Scale Archaeological Remains. American Antiquity 33(3):353–362. Stuiver, Minze, and Paula J. Reimer 1993 Extended 14C Data Base and Revised CALIB 3.0 14C Radiocarbon Calibration Program. Radiocarbon 35(1):215–230. Stuiver, Minze, Paula J. Reimer, Edouard Bard, J. Warren Beck, G. S. Burr, Konrad A. Hughen, Bernd Kromer, Gerry McCormac, Johannes van der Plicht, and Marco Spurk 1998 intcal98 Radiocarbon Age Calibration, 24,000–0 cal BP. Radiocarbon 40(3):1041–1083. Stuiver, Minze, Paula J. Reimer, and R. W. Reimer 2005 calib Radiocarbon Calibration. Internet resource http://calib.qub.ac.uk/calib/.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

489

Copyright © 2012. Nebraska Paperback. All rights reserved.

Styles, Bonnie W., and James R. Purdue 1984 Faunal Exploitation at the Cedar Grove Site. In Cedar Grove: An Interdisciplinary Investigation of a Late Caddo Farmstead in the Red River Valley, edited by Neal L Trubowitz, pp. 211–226. Research Series No. 23. Arkansas Archeological Survey, Fayetteville. Suhm, Dee Ann, and Edward B. Jelks (editors) 1962 Handbook of Texas Archeology: Type Descriptions. Texas Archeological Society, Special Publication No. 1, and Texas Memorial Museum, Bulletin No. 4, Austin tx. Suhm, Dee Ann, Alex D. Krieger, and Edward B. Jelks 1954 An Introductory Handbook of Texas Archeology. Bulletin of the Texas Archeological and Paleontological Society 25:1–562. Sullivan, Lynne P., and Robert C. Mainfort Jr. 2010 Mississippian Mortuary Practices and the Quest for Interpretation. In Mississippian Mortuary Practices: Beyond Hierarchy and the Representationist Perspective, edited by Lynne P. Sullivan and Robert C. Mainfort Jr., pp. 1–13. University Press of Florida, Gainesville. Sundermeyer, Scott, Timothy G. Baugh, Jane Whaley, and Myra L. McMinn 2004 Cultural Resource Investigations of 11 Archaeological Sites at Broken Bow Reservoir, McCurtain County, Oklahoma. Report of Investigations No. 5. LopezGarcia Group, Dallas. Swanton, John R. 1942 Source Material on the History and Ethnology of the Caddo Indians. Bulletin 132. Bureau of American Ethnology, Smithsonian Institution, Washington dc. Tamers, Murray A., F. J. Pearson Jr., and E. Mott Davis 1964 University of Texas Radiocarbon Dates II. Radiocarbon 6:138–159. Tanner, Helen 1993 The Caddos in the Era of the Republic of Texas. 3 parts. Caddo Nation News 2 (2–4). Texas Beyond History 2001 Upper Nasoni. http://www.texasbeyondhistory.net/nasoni/index.html. Thoburn, Joseph B. 1931 The Prehistoric Cultures of Oklahoma. In Archaeology of the Arkansas River Valley, by Warren King Moorehead, pp. 53–150, Yale University Press, New Haven. Thoburn, Joseph B., and Muriel H. Wright 1929 Oklahoma, A History of the State and its People. 6 volumes. Lewis Historical Publishing, New York. Thomas, Cyrus 1894 Report on the Mound Explorations of the Bureau of Ethnology. 12th Annual Report of the Bureau of Ethnology, Washington dc. Thomas, Prentice M., L. Janice Campbell, and Steven R. Ahler 1980 The Hanna Site: An Alto Focus Village in Red River Parish. Louisiana Archaeology 5:1–381. Thurmond, J. Peter 1990 Archeology of the Cypress Creek Drainage Basin, Northeastern Texas and Northwestern Louisiana. Studies in Archeology 5. Texas Archeological Research Laboratory, University of Texas at Austin.

490

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Tieszen, Larry L., and T. Fagre 1993 Effects of Diet Quality and Composition on the Isotopic Composition of Respiratory CO2, Bone Collagen, Bioapatite, and Soft Tissues. In Prehistoric Human Bone: Archaeology at the Molecular Level, edited by Joseph B. Lambert and Gisela Grupe, pp. 121–155. Springer-Verlag, New York. Tiller, Jim 2008 Was Timber Hill the Last Caddo Village in the Caddo Homeland? Caddo Archeology Journal 18:11–21. Tilley, Christopher 1994 A Phenomenology of Landscape: Places, Paths, and Monuments. Berg Publishing, Oxford. Tiné, Angela, and Larry Tieszen 1994 Bioarchaeology. In The McLelland and Joe Clark Sites: Protohistoric and Historic Caddoan Farmsteads in Southern Bossier Parish, Louisiana, edited by David B. Kelley, pp. 213–238. Report submitted by Coastal Environments, to U.S. Army Corps of Engineers, Vicksburg District, Contract No. DACW38–91-D-0014. 1997 Bioarcheology. In Two Caddoan Farmsteads in the Red River Valley: The Archeology of the McLelland and Joe Clark Sites, edited by David B. Kelley, pp. 121–137. Research Series 51. Arkansas Archeological Survey, Fayetteville. Trigger, Bruce G. 1980 Archaeology and the Image of the American Indian. American Antiquity 45:662–676. 1985 Natives and Newcomers: Canada’s “Heroic Age” Reconsidered. McGill-Queen’s University Press, Montreal. Triska, Frank J. 2008 Ecology and History of the Red River Raft. In Freeman and Custis Red River Expedition of 1806: Two Hundred Years Later, edited by Laurence M. Hardy, pp. 307–324. Bulletin of the Museum of Life Sciences, No. 14. Louisiana State University in Shreveport. Trubitt, Mary Beth 2002 Update on the Peeler Bend Canoe: An Ancient Dugout Canoe from Saline County. Field Notes (Newsletter of the Arkansas Archeological Society) 307:3–5. 2009 Burning and Burying Buildings: Exploring Variation in Caddo Architecture in Southwest Arkansas. Southeastern Archaeology 28(2):233–247. Trubowitz, Neal L. 1984 Research Goals. In Cedar Grove: An Interdisciplinary Investigation of a Late Caddo Farmstead in the Red River Valley, edited by Neal L. Trubowitz, pp. 8–16. Research Series No. 23. Arkansas Archeological Survey, Fayetteville. Trubowitz, Neal L. (editor) 1984 Cedar Grove: An Interdisciplinary Investigation of a Late Caddo Farmstead in the Red River Valley. Research Series No. 23. Arkansas Archeological Survey, Fayetteville. Tryon, Robert C. 1939 Cluster Analysis. Edwards Bothers, Ann Arbor. Tunnell, Curtis D. 1959 The Sam Roberts Site, Ferrell’s Bridge Reservoir, Texas. Manuscript on file, Texas Archeological Research Laboratory, University of Texas at Austin.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

491

Copyright © 2012. Nebraska Paperback. All rights reserved.

Turner, Christy G., II 1979 Dental Anthropological Indications of Agriculture Among the Jomon People of Central Japan. American Journal of Physical Anthropology 48:101–106. Turner, Mike 1993 A Two-Phase Caddo Mound at the Camp Joy Site (41UR144). Notes on Northeast Texas Archaeology, No. 2:66–75. Turner, Robert L. 1978 The Tuck Carpenter site and its Relation to Other Sites within the Titus Focus. Bulletin of the Texas Archeological Society 49:1–110. 1992 Prehistoric Mortuary Remains at the Tuck Carpenter Site, Camp County, Texas. Studies in Archeology No. 10. Texas Archeological Research Laboratory, University of Texas at Austin. Turner, Robert L., and James E. Smith III, with contributions by Timothy K. Perttula, Bo Nelson, Mark Walters, and Bobby Gonzalez 2002 The Harold Williams Site (41CP10) and the Texas Archeological Society Field School of 1967. Bulletin of the Texas Archeological Society 73:1–68. Valastro, Sam, Jr., and E. Mott Davis 1970 University of Texas at Austin Radiocarbon Dates VIII. Radiocarbon 12(2):617– 639. van der Veen, Marijke 2003 When Is Food a Luxury? World Archaeology 34:405–427. Vehik, Susan C. 2002 Conflict, Trade, and Political Development on the Southern Plains. American Antiquity 67(1):37–64. Vogel, Gregory 2005a Cavanaugh: A Late Prehistoric Platform Mound in Western Arkansas. Caddoan Archeology Journal 14:35–63. 2005b A View from the Bottomlands: Late Prehistoric Mound Centers and Their Relationship to the Landscape in the Northern Caddo Area. Ph.D. dissertation, Department of Anthropology, University of Arkansas, Fayetteville. Vogel, Gregory, Marvin Kay, and Louis Vogele Jr. 2005 A Platform Mound at the Norman Site (34WG2), Eastern Oklahoma. Southeastern Archaeology 24:28–45. Walker, Chester P. 2009 Landscape Archaeogeophysics: A Study of Magnetometer Surveys from Etowah (9BW1), The George C. Davis Site (41CE19), and the Hill Farm Site (41BW169). Ph.D. dissertation, Department of Anthropology, University of Texas at Austin. Walker, Chester P., and Timothy K. Perttula 2008 Geophysical Investigations on Caddo Sites in East Texas and Surrounding States. Bulletin of the Texas Archeological Society 79:159–176. Walker, Chester P., and T. Clay Schultz 2008 Magnetometer Survey and Results. In Integrated Cultural Resources Investigations for the Bowie County Levee Realignment Project, Bowie County, Texas, and Little River County, Arkansas, by Scott A. Sundermeyer, John T. Penman, and Timothy K. Perttula, pp. 167–183. Miscellaneous Reports, Report of Investigations No. 29. LopezGarcia Group, Dallas.

492

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Walters, Mark 2001 Archaeological Investigations and Oxidizable Carbon Ratio Dates from 41RK476, Rusk County, Texas. Journal of Northeast Texas Archaeology 15:9–27. 2003 The Wolf Site (41SM195), Smith County, Texas. Journal of Northeast Texas Archaeology 18:1–21. Walters, Mark, and Patti Haskins 2000 The Bryan Hardy Site (41SM55), Smith County, Texas. Journal of Northeast Texas Archaeology 12:1–26. Walters, Mark, and Patti Haskins, with contributions by David H. Jurney, S. Eileen Goldborer, and Timothy K. Perttula 1998 Archaeological Investigations at the Redwine Site (41SM193), Smith County, Texas. Journal of Northeast Texas Archaeology 11:1–38. Walthall, John A. 1981 Galena and Aboriginal Trade in Eastern North America. Scientific Papers 17. Illinois State Museum, Springfield. Waring, Antonio J., Jr., and Preston Holder 1945 A Prehistoric Ceremonial Complex in the Southeastern United States. American Anthropologist 47:1–34. Watson, Virginia 1947 A Report of Archaeological Reconnaissance in the Wister Reservoir Area, Le Flore County, Oklahoma. Appendix A in Preliminary Report on Archaeological Activities Conducted by the Department of Anthropology in the Wister Reservoir Area in the Summer of 1947. Report on file at Oklahoma Archeological Survey, Norman. Webb, Clarence H. n.d. Notes on file at Northwestern State University of Louisiana, Natchitoches. 1945 Second Historic Caddo Site at Natchitoches. Bulletin of the Texas Archeological and Paleontological Society 16:52–83. 1948 Caddoan Prehistory: The Bossier Focus. Bulletin of the Texas Archeological and Paleontological Society 19:100–143. 1959 The Belcher Mound: A Stratified Caddoan Site in Caddo Parish, Louisiana. Memoirs No. 16. Society for American Archaeology, Salt Lake City. 1963 The Smithport Landing Site: An Alto Focus Component in DeSoto Parish, Louisiana. Bulletin of the Texas Archeological Society 34:143–187. 1983 The Bossier Focus Revisited: Montgomery I, Werner, and Other Unicomponent Sites. In Southeastern Natives and Their Pasts, Papers Honoring Dr. Robert E. Bell, edited by Don G. Wyckoff and Jack L. Hofman, pp. 183–240. Studies in Oklahoma’s Past No. 11, Oklahoma Archeological Survey, and Contribution No. 2, Cross Timbers Heritage Association, Norman. 1984 The Bellevue Focus: A Marksville-Troyville Manifestation in Northwestern Louisiana. Louisiana Archaeology 9:251–274. Webb, Clarence H., and Monroe Dodd Jr. 1939 Further Excavations of the Gahagan Mound: Connections with a Florida Culture. Bulletin of the Texas Archeological and Paleontological Society 11:92–126. Webb, Clarence H., and Hiram F. Gregory 1986 The Caddo Indians of Louisiana, Second Edition. Anthropological Study No. 2. Department of Culture, Recreation, and Tourism, Louisiana Archaeological Survey and Antiquities Commission, Baton Rouge.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

493

Copyright © 2012. Nebraska Paperback. All rights reserved.

Webb, Clarence H., and Ralph McKinney. 1975 Mounds Planation (16CD12), Caddo Parish, Louisiana. Louisiana Archaeology 2:39–128. Weber, David J. 1992 The Spanish Frontier in North America. Yale University Press, New Haven. Weber, J. Cynthia 1972 Ceramics of the East Mounds (3CL21). Manuscript on file, Arkansas Archeological Survey, Fayetteville. 1973 The Hays Mound, 3CL6, Clark County, South Central Arkansas. Report on research conducted under cooperative agreement between the National Park Service Southeast Region and the Arkansas Archeological Survey. Manuscript on file, Arkansas Archeological Survey, Arkadelphia and Fayetteville. Weber, J. Cynthia, and Stephen G. Loring 1971 A Study of Friendship Engraved Vessels in the Henderson State College Museum. Manuscript on file, Arkansas Archeological Survey, Fayetteville. Wedel, Mildred M. 1978 La Harpe’s 1719 Post on Red River and Nearby Caddo Settlements. Bulletin 30. Texas Memorial Museum, Austin. Weinand, Daniel C., and Elizabeth Reitz 1994 Vertebrate Fauna. In The McLelland and Joe Clark Sites: Protohistoric-Historic Caddoan Farmsteads in Southern Bossier Parish, Louisiana, edited by David B. Kelley, pp. 165–186. Report prepared for the U. S. Army Corps of Engineers, Vicksburg District. Coastal Environments, Baton Rouge. 1997 Vertebrate Fauna and Freshwater Mussels. In Two Caddoan Farmsteads in the Red River Valley: The Archeology of the McLelland and Joe Clark Sites, edited by David B. Kelley, pp. 97–108. Research Series 51. Arkansas Archeological Survey, Fayetteville. Weinstein, Richard A., David B. Kelley, and Joe W. Saunders (editors) 2003 The Louisiana and Arkansas Expeditions of Clarence Bloomfield Moore. University of Alabama Press, Tuscaloosa. Weissner, Polly, and Wulf Schiefenhovel 1996 Food and the Status Quest. Berghan Books, Providence. Wheatley, David, and Mark Gillings 2000 Vision, Perception, and gis: Developing Enriched Approaches to the Study of Archaeological Visibility. In Beyond the Map: Archaeology and Spatial Technologies, edited by Gary Lock, pp. 1–27. NATO Science Series A, vol. 321, ios Press, Amsterdam. White, Nancy Marie, and Richard A. Weinstein 2008 The Mexican Connection and the Far West of the U.S. Southeast. American Antiquity 73(2):227–277. White, Patsy K. 1987 The Gordon Site: A Middle Coles Creek to Late Mississippi Period Occupation in Ashley County, Southeast Arkansas. The Arkansas Archeologist 25/26:1–51. Willey, Gordon R. 1953 Prehistoric Settlement Patterns in the Viru Valley, Peru. Bulletin 155. Bureau of American Ethnology, Smithsonian Institution, Washington dc.

494

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

1968 Settlement Archaeology: An Appraisal. In Settlement Archaeology, edited by Kwang-chih Chang, pp. 208–226. National Press Books, Palo Alto. Willey, Gordon R., and Philip Phillips 1958 Method and Theory in American Archaeology. University of Chicago Press, Chicago. Williams, G. Ishmael, Jr. 1993 Features and Dating. In Caddoan Saltmakers in the Ouachita Valley, edited by Ann M. Early, pp. 37–48. Research Series No. 43. Arkansas Archeological Survey Research Series, Fayetteville. Williams, Stephen 1964 The Aboriginal Location of the Kadohadacho and Related Tribes. In Exploration in Cultural Anthropology, edited by Ward Goodenough, pp. 545–570. Yale University Press, New Haven. Wilson, Diane 1997a Dental Paleopathology in the Sanders and Mitchell Populations. Bulletin of the Texas Archeological Society 68:147–160. 1997b Gender, Diet, Health, and Social Status in the Mississippian Powers Phase Turner Cemetery Population. In Women in Prehistory, edited by Cheryl Claassen and Rosemary A. Joyce, pp. 119–135. University of Pennsylvania Press, Philadelphia. 2003 Analysis of the Human Remains from Burials 1 and 2 from the Nawi haia ina site (41RK170). In The Nawi haia ina Site (41RK170): Archeological Investigations in the City of Henderson’s Southside Wastewater Treatment Plant, Rusk County, Texas, by Timothy K. Perttula and Bo Nelson, pp. 122–132. Report of Investigations No. 51. Archeological & Environmental Consultants, llc, Inc., Austin. 2004 Analysis of Human Remains from the 41CP71 Cemetery Area. In Archeological Investigations at the Shelby Site (41CP71) on Greasy Creek, Camp County, Texas, by Timothy K. Perttula and Bo Nelson, pp. 156–166. Special Publication No. 5. Friends of Northeast Texas Archaeology, Pittsburg and Austin. 2006 Human Remains in the Collection. In A Study of the Buddy Calvin Jones Collection from Northeast Texas Caddo Sites, by Timothy K. Perttula, pp. 110–133. Special Publication No. 6. Friends of Northeast Texas Archaeology, Pittsburg and Austin. 2008a The Human Remains from 41HS15, the Pine Tree Mound Site, Harrison County, Texas. Report on file at Prewitt and Associates, Austin. 2008b Human Remains. In Lake Naconiche Archeology, Nacogdoches County, Texas: Results of the Data Recovery Excavations at Five Prehistoric Archeological Sites, edited by Timothy K. Perttula, pp. 557–579. 2 vols. Report of Investigations No. 60. Archeological & Environmental Consultants, llc, Inc., Austin. 2011 Analysis of Human Remains from the Lang Pasture Site (41AN38), Anderson County, Texas. In Archeological Investigations at the Lang Pasture Site (41AN38) in the Upper Neches River Basin of East Texas, assembled and edited by Timothy K. Perttula, David B. Kelley, and Robert A. Ricklis, pp. 381–401. Texas Department of Transportation, Archeological Studies Program Report No. 129, Austin. Wilson, Diane, and Diane Cargill 1993 Stable Isotope Analysis from the Sanders Site (41LR2). Caddoan Archeology Newsletter 4(3):3.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

495

Copyright © 2012. Nebraska Paperback. All rights reserved.

Wilson, Diane, and D. Gentry Steele 1996 Prehistoric Human Remains from 12 Sites at U.S. Army Corps of Engineers Reservoirs in Bell, Delta, Denton, Ellis, Hill, Marion, and Navarro Counties, Texas. Technical Reports No. 23. Prewitt and Associates, Austin. 1997 Appendix D: Analysis of Human Remains from Cooper Lake. In Synthesis of the Prehistoric and Historic Archaeology of Cooper Lake, Delta and Hopkins Counties, Texas, by Ross C. Fields, Marie E. Blake, and Karl W. Kibler, pp. 177–244. Reports of Investigations No. 104. Prewitt and Associates, Austin. Wilson, Samuel M., and T. Clay Schultz, with contributions by Timothy K. Perttula 2009 2003 Investigations at Caddoan Mounds State Historic Site, the George C. Davis Site (41CE19), Cherokee County, Texas. MS in preparation, Department of Anthropology, University of Texas at Austin. Wolfman, Daniel 1984 Cedar Grove Chronometrics. In Cedar Grove: An Interdisciplinary Investigation of a Late Caddo Farmstead in the Red River Valley, edited by Neal L. Trubowitz, pp. 257–262. Research Series No. 23. Arkansas Archeological Survey, Fayetteville. Wood, W. Raymond 1963a The Crenshaw Site; A Coles Creek and Caddoan Mound Group in Miller County, Arkansas. Manuscript on file, University of Arkansas Museum, Fayetteville. 1963b A Preliminary Report on the 1962 Excavations at the Crenshaw Site, 3MI6. In Arkansas Archeology 1962: An Annual Report on Archeological Research, Investigation, Activities, and Progress in the State of Arkansas, edited by Charles R. McGimsey III, pp. 1–14. Special Publication of the Arkansas Archeological Society, Fayetteville. 1981 The Poole Site, 3GA3, with a New Foreword and Summary by Ann M. Early. Arkansas Archeologist 22:7–65. 1983 The Loftin Component, 235N42. Missouri Archaeologist 44:1–74. Wyckoff, Don G. 1963 Archaeological Survey of Pine Creek Reservoir, McCurtain County, Oklahoma. General Survey Report No. 3. Oklahoma River Basin Survey, Norman. 1965a The Biggham Creek of McCurtain County, Oklahoma. Archaeological Site Report No. 3. Oklahoma River Basin Survey, Norman. 1965b The Hughes, Lamas Branch, and Callahan Sites, McCurtain County, Oklahoma. Archaeological Site Report No. 4. Oklahoma River Basin Survey, Norman. 1967a The E. Johnston Site and Prehistory of Southeast Oklahoma. Archaeological Site Report No. 6. Oklahoma River Basin Survey, Norman. 1967b Woods Mound Group: A Prehistoric Mound Complex in McCurtain County, Oklahoma. Bulletin of the Oklahoma Anthropological Society 15:1–76. 1968a The Beaver Site and Archaeology of the Broken Bow Reservoir Areas, Southeast Oklahoma. Archaeological Site Report No. 9. Oklahoma River Basin Survey, Norman. 1968b The Bell and Gregory Sites: Archaeological Chronicles of Prehistory in the Pine Creek Reservoir Area, Southeast Oklahoma. Archaeological Site Report No. 11. Oklahoma River Basin Survey, Norman. 1970 The Archaeological Sequence in the Broken Bow Reservoir Area. Stovall Museum of Natural History, Norman. 1974 The Caddoan Cultural Area: An Archaeological Perspective. In Caddoan Indians I, pp. 25–280. Garland ny. 1980 Caddoan Adaptive Strategies in the Arkansas Basin, Eastern Oklahoma. Ph.D.

496

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

dissertation, Department of Anthropology, Washington State University, Pullman. 2008 Intertwined Legacies: The Wichita and Dr. Robert E. Bell. In “Land of Our Ancestors: Studies in Protohistoric and Historic Wichita Cultures,” edited by Timothy G. Baugh and Stephen M. Perkins, pp. 377–380. Memoir 40. Plains Anthropologist 53(208). Wyckoff, Don G., and Thomas P. Barr 1967 The Cat Smith Site: A Late Prehistoric Village in Muskogee County, Oklahoma. Bulletin of the Oklahoma Anthropological Society 15:81–106. Wyckoff, Don G., and Timothy G. Baugh 1980 Early Historic Hasinai Elites: A Model for the Material Culture of Governing Elites. Midcontinental Journal of Archaeology 5:225–288. Wyckoff, Don G., and Linda Ragland Fisher 1985 Preliminary Testing and Evaluation of the Grobin Davis Archeological Site, 34Mc-253, McCurtain County, Oklahoma. Archeological Resource Survey Report No. 22. Oklahoma Archeological Survey, Norman. Yaeger, Jason, and Marcello A. Canuto 2000 Introducing an Archaeology of Communities. In The Archaeology of Communities: A New World Perspective, edited by Marcello A. Canuto and Jason Yaeger, pp. 1–15. Routledge, New York. Yellen, John E. 1977 Cultural Patterning in Faunal Remains: Evidence from the !Kung Bushman. In Experimental Archaeology, edited by Dan Ingersoll, John E. Yellen, and William MacDonald, pp. 271–331. Columbia University Press, New York. Young, Biloine Whiting, and Melvin L. Fowler 2000 Cahokia, The Great Native American Metropolis. University of Illinois Press, Urbana. Young, Gloria A., and Michael P. Hoffman (editors) 1993 The Expedition of Hernando de Soto West of the Mississippi, 1541–1543: Proceedings of the De Soto Symposia, 1988 and 1990. University of Arkansas Press, Fayetteville.

References Cited

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

497

Copyright © 2012. Nebraska Paperback. All rights reserved. The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Contributors Robert L. Brooks, Oklahoma Archeological Survey, Norman James A. Brown, Northwestern University, Evanston Ann M. Early, Arkansas Archeological Survey, Fayetteville Jeffrey S. Girard, Northwestern State University of Louisiana, Natchitoches H. Edwin Jackson, University of Southern Mississippi, Hattiesburg Jami J. Lockhart, Arkansas Archeological Survey, Fayetteville David B. Kelley, Coastal Environments, Inc., Baton Rouge Duncan P. McKinnon, Department of Anthropology, University of Arkansas, Fayetteville Timothy K. Perttula, Archeological & Environmental Consultants, LLC, Austin Robert Rogers, PBS&J, Austin George Sabo III, Arkansas Archeological Survey, Fayetteville Frank F. Schambach, Arkansas Archeological Survey (emeritus)

Copyright © 2012. Nebraska Paperback. All rights reserved.

Susan L. Scott, University of Southern Mississippi, Hattiesburg Mary Beth Trubitt, Arkansas Archeological Survey, Henderson State University Research Station, Arkadelphia Gregory Vogel, Center for American Archeology, Kampsville, Illinois Chester P. Walker, Archaeo-Geophysical Associates, LLC, Austin Diane Wilson, A. M. Wilson Associates, Barnstable, Massachusetts

499

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved. The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Index

Copyright © 2012. Nebraska Paperback. All rights reserved.

Figures and tables are indicated by the letters F and T Adai language, 2 agriculture, 345, 363; droughts and, 236–237; in Early Caddo period, 345; maize, 5, 8, 87–88, 234–235, 427–428; rituals around, 441–443; and sedentary life, 7–8, 36, 314; and social complexity, 86; and soil, 352; in Titus phase, 366, 408, 427–428; and tropical cultigens, 7, 8, 19 A. H. Reagor Farm site, T4-2, T4-3 Alba arrow points, 423, 430n1 Albertson, Paul E., 245, 252 Alex Justiss site, 374, 402; stable isotope data from, 106, 115, T4-3, T4-6 Allison site, 249–250 Alto focus, 243, 254 Ambrose, Stanley H., 97, 103 American Bottom, 47, 109 Angelina River, 115, 236, 445, T4-6 apatite, 107, 108, 114, 115 appliqué ridges, 30, 255–256, 271, 273, 305–306, 310 Arcgis software, 318 archaeogeophysics: and George C. Davis site investigations, 181–189, F7-1, F7-2; and Hill Farm site investigations, 191–195, F7-7, F7-8; impact on Caddo archaeology of, 178–179, 181, 206–207; as investigative tool, 16–17, 208 architecture and architectural features: in Belcher phase, 415–417, F14-2; central hearth, 180, 183, 185, 193, 216, 218; circular structures, 213, 216, 228, 231, F8-4, F8-5, F8-6, F8-7, T8-1; extended entranceways, 192, 194, 213–214, 216,

231, 371, 415, 417; geometric shape, 179; geophysical, 183–186; in Oak Hill Village site, 212–218, 228, 231, F8-2, F8-3, F8-4, F8-5, F8-6, F8-7, T8-1; post hole patterns, 56, 203, 220; rectangular structures, 213, 214–216, 226, 231, F84; sub-square structures, 186; wooden structures, 213 Arikara language, 2 Arkansas Archeological Society, 313 Arkansas Archeological Survey, 196, 289, 291, 292, 295, 316, 414 Arkansas River basin, 16, 17, 25n3, 345–346; Region I, 165–169; Region II, 161 Arms Creek, 371, 393 arrow points, 9; Alba and Bonham, 423, 430n1; Bassett, 308, 309; in Belcher phase, 423–424; as funerary offerings, 396–397, 400, 401–402; Gary, 250–251; at Hughes site, 307–308, 311, F10-10; Nodena, 423–424; Perdiz, 308, 380; Talco, 308, 397, 400; Washita, 308 Ashdown Engraved, 308 Ashmore, Wendy, 336, 388 Atkins-Pope (ap) soil association, 353, 354 Automated Management of Archeological Site Data in Arkansas (amasda), 316, 329, 331, 333 Avery Engraved, 44, 418–419 axis mundi, 122, 127–128 Ayo-Caddi-Aymay, 22, 440, 443, 444 Baber, Quin, 292 Bailey Engraved, 305, 306 Baker, Louis, 247, 259–260 Baldwin site, 346

501

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Barnes, Mark R., 381 Barr, Thomas, 342 Barroto/Romero expedition, 431 basal notched arrow points, 308 Bassett arrow points, 308, 309 Battle Mound site, 60, 413, 414, 430n2, F14; about, 195–196; archaeogeophysical data from, 196–197, 207, F7-10; circular structures at, 180, 196, 198, 201, 203, F7-14, F7-15; community cemetery at, 206, F7-17; community organization at, 198–206, F7-11; concentric circular patterns at, 203, F7-16; early study of, xv, 196; farmsteads at, 198, F7-13 Bayou Sel site, 36 beans, 427–428 Beatty, William C., 191 Beaver site, 342, 345, 346 Beckman, Mike, 246 Beech Ridge site, T4-2, T4-3 Beene Plantation site, 420, 421, 430 Belcher Engraved, 307, 309, 417, F14-3; frequency of, 420, 421, 423, T14-1; var. Manchester, 42; var. Ogden, 418, 419 Belcher phase: about, 21–22, 411; architecture in, 415–417, F14-2; bone artifacts in, 424–425; cemeteries in, 412–413, 414; ceramics in, 254, 303, 417–423, F14-3; chronology of, 411–412; settlements in, 412–415, F14-1; shell artifacts in, 425, 427, 430n2, F14-4; stone artifacts in, 308, 423–424; subsistence and diet in, 427–429; tribal identity in, 429–430 Belcher Ridged, 254, 256, 284–285, 417, 418–419, F14-3; frequency of, 420, 421, T14-1 Belcher site, xv, 21, 201, 243, 423, F7-12 beliefs and myths, 439–447; on animals, 49; on birds, 82–83; on cultivated plants, 19; on direction and orientation, 192, 441–442, 446, 447; and feasting, 112; on life and death, 192, 407, 445; origin story, 439–440; Terán map and, 18, 431, 440–441. See also ritual and ceremony

502

Bell, Robert E., 120, 342 Bell site, 345 Bennett, Jack, 343 Bentsen-Clark site, T4-2 Bienville, Jean-Baptiste Le Moyne de, 254, 430 Big Cypress Creek basin, 21, 382, 385– 386; abandonment of, 409; cemeteries in, 393, 396, 407, 408; floodplain of, 371; mound building in, 384; settlements in, 367, 369, 371, 372–373; social complexity in, 363 Biggham Creek site, 342, 345, 346 Big Oaks site, 396 Bill Hughes site, 345 birds, 68–69, 82–83 bison, 8–9, 135 Blackland Prairies biotic region, 91, 93, 95, 109, T4-1, T4-2, T4-5 Blakely Engraved, 309 Blitz, John H., 4 Bob Williams site, 60 Bogan, Arthur, 81 Boggy Creek, 393 Bohannon site, 346 Bolton, Herbert E., 433 bone artifacts, 424–425 Bonham arrow points, 423, 430n1 borderland regions, 289, 310–312 Borrow Pit site, 165 Bossier Brushed, 255–256, 284 Bossier focus, 240, 243–244, 254, 285; ceramic types within, 284; spatial variation within, 284, 285–286 Bowie-Caddo-Boswell (bcb) soil association, 353, 354 Boyette site, T4-2, T4-3 Bradley, Raymond S., 142–143, 173, 174 Brain, Jeffrey P., 425 Brainerd-Robinson (br) coefficient of similarity, 225–226 Brazos River, 10–11 Briarfield Plain, 417 Broken Bow Lake, 340, 342 Brooks, Robert, 17, 175, 176, 335–362 Brown, James A., 22, 59, 117–138, 154, 447

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Brown Mound, 166 Brushed/Cowhide Stamped, 306 brushed pottery, 29, 225–226, 255–256, 271, 273, 303–304, 309 Brushy Creek, 373 Buikstra Jane E., 111 built environment, 17, 336–337, 349, 350; defined, 347; and elevation, 357, F12-3 Bullard Brushed, 29 burials. See cemeteries and burial sites; mortuary practices Burks site, 373 burned structures, 311, 369, 371, 383, 386; at Belcher phase sites, 412, 413; burned daub in, 415, 417; and ceremonial food preparation, 438 Burnett, Barbara A., 88, 96, 108 Button Houses, 180, 183–184, F7-3 Byrd, Kathleen M., 80 caddi, 9–10, 286n2, 435–436, 441 Caddo archaeology: and archaeogeophysics, 178–179, 181, 206–207; bibliography of, 24–25n2; bioarchaeological approach to, 19–20, 86–116; and Caddo Nation of Oklahoma, 19–20, 25n2, 54; chronology of, 1, 7, 11–13, 314, 344–346, T1-1; early study of, xv; future directions for, 20, 92, 208, 430, 446, 447; and Mississippian groups, 4–5; Oklahoma River Basin Survey and, 340, 341–343, 347; phases of, 12–13; as poorly understood, 3, 4; relic collecting by, 26–27, 291–292; Survey and Planning efforts and, 340, 343; technological framework of, 315–316; temporal-spatial systematics studied by, 11–12, 22–23, 335; and Terán map, 189, 296n2, 437–439, 446; terminology in, 1–2, 3; theoretical framework of, 314–315; traditional archaeological field methods and, 17, 179, 207–208, 209; Willow Chute Bayou studied, 246–249, F9-4; wpa and, 124, 149, 181, 191, 340–342, 437, 439 Caddo Conference, 2

Caddo Lake, 10 Caddo language, 2 Caddo Nation of Oklahoma, 11, 19–20, 24, 25n2, 54 Caddo removal, xvi, 1, 7, 10–11 “Caddo” term, 2, 5–6 Cahinnio group, 10 Cahokia site, 47, 50, 59, 97 Camp Joy Mound site, 367; dating of, 374, 377, 380; mound location at, 385, 386; mound structure at, 371, 383 canahas, 9, 440 Caney Creek valley, 373 Caney Punctated, 304, 306 Carden Bottom Cemetery site, 156, 161 Carroll Bell site, 293 Casañas de Jesús María, Francisco, 432 Cass Appliquéd, 396 Casselberry, Samuel E., 238n4 Cat Smith site, 164 Cavanaugh Mound site, 167, 171, F6-10 Cavelier, René-Robert, 431 C. D. Marsh site, 93, T4-1, T4-2 Cedar Grove site, 107, 412, 414, 438; bone and shell artifacts at, 424–425; ceramics at, 303, 417–418, 421, 423; Crenshaw site compared to, 76–80; floral and faunal assemblages at, 427–428 cemeteries and burial sites: at Battle Mound site, 206; in Belcher phase, 412–413, 414; community cemeteries, 365, 369, 371, 383, 395–396, 407; at Crenshaw site, 51, 59; at Ear Spool site, 372; family cemeteries, 400–401; at Great Mortuary, 125–127; at Hatchel site, 438; at Kuykendall Brake site, 307; and mounds, 8, 371, 373, 383, 386; at Oak Hill Village site, 232; at P. S. Cash site, 386; red clay at, 401; at Rookery Ridge site, 372; as sacred, 408; shaft graves at, 402, 403; at Shelby site, 369; size of, 206, 393, 394–395, F13-6; as social institutions, 407; social significance of, 388, 392, 407–408; in Spiro and Fort Coffee phases, 124; structure of, 242, 393–394, 396, 397–398, 401,

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

503

Copyright © 2012. Nebraska Paperback. All rights reserved.

cemeteries and burial sites (continued) 403, 405, 409, F13-7, T13-5; Terán map and, 436; in Titus phase, 365, 372, 373, 388, 393–396, F13-5, F13-6, T13-3. See also mortuary practices “center of the universe,” 121–122 ceramic decoration patterns: class A, 299, 301–302; class B, 299, 302; class C, 299, 302–303, T10-2; class D, 299, 303–304, T10-2; class E, 299, 305, F10-9, T10-2; class F, 299, 305, T10-2; class G, 299, 305, T10-2; class H, 299, 305–306, T10-2 ceramic design and decoration, 14; background embellishments, 38–39; bracketed, 37–38; Caddo principles of, 43–44; changes in, 29–30, 234; choices in, 34–40, F2-3, F2-4, F2-5, F2-6, F2-7, F2-8; conjoined oval element, 38, 39, 40; dimensionality in, 34, 46; and exterior surfaces, 299, 301–302, F10-7, T10-2; for fine ware, 28, 29, 419, 420, 421; finishing, 36–37, 39, F2-5; floating oval element, 37, F2-6; framing, 36; fused vessels and, 43–44; oval-withinan oval, 40; previews design, 28–29; scroll choices in, 38, 39, 40; stacking fields in, 43; symmetry and opposition in, 44; two- or three-panels, 36, F2-4; for utilitarian vessels, 27–28, 256, 279, 305, 306 ceramics: at Allison site, 249–250; in Belcher phase, 254, 303, 417–423, F14-3; in Bossier focus, 283–285; and Caddo culture, 7, 9, 26–27, 44–46; at Cedar Grove site, 303, 417–418, 421, 423; ceramic pipe, 423; clay sourcing of, 27; course utility wares, 419, 420, 421; and cultural interaction, 310–311; domestic assemblage of, 420; at Festervan site, 251, 253; as funerary objects, 396–397, 399–400, 403, 407, 417–418; geographic distribution of, 27; at Hughes site, 297–307; inter-site similarities of, 233; at Jim Burt site, 250; at Joe Clark site, 420; in Late

504

Caddo period, 40, 304; at McLelland site, 420, 423; in Middle Caddo period, 30, 256, 271, 304; at Oak Hill Village site, 225; at Rooker Ridge site, 380; technological sophistication of, 344; in Titus phase, 308, 364–365, 380, 405; trade in, 289, 420; at Vanceville site, 271, 279–282, T9-2; variation in, 14–15, 239 ceramic types, 29–30, F10-7; brushed and appliqué ridged, 29–30, 225–226, 255–256, 271, 273, 303–306, 309–310; engraved, 28, 29, 44, 258, 273, 279–83, 305, 419, 420, 421; as ethnic indicator, 288–289; incised, 28, 29, 256–257, 271, 283–284; Phillips’ type variety system of, 32; punctated, 29, 257–258, 271, 284, 302–303; ridged, 254, 256, 283, 284–285, 417, 418–419, 420, 421, F14-3; stamped, 27, 29, 258, 291, 305; undecorated, 251, 258–259, 271, 273, 279; Webb classification system of, 417, 419, 420 ceremonialism, 61, 234, 243, 331, 361–362, 365, 383. See also ritual and ceremony Chakanina phase, 415 Chastain site, 367, 371, 385, 386 Chesson, Meredith S., 392 Chickasaw, 70 chiefs, 4, 49–50. See also caddi chipped stone, 307–310, 423, 424, F10-10 Choctaw County, 337, 339, 346–347, F12-1 circular structures, 180, 185–186; at Battle Mound site, 198, 200, 203, F7-14, F715; at Hill Farm, 192–193; at Rookery Ridge site, 372 Claiborne group, 323, 324 clay, 401; and soils, 56, 339, 352 Clement mound center, 341, 345, 351, 354, 360, 361 Clements, Forrest, 341 Clements site, 400, 425 climate, 123, 211, 236–237, 363 cluster analysis, 320–321, 384, F11-1

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Coker Mound site, T4-3 Coles Creek Incised, 251, 257 collegiate system, 299 community: archaeological, 209; political, 367–373, 385–386, 398–399, 405, 408–409, F13-2 community organization: at Battle Mound site, 198–206, F7-11; in Caddo society, 8, 209, 345, 393; at George C. Davis site, 186–189, F7-4, F7-5; at Hill Farm site, 194–195 compound fences, 196, 198, 201–202, 203 concentric circles, 180, 203, 206, 302, 445, F7-16 conceptualized landscape, 336, 337 Conly site, 249 Cook, Sherburne F., 238n4 Cook Engraved, 36, 305, 306 Cook site, 341, 345 copper plates, 130 corn. See maize correspondence analysis, 272–273, 278–279; frequency seriation in, 273, F9-10; scatter plot for, 278–279, F9-9; sherd categories in, 278–283, T9-3; summary of, T9-4 cosmology, 128–129, 439–447 Cowhide Bayou, 253 Cowhide Stamped, 258, 288, 303, 305, 417 Cracker Road Incised, 420 Crenshaw site, xv, 80; about, 50–52, F3-1, F3-2; burials at, 51; faunal assemblage at, 48, 61–62, 67, T3-1; House of the Priest at, 20, 52–58; and inter-site comparisons, 75–80, F3-6, T3-4; and maize consumption, 105, 106 Crockett Curvilinear Incised, 29, 257 Culpepper site, 403 cultigens, 7, 8, 19, 427–428. See also agriculture cultural buffers, 329–330 cultural interaction, 310–312 cultural resources management, 23 Cunliffe site, 386 curvilinear incising, 256, 257

Cypress Creek, 96, T4-6 Dalton site, 367, 371, 383, 385, 386 Dan Holdeman site, 106, T4-2, T4-3 data mining, 320 Davenport, William, 292 Deaton site, 342 deer, 70, 84; antlers, 20, 54–55, 58–59, 61; ceremonialism around, 58–59, 61, 72–73; hides, 9; meat, 8, 71–72, 428, T3-2, T3-5 Dehahuit, 10 De León, Alfonso, 431–432 dental data, 89; bone chemistry, 97–98, 112–113, T4-3; caries, 93–97, 111, 112, 113, T4-2; from Crenshaw site, 57, 60–61; dental attrition, 89–92, T4-1; from Hughes site, 295 De Siard Incised, 420 de Soto, Hernando, 7, 291, 311 diagonal incised, 255, 256, 271, 273, 301 diet: beans in, 427–428; in Belcher phase, 427–429; in Early Caddo period, 96; fish in, 69, 428; in Historic Caddo period, 108; in Late Caddo period, 95, 103–104, 106–107, 108, 113; maize in, 83, 86, 87–88, 93, 95, 96, 97–98, 429; meat in, 8–9, 49, 71–72, 428; in Middle Caddo period, 95, 103, 107; nuts in, 428; in Titus phase, 106, 366–367. See also food di-polar signatures, 193, 194 directionality, 16, 192, 442–443; and burials, 127–128, 396, 397 Douay, Anastasius, 429 Dry Creek, 373, 393 dry paste incising, 29 Dunbar, Joseph B., 252 Dunkin Incised, 251, 256, 417 Dunnell, Robert C., 13, 273 Durant-San Saba-Tarrant (dst) soil association, 353–354 Dye, David H., 59 Early, Ann M., 14, 26–46, 54, 320; on archaeological research, 314; on

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

505

Copyright © 2012. Nebraska Paperback. All rights reserved.

Early, Ann M. (continued) ceremonial mounds, 381; class-patterndesign system of, 417; and Hughes site, 288, 291, 293 Early Caddo period, 91, 242, 254, T1-1; agriculture in, 345; ceramics in, 257, 271; maize consumption in, 96, 103, 106, 109, 113 Ear Spool site, 96, 372, 374, 380 Eastwood Bayou site, 307, 309 Ebarb Incised, 419 echelon centers, 142 Ed Johnson site, 246, 247 effigy vessels, 28 E. Johnson site, 342, 345, 346 elevation: and built environment, 357, F12-3; Caddo mounds and, 322; and viewsheds, 159–160, F6-7, F6-8, F6-9 Eliade, Mircea, 122 elite, 7, 49–50, 409; burial of, 242, 403, 405–407, 413, F13-8; faunal assemblages of, 48–50; greater access to resources of, 47, 67, 73, 111; midden deposits of, 81–83; and mound use, 381–388. See also hierarchy Emma Owens Farm site, T4-1, T4-2, T4-3 Emory Punctated-Incised, 419 Ender-Hector Conway (ech) soil association, 353, 354 engraving, 28, 44, 273, 279–283, 305; as decorative technique, 258; as fine ware, 28, 29, 419, 420, 421 environmental similarity model, 18, 326–327, 331, 333, T11-5; performance, 327–329, F11-2 epidemic diseases, 10, 381, 409 E. S. Dooley Farm site, 386 Espinosa, Isidro Felix de, 61, 394, 444–445 Etowah site, 118 Europeans, xvi, 3; Caddo trade with, 9, 10; and epidemic diseases, 381, 409 E. W. Hackney site, T4-1, T4-2, T4-3 E. W. Henry site, T4-1, T4-2, T4-3 Ewing Chapel Cemetery, 171, F6-12 extended entranceways, 192, 194, 213–214,

506

216, 231, 371; in Belcher phase, 415, 417; directionality of, 441–442 Fagre, T., 103 Farley, Barbara, 58–59 Farmer, Mary V., 307 Fatherland Incised, 420 fauna, 8–9, 20, 81, 352 faunal assemblages, 295, 339, 437–438; at Belcher phase sites, 427–428; elite, 48–50; at House of the Priest, 54–58, 61–62, 67–73, T3-1, T3-3; inter-site comparisons about, 75–80, F3-6, T3-4; religious concepts behind, 446–447; social and cultural factors in, 48, 83–84 feasting, 47, 311; and maize, 87, 112; as public ritual, 48–49, 388 Felsenthal region, 289, 330 Ferguson site, 293 Festervan site, 251, 253 Finn Bayou, 245 fire, 57 fish, 69, 428, F3-4 floral assemblages, 427–428 “Flotation Techniques for the Recovery of Small-Scale Archaeological Remains” (Struever), 62 Fontcuberta, Miguel de, 432 food: cultural messages in, 83, 112; differential access to, 47, 67, 73, 83–84, 111; preparation and cooking, 71, 437–438; and ritual, 84, 438; symbolism of, 84, 86, 112, 116. See also diet Foreman, Grant, 172 Forest Mound site, 93, T4-1, T4-2, T4-3 Formative Caddo period, 91; maize consumption in, 103, 104, 105, 107, 113; technological and cultural development in, 344–345 Fort Coffee phase, 124 Fort Davis site, 153–154, 164, 169, 171–172, F6-11 Foster site, 413, 425, 430 Foster Trailed-Incised, 257, 288, 302, 303, 306, 309, 417, 418–419; as characteristic of Social Hill phase, 307–308; frequency of, 420, 421, T14-1

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Fourche Maline, 7, 50, 89, 113, 240 Fowke, Gerard, 246 Frank Benson site, 380–381 Fred McKee site, 110, T4-1, T4-2, T4-3 Freeman, Thomas, 252 Freeman-Custis expedition, 252, 253, 429–430 French Fork Incised, 257 frequency seriation, 271, 273, F9-10. See also seriation analysis Friday site, 413 Friedman, Jonathan, 119 Friendship Engraved, 14, 43, 305, 306, 310, F2-1; about, 30–32; decorative options for, 34–41; var. Freeman, 32–34, 46; var. Meador, 42, F2-8 fruits, 428 Fulton, Robert, 246 Fulton aspect, 243 funerary objects, 22, 401–403, 407, 410n2, 412, T13-4; arrow points as, 396–397, 400, 401–402; ceramics as, 308, 396–397, 399–400, 403, 407, 417–418; as cult objects, 126, 128–130, 136–137, 138n2 fused vessels, 43–44, 46, F2-9 Gahagan phase, 254 Gahagan site, xv, 59 Gaignard, Jacques, 429 Galve, Conde de, 432 Gary dart points, 250–251 gateway communities, 176 Geller, Pamela L., 388 General Allotment Act of 1887, 11 Geographic Information Systems (gis), 176, 439; about, 149; Monte Carlo randomization method of, 157, 159, T6-5; technology of, 315–316; and Tom Jones site data, 316, 318, 326, T11-1; viewshed analysis by, 143, 149–152 geophysical surveys, 23–24, 207, 438 George C. Davis site, xv, 181; about, 180– 181; archaeogeophysical investigations at, 181–189, F7-1, F7-2; architectural features at, 180, 183–186, F7-3; community organization at, 186–189, 233,

F7-4, F7-5; dental information from, 93, 96, 97, 105, 114, T4-1, T4-2, T4-3 Gettys, Marshall, 342 Ghost Dance, 118 Gibson, Jon, 284 Gibson Aspect, 243 Gillings, Mark, 148 Girard, Jeffrey S., 14–15, 239–287, 420, 421 Glassell Engraved, 417, 418–419, F14-3; frequency of, 421, 423, T14-1 Glendora site, 44 Glover River, 339–340, 349–350, 354–355, 360 Goforth-Saindon site, 143, 165, 171, 174 Goldsmith site, 373 Gold Star Ballroom site, 396 Grace Brushed, 304, 306 granaries, 218, 232, 387 Grappé, François, 436 grass software, 318 Greasy Creek, 369, 393 Great Mortuary, 4, 22, 447; cult objects in, 128–131, 136–137, 138n2; and devotional economy, 121–123, 133–136; as economic institution, 119–121; mortuary goods recovered from, 22; orientation of, 127–128; as sacred monument, 124, 125–127, 136–137; treatment of dead at, 131–133, 137, 398–399; viewed as chiefly monument, 117, 118–119, 137. See also Spiro site Great Raft, 252 Greenlee, Diana M., 104–105, 115–116 Gregory, Hiram F., 419 Gregory site, 345, 346 Griffith, William Joyce, 186 Grobin Davis Mounds site, 342, 345, 351, 360, 361 grog tempering, 304, 305, 310 Groseclose site, 125 ground stone, 424 Gulf Coastal Plain, 349, 350; natural landscape of, 337–340, F12-1; prehistoric Caddo homeland within, 330–332, F11-3

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

507

Guy Brittain site, 153

Copyright © 2012. Nebraska Paperback. All rights reserved.

Haley Complicated Incised, 30, 258 Haley phase, 414 Hally, David J., 351 Handy Place, 196, 201 Hanna site, 76–80, 106 Hardman Engraved, 305, 306, 307, 309 Hardman site, 36, 42, 43, 201 Hardy Incised, 284 Harkey Mound site, 344 Harlan site, xv, 125, 161, 169–171, F6-11 Harleton Appliqued, 29 Harmon, Anna M., 88 Harold Williams site, 367, 369, 374, T4-2, T4-3 Harrison Bayou Incised, 256 Harris, R. K., 59–60 Harroun site, 367, 371, 377, 387; mounds at, 382, 385, F13-4 Hasinai, 57, 431–432; geographic area of, 6–7, 10; rituals of, 58–59, 61, 444 Hatchel complex, 191 Hatchel site, 189, 201, 203, 240–241, 438 Hatcher, Mattie Austin, 433 Hatinu Engraved, 399–400 Hawkins Creek, 407 Hays site, 293 headdresses, 58 hearth, central, 180, 183, 185, 193, 216, 218 Helms, Mary, 49 Henderson-Southall site, 396 Henderson State University Museum, 32, 33 hierarchy, 9–10, 46, 47, 242, 314, 345. See also elite High Island site, 284 Hill Farm site, 180; about, 189, 191, F7-6; archaeogeophysical investigations at, 191–194, F7-7, F7-8; community organization at, 194–195 Historic Caddo period, 12–13, 108, T1-1 Hodges Engraved, 40, 305, 417, 418–419; frequency of, 420, 421, 423, T14-1 Hodges, Thomas L. and Charlotte, 30–32

508

Hoffman, Wanda, 291 Holder, Preston, 119 Horace Cabe site, 437 horses, 9 horticulture, 7, 8, 364. See also agriculture House, John H., 307 house construction, 23, 180, 185–186, 192–193, 198, 200, 203, 372, F7-14, F7-15 Howard, Lynn E., 196, 414 H. R. Taylor site, 367, 371, 396 Hudnall-Pirtle site, 233 Hughes site: about, 289–291; archaeological investigations at, 21, 291, 292–297, F10-5; artifact collectors at, 292; as borderland region, 310–312; burned structures at, 294–295, 311; ceramics at, 297–307; chipped stone at, 297, 307–310, 311, F10-10; landscape position of, 164; main mound at, 291, 292, 293, F10-2; midden deposits at, 291, 294, 295, 305; occupation of, 290–291; radiocarbon analyses from, 295, 297, F10-6, T10-1; topographic map of, 292–293, F10-4 Hugo Lake, 339, 342, 346 Hunter, Donald G., 423 hunting and gathering, 86, 89 Hurricane Hill site, 95, 96, 105, 106, 109, 115, T4-1, T4-3 hydrography, 316, 318, 321–322 iconography, 23, 59, 82–83 ideational landscape, 336, 337 idrisi software, 318 Iesh (José María), 10 incising, 28, 29, 271, 283–284 instrumental neutron activation analysis, 289 interpretive hypotheses, 318 interregional interaction, 345–346 Iroquois Feast of the Dead, 121 Jackson, A. T., 191 Jackson, H. Edwin, 20, 47–85, 447 Jake Martin site, 386

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Jeane, David, 54 J. E. Galt site, 403 Jelks, Edward B., 28, 29 Jim Allen site, 93, 96, 115, T4-1, T4-2, T4-3 Jim Burt site, 250 J. M. Riley site, 396 Joe Clark site, 411–412, 415, 430; ceramics found at, 420; cultigens at, 427–428; extended entranceway at, 415, 417 Joint Educational Consortium of Arkadelphia, 3 Jones, Lindsay, 122 Jordan site, 310 Joutel, Henri, 429, 436 J. W. Blackburn site, 106, T4-1, T4-2, T4-3 Kadohadacho, 59, 429, 430, 443; ceremonies and rituals of, 61, 445; geographic location of, 6–7, 10 Karnack Brushed-Incised, 304, 417, 418–419; frequency of, 421, T14-1 Kay, Marvin, 174 Kelly, David B., 21–22, 411–430 Kelly, Lucretia, 295 Kelsey Creek, 373–374 Keno Trailed, 254, 288, 302, 419; frequency of, 420, 421, 423, T14-1; var. Curtis, 45, 308, F2-9 Kent, Susan, 49 Kiamichi River, 339 Kiam Incised, 29, 251, 257, 284 King, Nicholas, 429 kinship, 9, 288, 393, 401 Kirvin-Cuthbert-Bowie (kcb) soil association, 353, 354 Kitsai language, 2 Knapp, A. Bernard, 336 Knight, Vernon J., 143 Krieger, Alex D., 11, 29, 119, 196, 271, 411, 414 Kruskal-Wallis one-way anova analysis, 91, 96 Kuykendall Brake site, 306–307, 310 Kvamme, Kenneth L., 157, 316, 328

La Flesche, Francis, 442 La Harpe, Jean-Baptiste Bénard de, 436 Lake Bob Sandlin, 373 Lake O’ the Pines, 386 Lake Providence Mounds, 50 Landsat Thematic Mapper (tm), 324 Lang Pasture site, 19, T4-1, T4-2; stable isotope data from, 107–108, 115, T4-3 languages and dialects, 2, 288 La Rue Neck Banded, 29 Late Caddo period, 91, 361, T1-1; ceramics in, 40, 304; diet in, 95, 103–104, 106–107, 108, 113; mound construction in, 346, 384; Texarkana phase, 438; at Willow Chute Bayou locality, 254–255. See also Titus phase Lawton, Sherman, 342 leadership, 9–10, 50, 409, 439–441. See also elite Lee, Christine, 89, 115 Lee Ellis Farm site, T4-3 Le Moyne, Jean Baptiste, 430 Lewis, R. Barry, 336 Lillie Creek site, 164 Lindsay Park site, T4-3 linear punctated, 257–258, 273 Little Cypress Creek, 372–373, 405 Little Missouri River, 30 Little River, 339, 349, 350, 360 Location of High Devotional Expression (lhde), 121, 123 Lock, Gary, 148 Lockhart, Jami J., 18, 176, 313–334, 350, 354, 438 Logan Eddy site, 165 looting, 292 Lost Prairie phase, 53 Lower Peach Orchard site, 367, 369 Lubbub Creek site, 81, T3-5 Lukfata Reservoir, 342 MacDonald I and II sites, 341 Mahaffrey site, 105, 342, 345 maize: agricultural economy around, 5, 8, 87–88, 234–235, 427–428; and Caddo

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

509

Copyright © 2012. Nebraska Paperback. All rights reserved.

maize (continued) cultural identity, 112; communal harvesting of, 87; consumption in Early Caddo period, 96, 103, 106, 109, 113; consumption in Formative Caddo period, 103, 104, 105, 107, 113; consumption in Middle Caddo period, 95; in diet, 86, 87–88, 93, 95–98, 429; and feasting, 87, 112; history of Caddo production of, 8, 87–88, 232; and natural environment, 88; regional differences around consumption of, 83, 89, 111; symbolism of, 19; types cultivated, 8 marriage, 9 Marston site, 283–284 Massanet, Damián, 432, 433–434 matrilineality, 9 Maud arrow points, 307, 308, 309 Maydelle Incised, 29 McClure site, 413 McCrocklin, Claude, 247 McCurtain County, 337, 339, 346–347, F12-1 McDonald site, 346 McKinney site, 346 McKinnon, Duncan, 16–17, 177–208 McLelland site, 411, 415, 430, F7-15; arrow points at, 423–424; bone artifacts at, 424–425; ceramic assemblage at, 420, 423; Crenshaw site compared to, 76–80; extended entranceways at, 415, 417; floral and faunal assemblages at, 427–428; shell artifacts at, 425 Means Engraved, 44 meat: deer, 8, 71–72, T3-2, T3-5; in diet, 428; elite’s greater access to, 47, 67, 73, 111; sources of, 8–9, 428; symbolic potential of, 49 Meddlin Creek, 371, 385, 393 midden deposits: at Crenshaw site, 56, 67; elite, 81–83; at Hatchel site, 438–439; at Hughes site, 291, 294, 295, 305; at Oak Hill Village site, 232–233; at Rookery Ridge site, 372; at Sam Roberts site, 386; at Shelby site, 369; at Spiro site, 123–124; in Titus phase, 369, 372, 373;

510

at Vanceville mound site, 259, 263, 271, 284; at Whelan site, 387–388 Middle Caddo period, 91, 247, 254–255, F8-12, T1-1; and Bossier focus, 285; ceramics in, 30, 256, 271, 304; diet in, 95, 103, 107; interregional interaction in, 345–346 Mid-Ouachita phase, 308, 310 Midwestern Taxonomic Method/System, 11, 12, 254, 411 Military Road Incised, 30, 302, 306, 310 Mill Creek site, 7, 283–284 Miller, John E., 303, 417, 419 Miller’s Crossing site, 106 Miller-Yahola-Teller (myt) soil association, 353 Milner, George, 111 miscellaneous incised, 256–257, 273 miscellaneous punctated, 257, 273 Mississippian period, 3–5, 114–115, F1-1 Mitchell, Sharon A., 397 Mitchell site, T4-1, T4-2 Mockingbird site, 374, 380, 403 Monte Carlo randomization method, 157, 159, 173, T6-5 Montgomery site (la), 283–284 Montgomer site (ok), 342 Moore, Clarence B., 9; and Battle Mound site, 196, 203; and Belcher phase sites, 413–414, 423, 425, 427; and Crenshaw site, 50, 60 Moore 6 site, 342 Moore/Higginbotham site, 429 Moore site, 293 Morrow, Juliet E., 143 mortuary practices: arrow point offerings, 396–397, 400, 401–402; ceramic offerings, 308, 396–397, 399–400, 403, 407, 417–418; collective rites, 132, 137; cosmological principles in, 128–129, F5-2; and cult objects, 126, 128–130; multiple burials, 397–398, F13-7; orientation and directionality in, 396, 397, 407; and personhood, 131–133, 137; ritual in, 8, 365, 388, 392, 401–402; and social complexity, 119; social

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

hierarchy and, 120–121, 242, 400–401, 403, 405–407, 413, F13-8; spatial dimensions of, 388, 392, 407; at Spiro site, 125–127; symbolism in, 126–127, 407–408; wood litters in, 398–399. See also cemeteries and burial sites Moscoso Alvarado, Luis de, 7 Mosquito Hollow site, 386 Mossy Grove culture group, 7 mound construction, 5, 55, 314; accretional burial type, 140, 181; of ceremonial fire mounds, 8; dismantling and burning of, 383; platform type, 140–141, 143, 155–156, 181; as ritual act, 121–122; structural type, 140; in Titus phase, 381–383; of two-stage form, 312 Mound Place Incised and Brushed, 291, F10-3 mound placement: and aspect data, 323; cluster analysis of, 320–321, 384, F11-1; in concealed areas, 171; cultural factors in, 329–330; density of, 330–331; differential distribution of, 350, 351; and elevation, 322; elites and, 387; environmental similarity model and, 318–320, T11-2; environmental variables in, 176, 325, T11-4; and gateway communities, 176; general pattern of, 314–315, 332–333; and geology, 323–324; and habitation areas, 175–176, 331, 360–361, 373, 386–387; and hydrography, 316, 318, 321–322; and landform, 355–357, T12-3; numerical distribution of, 347, 349; predictive model of, 158–159, 164–165, T6-3, T6-4; and slope, 322; and soil, 324, 354; in Titus phase, 385–386; and vegetation, 324–325, T11-3; and viewsheds, 140, 158–159, 164–165, 173–174, 175, T6-3, T6-4; and water availability, 359–360; and watersheds, 320–321, F11-1 mounds, 234, 235, 345, 412; in Arkansas Basin, 140–142, F6-1; and cemeteries, 8, 371, 373, 383, 386; ceremonialism around, 365, 381; multiple, 250, 350–351; significance to population of, 139; and social complexity, 8, 124;

symbolism in, 142–143, 174–175; and visibility, 17, 142–143 Mounds Plantation site, xv, 242, 287n2 Mound Tract Brushed-Incised, 305, 306, 307, 309 Moundville site, 47, 118 Mountain Fork River, 340, 349–350, 354–355 Mountain Fork site, 351 Museum of the Red River, 343 Nacogdoches, 10 Nakasas, 430 Nash Neck Banded, 29 Nasoni, 189, 191, 357, 436 Natchitoches, 7, 10 Natchitoches Engraved, 254, 396, 419; frequency of, 420, 421, 423, T14-1 National Historic Preservation Act, 343 National Land Cover Data (nlcd), 324 National Register of Historic Places, 291, 343 Natural Resource Conservation Service (nrcs), 321, 344 Nawi haia ina site, 96, 114, T4-2, T4-3 Neches River, 115, 180, 432, 445, T4-6 Nelson, Bo, 372–373 Nelson site, 341, 345 Neosho River regions, 161, 164, 169–171 Newell, H. Perry, 29 Nodena arrow points, 423–424 Norman site, 141, 161, 169–171, F6-11 Norr, Lynette, 103 North American Datum 1927, 316 novaculite, 307, 311 Nunamiut, 72 nuts, 428 Oak Hill Village site: about, 209–212, F81; architecture at, 212–218, F8-2, F8-3; community structure at, 16, 230–233, F8-10, F8-11; excavations at, 15–16, 178; occupational history of, 220, 225–230, 233–237; radiocarbon dates from, 218, T8-2; structure distribution at, 226–230, F8-9, F8-10

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

511

ocr dating, 380–381 Oklahoma River Basin Survey (orbs), 340, 341–343, 347 O. L. Ellis Farm site, T4-3 Osage Orange bow wood, 135 Ouachita Mountains, 159, 165, 337, 339, 349, 350 Ouachita National Forest, 344 Ouachita River, 10, 16, 30, 299, 310 Ozark Plateaus region, 17, 159, 164, 174, 175

Copyright © 2012. Nebraska Paperback. All rights reserved.

Palmer, Edward, 291 Parsons, Mark L., 372 Partitioning Around Medoids (pam), 320–221 passenger pigeons, 68–69 Pat Boyd site, 345 Pate-Roden site, 342 Pawnee language, 2 Payne site, 345 Pea Patch site, 371, 396 pearl beads, 57–58 Pearson, F. J., Jr., 252–253 Pease Brushed-Incised, 255, 257, 260, 284, 304, 306, 380 Pease site, 283–284 Pennington Punctated Incised, 258 percentage stratigraphy, 271, F9-8 Perdiz arrow points, 308, 380 Perino, Gregory, 60 Perttula, Timothy K., 1–25, 209–238, 254, 288, 289; on entranceway orientation, 442; on maize consumption, 88, 95, 98, 104, 116; on Terán map, 195, 241; on Titus phase, 21, 363–410 Phillips, Philip, 32, 130, 425 Pierce Freeman Farm site, T4-3 Pilgrim’s Pride site, 367, 386; cemetery at, 369, 401, 402, 403; ceramic and lithic types at, 380; dating of, 374, 380 Pine Creek Mounds site, 342, 345, 350, 356 Pine Tree Mound site, 96, 382, T4-2 Pineywoods: abandonment of, 409; biotic habitat of, 211; community cemeteries

512

in, 408–409; dental data from, 91, 93, T4-1, T4-2, T4-5; mounds in, 381–383, T13-2; population growth in, 363. See also Oak Hill Village site pipes: ceramic, 423; stone sculptured, 130–131 Pitman site, 341 Plaquemine Brushed, 304 plazas, 189, 228, 230, 233, 352, F7-5 Pleasure Point site, 396, T4-1 Polk Estates site, T4-3 population: density of, 373; dispersal of, 51, 234, 366, 393, 408; growth of, 21, 363; loss, 10, 381; measurement and estimates of, 238n4 Post Oak Savannah, 88, 373; dental information from, 93, 95, 112, T4-1, T4-2, T4-5; mounds in, 381–382, T13-2 Poteau Plain, 308 pottery. See ceramics Powell, John Wesley, 2 precipitation, 211, 339 predictive models, 326, 328; of mound location, 158–159, 165, T6-3, T6-4 prestige goods thesis, 119, 121 Prewitt, Terry J., 27 P. S. Cash site, 367, 386 Pursell, Colin C., 143 Quatenary (Qal) alluvim, 323–324 radiocarbon analysis: at Belcher phase sites, 411–412; at Hughes site, 295, 297, 310, F10-6, T10-1; at Oak Hill Village site, 218, 238nn1–2, T8-2; at Titus phase sites, 374, 377, 381, 382, 409n1, T13-1, T13-2; at Willow Chute Bayou, 245 Ragland Fisher, Linda, 344 Red Chute Bayou, 253 Red River, 244; as boundary, 340; floodplain of, 243–244, 252, 350; Great Raft of, 54, 251–252; habitation sites along, 10, 16, 21–22, 243–244; Terán map and, 435 Reitz, Elizabeth, 77–78

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

relational database (rdbms), 316 remote sensing, 56, 316, 438, 439. See also archaeogeophysics Renfrew, Colin, 122–123 reptiles and amphibians, 69 revitalization cult model, 119 Riley, Carroll L., 58 Ripley Engraved, 380, 396 ritual and ceremony, 7, 23, 235, 442–444, 446–447; and agricultural cycle, 443–444; animals and, 50, 58; eastfacing, 441; feasting and meals, 48–49, 84, 112, 388, 438; fire in, 57; of Hasinai, 58–59, 61, 443–444; human heads in, 58–59; mortuary, 8, 365, 388, 392, 401–402; mound building as, 121–122; responsibilities at, 49–50, 439–440. See also beliefs and myths R. L. Cason site, 402 rock shelters, 322, 323 Roden site, 60, 346 rodents, 68, 82 Rogers, Robert, 4, 16, 115, 209–238 Roitsch site, 91, T4-1, T4-2, T4-3 Rookery Ridge site, 372, 374, 380 Rose, Jerome C., 88, 98, 106, 108, 113, 295 Round Mountain, 166, 168–169, 174, 175 Rowland Clark site, T4-2, T4-3 Sabine River, 110, 211, 233, 234, 373, 409, T4-6 Sabo, George, III, 18, 45–46, 286–287n2, 351, 362, 431–447 sacred economy, 121–123 sagamite, 112 Saline River, 20–21, 289, 310–312 Sam Gray sites, 386 Sam Kaufman site, 56, 59–60, T4-2, T4-3. See also Sam Kaufman site Sam Roberts site, 367, 377, 380, 386 Sanders site, 106, 134, T4-1, T4-2, T4-3 Sandlin Dam site, 367, 373, 396 Sandy Creek site, 396 satellite imagery, 324 Saucier, Roger T., 244 Schambach, Frank, 20, 32, 36, 47–85,

330, 447; on archaeological research, 314; and Cedar Grove site, 303, 417, 419; class-pattern-design system of, 417; collegiate system of, 299; and Crenshaw site, 20, 54, 56, 58; on Foster site, 430; on Spiro site, 134, 135; and Terán map, 241, 437–438 Scott, Susan L., 20, 47–85, 447 Second Old River, 52, 53 seriation analysis, 407. See also frequency seriation settlement patterns, 8, 15, 331, 360–361; archaeogeophysics and, 178; archaeological analysis of, 18, 23, 335–336; cluster analysis of, 320–321; gis analysis of, 149, 152, 316, 318, 347–351, F12-2; interior features, 179–180; and landform, 355–359, T12-3; linearity of, 360; and soils, 351–353, 361, T12-1; Terán map and, 177–178, 436–438; in Titus phase, 366–374. See also mound placement Shelby site, 367, 374, 386; community cemetery at, 369, 396, 397, 399–400; dental information from, T4-1, T4-2; social and political center of, 369, 371, F13-3 shell artifacts, 134, 136, 425, 427, 430n2, F14-4 shell tempering, 291, 299, 303–304, T10-2; in Belcher phase, 419, 420; predominance of, 297, 299 Sibley, John, 436 Sierzchula, Michael C., 429 Simms Engraved, 396 Sinner Linear Punctated, 257, 258 Sinner site, 283–284 Skidgel mound, 166, 167, 168, F6-10 slope, 322, 323, 357–358, T12-4, T12-5; direction, 357–359, T12-5 Smith, Bruce D., 238n4 Smithsonian Institution, 291 Snead, John I., 244 social complexity, 24, 134, 345, 363; agriculture and, 86; burial practices and, 119; Caddo development of, 5, 7, 8;

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

513

Copyright © 2012. Nebraska Paperback. All rights reserved.

social complexity (continued) chiefly authority and, 50; and differential resource access, 49, 83–84, 114; mounds and structures and, 8, 124 Social Hill phase, 36, 307–308 social status, 8, 20, 48, 111, 120–121, 400–401. See also elite sociopolitical organization, 23, 234–235, 385 soils: and Caddo cultural landscape, 351–353, 361; clay content in, 56, 339, 352; and mound construction, 143; and mound site selection, 324; soil associations, 353–354, T12-1, T12-2 Solon Stanley Farm site, 109, 110, T4-1, T4-2, T4-3 Soule, William S., 177, 437 Southeastern Ceremonial Complex, 4, 118, 425 Spanish, 431–433; de Soto expedition of, 7, 291, 311; economic symbiosis with, 10; Terán expedition of, 18, 177, 432–433, 435 spatial variation, 263, 283, 284, 285–286, 288, 289 Spencer Spring site, 342 Spider Lilly site, 396 Spiro phase, 59, 124 Spiro site: about, 123–124; chiefdom at, 118–119; Copple mound at, 156, 166; Craig mound at, 117–118, 125, 153–154, 166, F5-1; early study of, xv, 117; four periods of, 125, 133–134; population density of, 124; topographic profile of, 167, F6-10; as trade-exchange center, 119–121, 135–136; viewsheds at, 166. See also Great Mortuary Spoonbill site, 373 squash and pumpkin, 19, 427 stable isotope data, 97–103, 116, T4-3; by biotic region, 108–109; by river basin, 109–110, T4-6; by site type, 111–112, 114, T4-7; by time period, 103–108, 114–115, F4-1, F4-2, T4-4 stamping, 27, 29, 258, 291, 305 Standridge site, 308, 309

514

A State Plan for the Conservation of Archaeological Resources in Arkansas (Schambach and Early), 314 Steck site, 373 Steele, D. Gentry, 96 stone artifacts, 9, 311; in Belcher phase, 308, 423–424; from chipped stone, 307–310, 423, 424, F10-10; from ground stone, 424; pipes, 130–131. See also arrow points Story, Dee Ann, 186, 233 Struever, Stuart, 62 Suhm, Dee Ann, 28, 29, 31–32 Sulphur River, 96, 113, 373, 409, T4-6 Survey and Planning era, 340, 343 Swan Lake site, 246–247, 250 Swanton, John R., 433 Swauano Creek, 393, 396 Taddlock site, 76–80 Talco arrow points, 308, 397, 400 Tallow Grove site, 93, T4-1, T4-2, T4-3 Tanner, Helen, 4–5 Taylor Engraved, 305, 396 temporal-spatial systematics, 11–12, 22–23, 324 temporal variation, 115, 259–263, 271, 273 Terán de los Ríos, Domingo, 18, 177, 431, 432, 435, 443; diary of, 433–434, 440–441 Terán map (Mapa de la provincia donde habita la Nacion Casdudacho), 357, 414, F9-2; archaeological uses of, 189, 296n2, 436–439, 446; as Caddo cosmogram, 18, 431, 439–447; creation of, 433–434; depiction of farmsteads in, 177, 431, 435–436; facsimile of, F15-1; and Hatchel site, 240; and Hill Farm site, 194–195, 207, F7-9; three elements of, 435–436 terrace deposits, 323, 324 Texarkana phase, 438 Texas in the Middle Eighteenth Century (Bolton), 433 textiles, 126, 130, 136 Thomas P. Caldwell site, 492

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

Copyright © 2012. Nebraska Paperback. All rights reserved.

Thurmond, J. Peter, 384, 386, 400 Tiddle Lake site, 380, 386–387 Tieszen, Larry L., 103 Tinhiouen, 10 Titus phase, 21, 311; about, 364–365, F13-1; age of, 374–381, T13-1, T13-2; agriculture in, 366; cemeteries and burials in, 373, 388, 395–396, T13-3; ceramics and arrow points in, 308, 364–365, 380, 407; demise of, 409; density of sites in, 373; diet in, 106, 366–367; mound construction and location in, 369, 373, 381–383, 385–386; political communities and settlements in, 366–374, 385–388, 405, 409, F13-2; population dispersal in, 366, 367, 373–374, 409; sociopolitical diversity in, 385 Tom Hanks site, 367, 369 Tom Jones site, 18, 311, 313, 322, 333, 438; about, 316 Tonti, Henri de, 254, 429, 430 Toqua site, 81, T3-5 trade, xvi, 314, 363; in ceramics, 289, 420; in deer hides, 9; with Europeans, 10; as exchange, 136; in exotic artifacts, 135, 137; interregional, 8; networks of, 9, 23; and sacred gift exchange, 136; and Spiro Great Mortuary, 119, 133–136 transportation, 311, 359 tribal identity, 429–430 Trinity River basin, 96 Trubitt, Mary Beth, 20–21, 288–312 Trubowitz, Neal L., 241, 414–415, 438 Tuck Carpenter site, 367, 369, 374, 396 turkey, 68 Tyler, James, 295 Universal Transverse Mercator (utm), 316 University of Arkansas, 32 University of Texas, 191, 414, 437 Upper Nasoni, 177, 207, 357, 433, 436 U.S. Army Corps of Engineers, 343, 344 Vanceville mound site, 246, 248, 284,

286; contour map of, F9-5; midden deposits at, 259, 271, 283; radiocarbon assays from, 261–263, T9-1; sherds at, 271, 284, T9-2; temporal variation at, 259–271; test pits at, 260, 261–262, 271, F9-6, F9-7 Vanceville Village site, 248 Vehik, Susan C., 135 Victor Area site, 155 viewsheds: in Arkansas River Basin, 143– 148, 165–169, T6-1, T6-2; elevation and, 159–160, F6-7, F6-8, F6-9; gis analysis of, 143, 149–152; and mound location, 140, 158–159, 164–165, 173–174, 175, T6-3, T6-4; in Neosho River Region, 169–171, F6-11; reciprocity of, 148; size of, 152–157, 173–174, F6-2, F6-3, F6-4, F6-5 Vogel, Gregory, 17, 139–176 W. A. Ford site, 402 Walker, Chester P., 16–17, 177–208 Walnut Creek, 396 warfare, 23, 408 Waring, Antonio J., Jr., 119 Washington Square site, T4-2 Washita arrow points, 308 Washita River, 11 Waterfall-Gilford watershed, 344 water resources, 318, 339–340, 349, 359–360, T12-6 watersheds, 318, 344; and mound site location, 320–321, F11-1 Watkins site, 386 Webb, Clarence H., 256, 271, 279; and Belcher site, 21, 411, 412, 413, 423, 425; on Bossier focus, 242–243, 284; classification of ceramics by, 417, 419, 420; on Freeman-Custis Expedition, 430; and Jim Burt site, 250; typological and chronological schemes of, 11, 284, 417; and Vanceville site, 246, 259, 284 Weber, J. Cynthia, 32; class-patterndesign system of, 417 Weches Fingernail Impressed, 257 Wedel, Mildred Mott, 429, 433, 436

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,

515

ment variation across, 272–283, F9-11; temporal variation at, 259–271 Wilson, Diane, 19–20, 86–116 Winding Stair site, 308 Woodland period, 89, 113, 240, 246, 250, 345 Woods Mound Group site, 342, 346 Woodward Plain, 308 Works Progress Administration (wpa), 124, 149, 181, 191, 340–342, 437, 438 W-S site, 396, 397 Wyckoff, Don G., 25n3, 342, 344, 347 xinesi, 9–10, 87, 286n2, 351, 440 Yatasi, 430 Young site, 346 zoned punctated, 257, 273

Copyright © 2012. Nebraska Paperback. All rights reserved.

Weinand, Daniel C., 77–78 Werner site, 246, 279–283, 284, 286, F7-16 Wheatley, David, 148 Whelan phase, 377, 380, 384 Whelan site, 311, 367, 371, 383, 385, 387–388 White Oak Creek, 372–373 Wichita people, 25n3; language of, 2 Wilkinson Punctated, 257 Willey, Gordon R., 318, 335 Willow Chute Bayou locality: archaeological investigations of, 14–15, 242, 246–249, F9-4; ceramic diversity at, 254–259; geography of, 239, F9-1; human occupation at, 249–250, 251–254; natural setting of, 244–245, F9-3; and Red River flooding, 252–253; in regional perspective, 283–285; settle-

516

Index

The Archaeology of the Caddo, Nebraska Paperback, 2012. ProQuest Ebook Central,