Chaco's Northern Prodigies : Salmon, Aztec, and the Ascendancy of the Middle San Juan Region after AD 1100 [1 ed.] 9781607817918, 9780874809251

Citation preview

Chaco’s Northern Prodigies Salmon, Aztec, and the Ascendancy of the Middle San Juan Region After AD 1100 Edited by Paul F. Reed

Chaco’s Northern Prodigies

Chaco’s Northern Prodigies Salmon, Aztec, and the Ascendancy of the Middle San Juan Region after AD 1100 Edited by

Paul F. Reed

The University of Utah Press Salt Lake City

© 2008 by The University of Utah Press. All rights reserved. The Defiance House Man colophon is a registered trademark of the University of Utah Press. It is based upon a four-foot-tall Ancient Puebloan pictograph (late PIII) near Glen Canyon, Utah. 12  11  10  09  08    1  2  3  4  5 Library of Congress Cataloging-in-Publication Data Salmon Working Conference (2004:Farmington, N.M.)   Chaco’s northern prodigies : Salmon, Aztec, and the ascendancy of the   middle San Juan region after ad 1100  /  edited by Paul F. Reed.     p.  cm.   Proceedings of the Salmon Working Conference held in Farmington, New Mexico, in April 2004.   Includes bibliographical references.   isbn 978-0-87480-925-1 (cloth : alk. paper)  1.  Salmon Site (N.M.)— Congresses.  2.  Pueblo Indians—New Mexico—Antiquities—Congresses. 3.  Pueblo Indians—Material culture—New Mexico—Congresses. 4.  Excavations (Archaeology)—New Mexico—Congresses.  5.  Pueblo pottery— New Mexico—Congresses.  6.  Plant remains (Archaeology)—New Mexico— Congresses.  I. Reed, Paul F.  II. Title.   E78.N65S325 2008   978.9'8201—dc22 2008005508 Printed and bound by Sheridan Books, Inc., Ann Arbor, Michigan Interior printed on recycled paper with 50% post-consumer content.

Contents

List of Figures   vii List of Tables   xi Preface   xiii Part 1: Introduction 1. Setting the Stage: A Reconsideration of Salmon, Aztec, and the Middle San Juan Region in Chacoan and Post-Chacoan Puebloan History   3 Paul F. Reed Part 2: Salmon Pueblo 2. Salmon Ruins: Architecture and Development of a Chacoan Satellite on the San Juan River   29 Larry L. Baker 3. Salmon Pueblo as a Ritual and Residential Chacoan Great House   42 Paul F. Reed Part 3: Material Culture Analyses: Salmon Pueblo 4. Subsistence and Plant Use During the Chacoan and Secondary Occupations at Salmon Ruin   65 Karen R. Adams 5. Parasite Pathoecology of Salmon Pueblo and Other Chacoan Great Houses: The Healthiest and Wormiest Ancestral Puebloans   86 Karl J. Reinhard 6. Animal Bone from Salmon Ruins and Other Great Houses: Faunal Exploitation in the Chaco World   96 Kathy Roler Durand and Stephen R. Durand 7. Sighting Along the Grain: Differential Structural Wood Use at the Salmon Ruin   113 Thomas C. Windes and Eileen Bacha 8. Human Remains Recovered from the Tower Kiva at Salmon Ruins   140 Nancy J. Akins v

Part 4: Material Culture Analyses: Across the Middle San Juan 9. An Initial Assessment of Perishable Relationships Among Salmon, Aztec, and Chaco Canyon   167 Laurie D. Webster 10. Ceramics of the Middle San Juan Region: Potters, Recipes, and Varieties   190 Lori Stephens Reed Part 5: New R esearch in the Middle San Juan R egion 11. Puebloan Communities on the South Side of the Middle San Juan River   209 Linda Wheelbarger 12. Animas Anamnesis: Aztec Ruins or Anasazi Capital?   231 Gary M. Brown, Thomas C. Windes, and Peter J. McKenna 13. Comparing Great House Architecture: Perspectives from the Bluff Great House   251 Catherine M. Cameron Part 6: Sy nthetic Views of the Middle San Juan R egion 14. Chacoan Society: The View from Salmon Ruins   273 Cynthia Irwin-Williams 15. The Position of Salmon Ruins in the Middle San Juan, ad 1000–1300: A Perspective from Ceramic Design Structure   284 Dorothy Washburn 16. The La Plata, the Totah, and the Chaco: Variations on a Theme   309 H. Wolcott Toll 17. Sacred Landscapes: The Chaco-Totah Connection   334 Ruth M. Van Dyke Part 7: F rom the Outside: Salmon and the Middle San Juan R egion in a Larger Context 18. Population Dynamics among Salmon’s Northern Neighbors in the Central Mesa Verde Region   351 Mark D. Varien, Scott G. Ortman, Susan C. Ryan, and Kristin A. Kuckelman 19. The Middle San Juan and Chaco Canyon   366 R. Gwinn Vivian Part 8: Conclusion 20. Prodigy, Rebel, or Stepchild? The Middle San Juan Region vis-à-vis Chaco Canyon   379 Paul F. Reed References   389 Contributors   429 Index   431 vi

Figures

1.1. Map of the Middle San Juan region   4 1.2. Timothy O’Sullivan photo of Salmon Pueblo   9 1.3. O’Sullivan photo of Salmon Pueblo   15 2.1. Location of Salmon Ruins in northwest New Mexico   31 2.2. Location of rooms with early tree-ring dates in Salmon’s east wing   32 2.3. Rooms with at least two T-shaped doorways   33 2.4. Abutments and continuous wall sections of Salmon Pueblo   34 2.5. Astronomical observations and architectural reference points at Salmon Pueblo   36 2.6. Reconstructed Palmer Drought Severity Index (PDSI) values for Chaco Canyon, ad 1000–1100   39 2.7. Comparison of forms and ground plans of Salmon Pueblo and Hungo Pavi   40 3.1. Map showing Chacoan (Primary) occupation at Salmon Pueblo   46 3.2. Distribution of Chacoan floor hearths at Salmon Pueblo   48 3.3. Map of Salmon showing Chacoan rooms with hearths   49 3.4. Distribution of Chacoan milling bins at Salmon Pueblo   50 3.5. Map of Salmon showing Chacoan rooms with milling bins   51 3.6. Distribution of storage features in Chacoan rooms at Salmon Pueblo   52 3.7. Map of Salmon showing Chacoan rooms with storage features    53 3.8. Distribution of trash deposits in Chacoan rooms and units at Salmon Pueblo   54

3.9. Map of Salmon showing Chacoan rooms with trash strata   55 3.10. Map of Salmon showing Chacoan rooms with burials   56 3.11. Distribution of co-occurring residential features in Chacoan rooms at Salmon   59 3.12. Map of Salmon showing Chacoan rooms with co-occurring residential features   60 5.1. Diagram showing the wide spectrum of parasites that infected ancestral Puebloans   87 5.2. Map showing variation in percentages of pinworm parasitism among ancestral Pueblo villages    88 5.3. Diagram showing modes of pinworm transmission to human hosts   88 5.4. Graph comparing pinworm parasitism with porotic hyperostosis prevalence for several southwestern locales   89 6.1. Locations of sites and localities in the San Juan Basin and the Four Corners region discussed in the text   97 6.2. Generalized ground plan of Salmon Ruin   99 6.3. Turkey burial from Room 100   106 6.4. The second turkey burial from Room 100   107 6.5. The relationship between sample size and diversity of all fauna taxa    108 6.6. The relationship between sample size and diversity of ritual fauna   109 7.1. The region around Salmon Ruins showing major drainages and the closest stands of trees for construction   118 vii

Figures 7.2. Tree species used in the Chacoan construction of Salmon Ruin    119 7.3. Beam end treatments recorded at Salmon Ruin   121 7.4. Profiles of the west and east walls of Room 62   122 7.5. Collapsed roofing in Room 118   124 7.6. Tree-ring dates from Salmon Ruin   125 7.7. Map of Salmon showing probable thirdstory additions   126 8.1. Distribution of subadult remains in the Tower Kiva at Salmon   144 8.2. Distribution of burning for subadult remains   144 8.3. Distribution of mature remains   160 8.4. Distribution of burning for mature remains   160 9.1. Coiled baskets   170 9.2. Plaited ring baskets   171 9.3. Plaited sandals   172 9.4. Twined sandals   173 9.5. Cotton fabrics and looped turkey feather shoe-socks   174 9.6. Painted and turquoise-covered coiled baskets   178 9.7. Plaited pouches   180 9.8. Twined reed objects   181 9.9. Fabric-wrapped bundles   182 9.10. Weaving tools   185 10.1. Scatter plot showing ICP signature group membership   195 10.2. Distribution of paint types for Northern San Juan Animas Variety Pueblo II bowls from Salmon, Aztec, and the Tommy site   198 10.3. Distribution of paint types for Northern San Juan Animas Variety Pueblo III bowls from Salmon, Aztec, and the Tommy site   199 10.4. Distribution of paint types for Cibola Animas Variety Pueblo II and III bowls from Salmon, Aztec, and the Tommy site   200 10.5. Distribution of slip type for Northern San Juan Animas Variety Pueblo II bowls from Salmon, Aztec, and the Tommy site   201

10.6. Example of Northern San Juan Mancos Black-on-white (Dogoszhi style) Animas Variety and McElmo Black-on-white Animas Variety from Salmon Ruins   201 10.7. Distribution of slip types for Cibola Animas Variety Pueblo II and III bowls from Salmon, Aztec, and the Tommy site   202 10.8. Cibola Chaco Black-on-white Animas Variety bowl   203 11.1. View of the B-Square Ranch from the Point site   210 11.2. Boundaries of the B-Square Ranch   211 11.3. Puebloan communities on the south side of the middle San Juan River   212 11.4. Plan view of the Fort site   214 11.5. Photograph of the Fort site   215 11.6. Photograph of Needle Rock   217 11.7. Plan view of the Tommy site   218 11.8. Plan view of the major roomblock at the Tommy site   219 11.9. View of the Point site, looking northwest over site area on bench   220 11.10. Bolack’s plan view of Mine Canyon   224 12.1. Distribution of tree-ring dates from Aztec West Ruin   234 12.2. Distribution of tree-ring dates from Aztec East Ruin   235 12.3. Plan maps of Aztec West Ruin and Salmon Ruin at time of ultimate construction, ad 1200s   236 12.4. Late Bonito phase construction sequence at Aztec West   239 12.5. McElmo phase construction sequence and major remodeling at Aztec West   241 12.6. Mesa Verde phase construction and major remodeling at Aztec West   243 12.7. Plan map of Aztec East Ruin showing pattern of latest tree-ring dates within individual structures   244 12.8. Distribution of sites in the main Aztec Ruins complex and on the Pleistocene terrace to the north   246 12.9. Schematic depiction of symmetrical layout of major architectural features in the main Aztec Ruins site complex   247 viii

Figures 13.1. Map of the northern Southwest showing the locations of excavated great houses   252 13.2. Map of the Bluff great house site   253 13.3. Map of the Bluff great house showing the location of excavation units and walls   254 13.4. Plan view of the Bluff great house showing two construction episodes   255 13.5. Photo of edge-flaked or “scabbled” masonry from the western section of the Bluff great house   256 13.6. Core-and-veneer wall at the Bluff great house   257 13.7. Kiva roof construction methods for Chaco style domed roofs and Mesa Verde style cribbed roofs   259 13.8. Plan views of Wijiji in Chaco Canyon, Ida Jean Ruin, and Wallace Ruin   261 13.9. Plan views of Salmon Ruins and Aztec West, the largest great houses outside of Chaco Canyon   262 13.10. Plan views of Tsin Kletzin and Kin Kletso in Chaco Canyon and Escalante and Morris 39, termed McElmo style buildings   263 13.11. Plan view of Chimney Rock Pueblo in the far eastern portion of the Northern San Juan region   264 13.12. Plan views of Edge of the Cedars, Far View, and Lowry Ruins in the Northern San Juan region   265 15.1. Designs displaying element and motif variety but structural homogeneity in the local San Juan design system   289 15.2. Examples of the limited pattern and symmetry repertoire of the Chaco design system   291 15.3. Mistakes in executing the Chaco design system   292 15.4. Chaco Black-on-white p2 pattern on cylinder jar from Pueblo Bonito   292 15.5. Chaco Black-on-white p112 pattern on cylinder jar   293 15.6. Chaco Black-on-white pm pattern on cylinder jar   293

15.7. Local San Juan simplification of the Chacoan design system   295 15.8. Graph of the use of p112 and p1a1 symmetries at Salmon and Aztec West, and at Mug House and Long House   302 15.9. Designs showing element and motif variety but homogeneity in p112 structure from Salmon’s Secondary occupation   303 15.10. Two popular p1a1 motif combinations at Salmon and Aztec West   304 15.11. One-dimensional p1a1 pattern (a) that results from truncation of the twodimensional­pgg pattern (b)(after Zaslow 1977: Figure 4)   305 16.1. Map of the Totah region in the ad 1000s through 1200s   312 16.2. Map showing locations of great houses in the La Plata Valley and communities defined during the La Plata Highway Project   330 17.1. Map of the Middle San Juan and northern Chacoan world   335 17.2. Aerial photo of Salmon Pueblo   340 17.3. Stein and McKenna’s (1988) map of the Aztec ritual landscape   342 17.4. Plan views of McElmo great houses in Chaco   345 17.5. Photo of Kin Ya’a   347 18.1. Map of the Northern San Juan region   352 18.2. Map showing locations of surveys in the study area for the Village Ecodynamics Project   353 18.3. Momentary population estimates for the study area   354 18.4. Maps showing the intensity of occupation at Yellow Jacket Pueblo during four successive time periods   361 18.5. Map of the visible architectural remains and electrical resistivity anomalies at Albert Porter Pueblo   363 18.6. Plan view of the great house at Albert Porter Pueblo showing structures mentioned in the text   364

ix

Tables

1.1. Scholars associated with the Salmon Research Initiative and the Salmon Working Conference   5 1.2. Chronological sequences used in the Middle San Juan region   18 2.1. Astronomical observations and architectural reference points at Salmon Pueblo   37 3.1. Generic categories used in Salmon feature analysis   48 3.2. Chacoan age trash pit features at Salmon Pueblo   54 3.3. Estimated Salmon population and deaths by period   58 4.1. Publications and master’s theses focusing on plant remains from Salmon Ruin   66 4.2. Methodological contributions to the ethnobotany of Salmon Ruin   67 4.3. Plant taxa identified from Salmon Ruin   70 6.1. References for sites included in this study   98 6.2. Dietary fauna from Chacoan sites   99 6.3. Artiodactyl, lagomorph, and turkey indices   100 6.4. NISP data and indices for selected sites in the Sand Canyon region   101 6.5. Salmon Ruin strata with eggshell present   102 6.6. Summary of macaw and turkey burials at Salmon Ruin   105 7.1. Construction wood recovered from prehistoric small sites in the Salmon region, ad 600s–1500s   116

7.2. Construction wood recovered from historic sites in the Salmon region   117 7.3. Unanalyzed dendrochronological samples from the Salmon excavations of 1972   120 7.4. Construction elements and total trees needed for the initial and secondary constructions at Salmon Ruin   123 7.5. Use of tree species in the architectural core units at Salmon and Aztec West   128 7.6. Tree harvest periods by species and selected years   132 8.1. Subadult remains by age from the Tower Kiva at Salmon   145 8.2. Summary of immature remains recovered from the Tower Kiva   153 8.3. Age and completeness of subadult remains from the Tower Kiva   156 8.4. Distribution of elements of mature remains from the Tower Kiva   157 8.5. Comparison of age distributions of Tower Kiva remains and inhumations   159 8.6. Tower Kiva cumulative age proportions compared to other Southwest populations   159 8.7. Age by predominant burn type from the Tower Kiva   161 8.8. Comparison of burn types on Skeleton 34 and faunal assemblages from burned roofs at LA 6170   162 9.1. Worked-fiber artifacts from Chaco Canyon, the West Ruin of Aztec, and Salmon Pueblo   175 10.1. Local and imported varieties of the Northern San Juan and Cibola ceramic xi

Tables traditions in the Middle San Juan region   197 11.1. Tommy site decorated ceramic counts   221 12.1. Chronology of Puebloan occupation at Aztec Ruins   233 12.2. Main stages in architectural evolution at Aztec West   237 12.3. Outer ring classification on twelfth-century cutting and near-cutting dates from Aztec East   249 13.1. Great house layout styles and number of Chaco era and post–Chaco era kivas   260 13.2. Internal characteristics of great house Chaco era kivas   267

15.1. Comparison of one- and two-dimensional design symmetries at Salmon and Aztec   290 15.2. Sample of Chaco wares with classifiable symmetries from the Primary occupations of Salmon and Aztec   294 15.3. Distribution at Salmon of axial categories by ceramic type   300 15.4. Comparison of one-dimensional symmetries on Secondary occupation ceramics from San Juan and Mesa Verde sites   301 16.1. La Plata Highway Project sites by component age assignments   310 16.2. Material occurrences by period at La Plata small sites, Chaco small sites, and Pueblo Alto   316

xii

Preface

The chapters in this volume flow from the Salmon Working Conference, convened in Farmington, New Mexico, in April 2004. I organized the conference with the primary goal of gathering archaeologists working in the Middle San Juan region to explore and discuss renewed archaeological research over the last 10 years. I also invited conference discussants with extensive knowledge of Chacoan and Puebloan archaeology. The result was a dynamic meeting of minds, with much consensus but also much debate. It is my hope that this volume re-creates for readers the lively discourse at the conference in its presentation of the latest findings of Middle San Juan archaeology. As the final, synthetic part of the original Salmon Research Initiative, this volume builds upon the original Salmon research from the 1970s and the three-volume, comprehensive Salmon report completed in July 2006, Thirty-Five Years of Archaeological Research at Salmon Ruins, New Mexico (edited by P. Reed), published jointly by the Center for Desert Archaeology and Salmon Ruin Museum. Much of the research discussed in this volume originated because of the foresight, brilliance, and persistence of Cynthia Irwin-Williams. Over the last six years, I have come to a great appreciation for Cynthia’s ideas about Salmon and the Middle San Juan region. I think she correctly identified many of the region’s core research issues. She also coined the term Chaco Phenomenon. Although it is now out of vogue—having been replaced over the years by Chaco System, Chaco World, Chaco ­Experience, and others—in my mind it may be

the most descrip­tive label. Cynthia proposed an explicit colonization model for the founding of Salmon by Chacoan migrants, and she also inspired and touched scores of people during her short 54 years of life. Many of the contributors to this volume are former students of Cynthia or worked with her in the field at Salmon: Karen ­Adams, Larry Baker, Gary Brown, Cathy Cameron, Steve Durand, Steve Lekson, ­Peter ­McKenna, Karl Reinhard, and ­Dorothy Washburn. The Center for Desert Archaeology and Salmon Ruins have supported new research at Salmon and across the Middle San Juan region for more than six years. In addition, the Center funded the Salmon Working Conference and the research that went into this publication. I thank Bill Doelle and Linda Pierce, and other Center staff, for their encouragement and assistance over the last several years. Doug Gann of the Center created the digital reconstructions of Salmon Pueblo and the Salmon site maps used in this book. Catherine Gilman of the Desert Archaeology Center drafted several additional maps. Larry Baker, Salmon’s executive director, deserves my hearty thanks for encouraging and supporting renewed research at Salmon and across the Middle San Juan over the last six years. Other Salmon staff and volunteers helped with this effort, including Diane Hayden, Nancy Sweet Espinosa, Marissa Miller, and the late Jim Snyder. My work at Salmon has benefited from extended conversations with the following archaeological colleagues: Larry Baker, Gary Brown, Steve Durand, Kathy Roler Durand, Lori Stephens Reed,

xiii

Preface Dorothy Washburn, Laurie Webster, and Tom Windes. I wish to thank all the volume authors for preparing and revising their chapters several times and sticking with the publication schedule. Our work in this volume has benefited from reviews by Chuck Riggs and Carroll Riley. During the conference, Gwinn Vivian and Linda Cordell served as discussants. Linda was not able to contribute written insights to this volume, but Gwinn’s discerning comments appear in Chapter 19. I thank Gwinn and Linda for helping me and the other authors stay on track and keep our Middle San Juan research properly grounded in the larger Southwest.

Alexis Mills tackled the thankless job of editing this enormous volume, doing her usual magic and producing an excellent manuscript. John Herbert and his University of Utah Press staff carried the project to completion with great skill, resulting in a high-quality book. Finally, I thank my wife, Tristan Kwiecinski, for her help with editing and reference chasing, and her unflagging support and love throughout this long process. Paul F. Reed January 2008

xiv

Pa rt 1

Introduction

1

Setting the Stage A Reconsideration of Salmon, Aztec, and the Middle San Juan Region in Chacoan and Post-Chacoan Puebloan History Paul F. Reed In the late eleventh and early twelfth centuries, the ancient pueblos at Aztec and Salmon in the ­Middle San Juan region rapidly emerged as population and political centers during the closing stages of Chaco Canyon’s ascendancy. Some archaeologists have attributed the development of these centers to migration and colonization by people from Chaco Canyon, 75 km to the south (e.g., Irwin‑Williams and Shelley 1980; Morris 1915). Others have suggested that the so‑called Chacoan “system” was largely the result of emulation of Chacoan characteristics by local groups in outlying areas (e.g., Van Dyke 1999; papers in Kantner and Mahoney 2000). Research over the last five years in the ­Middle San Juan region suggests that both of these processes were operating: some sites (e.g., Salmon and Aztec) exhibit architecture, ceramics, and other material culture that support direct Chacoan migration. Other sites seem to reflect local choices to emulate Chacoan culture but do not contain clear Chacoan hallmarks (e.g., fine Chacoan banded masonry). The chapters in this volume explore this issue, as well as many others, as the authors detail the latest findings regarding the Chacoan and post-Chacoan archaeology of Salmon, Aztec, and the Middle San Juan region. Some archaeologists use the Navajo name Totah to describe the ancient Puebloan homeland in the Middle San Juan region (McKenna and Toll 2001; Toll, Wheelbarger, this volume). For the research reported in this volume (as well as other projects completed by the Center for Desert Archaeology

and Salmon Ruins Museum), we prefer the longer, more inclusive, and neutral term: “Middle San Juan region.” The boundaries of this region go well beyond the area outlined by Toll (this volume) and others as the “Totah.” Figure 1.1 shows the Middle San Juan region and the locations of Aztec, Salmon, and other ancient Puebloan communities. The ruins of Salmon Pueblo lie on the north bank of the San Juan River, approximately 3.5 km west of Bloomfield and 15 km east of Farmington, New Mexico. Salmon Pueblo was constructed as a Chacoan outlier in ad 1090 with at least 275 original rooms, an elevated kiva in its central roomblock, and a great kiva in its plaza. The Chacoan occupation lasted into the 1120s. Subsequent use by local San Juan populations (from the mid-1100s through the late 1200s) resulted in extensive modification to the original building, with use of approximately 200 room spaces, subdivision of original large, Chacoan rooms, and emplacement of more than 25 small kivas into rooms and the plaza. Occupation of Salmon during this San Juan period continued until the late 1280s or 1290s, when a catastrophic fire terminated the use-life of the pueblo. Through a partnership with the San Juan County Museum Association, the Salmon site was excavated in the 1970s by Dr. Cynthia IrwinWilliams, her colleagues, staff, and students from Eastern New Mexico University. Approximately 30 percent of the site was excavated by field school crews, paid professionals, and local volunteers. In all, more than 700 individuals participated in the 3

Paul F. Reed

Figure 1.1.  Map of the Middle San Juan region in relation to Chaco Canyon and Mesa Verde.

fieldwork. Concurrent with the field operations, laboratories were operated on the premises to wash, catalog, and rough-sort artifacts and samples. More than 1.5 million artifacts and samples were recovered from Salmon. In 1980, Irwin-Williams and coprincipal investigator Phillip Shelley compiled and edited a five-volume report to the funding agencies (including the National Science Foundation, the National Endowment for the Humanities, and the Four Corners Regional Commission, among ­others; Irwin-Williams and Shelley 1980). Although comprehensive and voluminous, the 1980 report was assembled quickly, with incomplete editing, and its distribution was extremely limited. Only 100 copies were printed and distributed, primarily to universities and research libraries. In the summer of 2001, the Center for Desert Archaeology and Salmon Ruins Museum initiated

a multiyear partnership to renew the research potential of the Salmon site. The Salmon Project, part of the Center’s Heritage Southwest Project, was intended to bring to fruition to the great effort put forth by Irwin‑Williams and her staff during the 1970s. A major thrust of the research effort involved a new look at Salmon Pueblo, not only as a Chacoan outlier, but in its regional, thirteenth‑century­Middle San Juan context as well. Publication of a multivolume technical report and a synthetic, summary volume was the primary goal of the new Salmon Project. The three-volume technical report is published and available (P. Reed 2006a); this volume is the synthetic portion. An additional goal of the Center-Salmon partnership involved reassessment of Middle San Juan Pueblo archaeology and its role in the Chacoan and post-Chacoan worlds. To achieve these research and publication goals, 4

Setting the Stage Table 1.1.  Scholars Associated with the Salmon Research Initiative and the Salmon Working Conference

Name

Specialty

Karen Adams Nancy Akins Larry Baker Gary Brown Cathy Cameron Linda Cordell Kathy Roler Durand Steve Durand Jeff Eighmy Nancy Espinosa Hayward Franklin Cynthia Irwin-Williams­ Steve Lekson Peter McKenna Tori Myers Lori Stephens Reed Paul Reed Karl Reinhard H. Wolcott Toll Ruth Van Dyke Mark Varien Gwinn Vivian Dorothy Washburn Laurie Webster Tom Windes

ethnobotany osteology architecture Aztec Ruins architecture Bluff site, Chaco Puebloan archaeology faunal remains, Chaco faunal remains, Chaco archaeomagnetism osteology, curation Sterling, ceramics Salmon, Chaco Chaco, big picture Sterling, ceramics ceramics ceramics Salmon, Middle San Juan coprolites La Plata, Chaco Chaco northern San Juan Chacoan archaeology ceramic symmetry textiles, perishables wood, tree-ring dating

Current Salmon Researcher

Salmon Conference Attendee

Original Salmon Project

yes yes yes no no no yes yes yes yes no no no no yes yes yes yes no no no no yes yes yes

yes yes yes yes yes yes yes no no yes yes no yes yes no yes yes yes yes yes yes yes yes yes yes

yes no yes yes yes no no yes no no yes yes yes no no no no yes no no no no yes no no

I recruited two groups of scholars: one group to research different aspects of Salmon’s archaeological record, and a second group of outside researchers to contribute data and interpretation of the surrounding Middle San Juan region (Table 1.1). Four years of work by these two groups of contributors resulted in this synthetic volume, interpreting 35 years of research at Salmon Ruins, as well as reassessing and reinterpreting the ancient Pueblo II–III occupation of the Middle San Juan region. To set the stage for this volume, the Salmon Working Conference was convened in April 2004 in Farmington, New Mexico. Invited scholars presented papers on a variety of topics; more than half of them focused directly on aspects of Salmon’s archaeology, history, and material culture. The remaining papers were

presented by the group of outside scholars working across the Middle San Juan region (and beyond). A central goal of the Salmon Conference was to provide a forum for extended discussion of research issues critical to understanding Middle San Juan Pueblo archaeology. Extended presentation and discussion windows allowed these interactions to take wing. The result was an invigorating seminar, going well beyond the basics of a standard conference. After the conference, participants were asked to revise their papers in light of the ongoing discussions and submit finished manuscripts. Except for a paper by the late Cynthia Irwin‑Williams (a combination of her work from 1980 and 1983), the chapters in the book are new, original compositions. I want to introduce this volume by discussing 5

Paul F. Reed several themes and issues. I first briefly review the history of archaeological exploration and research in the Middle San Juan region. Next, I discuss new research at Salmon that has broadened and changed our interpretations, particularly with regard to the timing of its occupation (P. Reed 2006b). The chronological refinements realized during the new Salmon Research Initiative have implications for how we view chronology across the Middle San Juan region. Thus, a discussion of the Salmon sequence and the revised regional chronology is necessary. Third, I continue the introduction to Middle San Juan archaeology by delineating the timing and nature of late eleventh and early twelfth century Chacoan expansion to the north and the subsequent evolution of the local communities. I also explore the reasons behind this northward movement. Fourth, the function and use of Chacoan great houses have captured the interest of archaeologists for more than 100 years. Explanations of the great houses have varied from massive apartment buildings accommodating thousands of people to empty ceremonial-administrative complexes with few residents (see Lekson 2006b; Reed 2004a). New research at Salmon Pueblo has revealed that the site was used as a residential and ritual center by the Chacoans (e.g., Durand and Durand, this volume; Reed, this volume). Recent work at Aztec Ruins (Brown et al., this volume) also suggests a more complicated view of great house function and use. Fifth, attention has been refocused on Chaco Canyon and Chacoan sites across the Southwest over the last 10 years (e.g., Cameron and Toll 2001; Cordell and Judge 2001; Kantner and Mahoney 2000; Lekson 2006a; Reed 2004a; Van Dyke 2003a). As a result, interpretations of great houses have become more refined. Salmon, Aztec, and other Middle San Juan great houses thus far have been only peripherally invoked in these new interpretations. Nevertheless, consideration of the role of the Middle San Juan communities in the ancient Chacoan and Puebloan worlds is critical for understanding the Pueblo II–III interval. Aspects of the revised view of the Middle San Juan and the larger implications are highlighted below. Finally, following the exploration of the Middle San Juan’s place in Puebloan history, I consider the larger

i­ ssue of social and ethnic identity and relationships across the region. I want to offer two caveats prior to delving into the issues. My primary desire in putting together the primary team of researchers, and in recruiting additional scholars for the Salmon Working Conference, was to refocus attention on Salmon, Aztec, and the Middle San Juan region. This goal flowed directly from the Center’s purpose in creating the Salmon Research Initiative: to jumpstart the longdormant, sleeping giant of Salmon and Middle San Juan research. As we began this process, it became clear that some aspects of the original research completed by Irwin-Williams and her staff were supported by our new research, whereas ­others were not. This is reflected in a number of ­areas where I or other members of the new Salmon research team have departed from the conclusions of the original project. As papers were presented at the Salmon Working Conference in 2004, it became clear that, in addition to interpretive disagreements between old and new Salmon research, many different view­points were represented among the attendees. These differences are apparent in the chapters comprising this volume. Genuine disagreements among the scholars exist for a number of issues, including, among others, the nature of relations between the Middle San Juan and outside areas, the nature of the Chacoan “system” and the outliers, and the function and use of Chacoan great houses. Thus, anyone picking up this volume and expecting to read 20 chapters that present a unified view of Salmon, Aztec, and the Middle San Juan region will be disappointed. This volume does not offer such a consensus. As I hope will become clear as one reads through the book, Middle San Juan Puebloan archaeology is in a state of flux. Many long-dormant issues have been reawakened, and research energy and attention has been refocused on this area in a manner not seen for some 30 years. Therefore, I believe that readers of this book will be rewarded with invigorating, stimulating chapters that convey a wealth of new and revised data on the ancient Puebloan archaeology of the long-overlooked­, yet important Middle San Juan region. The second caveat relates to use of the terms 6

Setting the Stage ­ hacoan and Chacoans. Many of us working in the C Middle San Juan (and beyond) recognize the pitfalls of (1) using the term Chacoan and (2) describing ancient people as Chacoans in our research at Salmon. Chacoan means “derived from Chaco Canyon,” and its use should be limited to a specific series of material culture categories (great house architecture, great kivas, specific ceramic types such as Chaco Black-on-white, and so on). In usage, however, the term has grown and, in popular circles, has taken on an ethnic or tribal dimension. Lay­people often wonder what happened to the Chacoan “tribe.” Recent research shows the danger of considering the Chacoans as a single, distinct ethnic group. Skeletal studies by Akins (1986) and more recently by Schillaci (2003) have revealed the presence of distinct biological populations in Chaco Canyon. Akins’s work also suggests that status differences were present among the Chacoan populations, with some residents (possibly elites) taller and healthier than others (perhaps ­commoners). These findings indicate that considering the “Chacoans” as a homo­ geneous group is problematic and that at least two or more distinct biological and social groups were present in Chaco Canyon. The data suggest that Salmon Pueblo was similarly composed of distinct groups of people: migrants from Chaco Canyon and local, San Juan groups (Reed 2006d). In this volume, then, use of the term Chacoan is not meant to indicate connection to a ­single ethnic or biological population. Instead, use of Chacoan and Chacoans, unless otherwise indicated, refers to specific suites of traits including architecture; projectile points; pottery made with a specific recipe; certain textiles, baskets, and other perishables; other classes of material cultural; and other traits (including the presence of macaws), and the people represented by these traits.

of Spanish exploration and colonization in the sixteenth through mid-nineteenth centuries, we have no written accounts of the pueblo. Roque Madrid’s military campaign of 1705 set out from Santa Fe and traveled west to the Navajo country. This company, however, did not travel beyond the confluence of La Jara Canyon with the San Juan River and thus never reached the vicinity of Salmon (Hendricks and Wilson 1996). The 1776 exploration by the Dominican priests Dominguez and Escalante proceeded west from Santa Fe into Colorado and Utah on a northerly route that missed the San Juan by about 40 miles (Bolton 1950; Warner 1996). The Listers (1987) report that the mapmaker for the Dominguez-Escalante expedition, the famous Bernardo de Miera y Pacheco, mentioned Aztec Ruins, but the Listers doubted, given the expedition’s northerly route, that any of its members actually visited Aztec. Because Aztec lies only 10 miles north of Salmon, we can infer that the DomínguezEscalante Expedition did not visit Salmon either. The earliest documented visit to and account of the Salmon and Aztec sites resulted from an 1859 Corps of Topographical Engineers expedition along the Animas and San Juan rivers, en route to the Green River in Utah (Macomb 1876). Geologist John S. Newberry spent several days studying the ruins that came to be known as Aztec Ruins. He made a number of original observations, including the site’s great size, architectural details, and the amazing preservation in several of the interior rooms. Most importantly, from our perspective, Newberry correctly identified the builders of the ruined pueblos as the ancestors of the modern Pueblo Indians. By examining the details of broken pottery and the architecture of the site, Newberry concluded that continuity existed between the site and the occupied Pueblos of the Southwest. Newberry (1876:109) also made the first recorded visit to and mention of the site on the San Juan River that came to be known as Salmon Ruins: “Near Camp 43 also are the remains of several large structures, of which portions of the walls of stone are still standing, twenty or more feet in height.” Camp 43 of the expedition was identified by Newberry as being a few miles below the mouth of Largo Canyon at the San Juan River. Thus, there

The Middle San Juan Region in Historical Context Archaeological research in the Middle San Juan area has a long history, going back more than 100 years. In the early nineteenth century, several Spanish and Mexican exploration and military campaigns (e.g., Vizcarra in 1823; Brugge 1980) crossed through region.1 However, from the initial period 7

Paul F. Reed can be little doubt that the site described, with standing walls over 20 feet in height, was Salmon Pueblo. Another large, ancient pueblo site identified by Newberry at the confluence of the Animas with the San Juan River was later investigated briefly by Earl Morris in the early 1900s. This site, called by Morris the “Old Fort,” was a Pueblo III period (dating from ad 1150 to 1300) village associated with a long irrigation ditch that ran up the Animas River toward Aztec. This ditch, along with at least two others identified in the vicinity of Aztec Ruins, provides evidence of ancient Pueblo irrigation activities. Accounts of early settlers in the Aztec, Bloomfield, and Farmington areas of New Mexico also mention irrigation ditches, many located in ideal settings for the settlers’ own water needs. Renowned early photographer Timothy O’Sullivan visited the ruins of Salmon Pueblo in 1874 on the Wheeler Survey. In 1874 the survey party traveled up Largo Canyon, to the east of Salmon, and upon reaching the San Juan River, turned west. O’Sullivan was apparently quite impressed with the ruined pueblo. He made eight wet-plate photographs of the site and wrote the following brief account (included as a caption for one of the Salmon photographs included in the Wheeler Survey’s 1879 final report):

in its integrity, centuries ago, inhabited this San Juan region whence they were compelled to migrate by the absence of rainfall and the increasing aridity of their land. A portion of them, it is believed, journeyed to Mexico and built cities there, where they were found by Montezuma, who promised to return for the remaining tribes, and some of the living Pueblos, of New Mexico, who have by some been considered as the abandoned remnant of his people, still believe, or profess to do so, in the ultimate fulfillment of that promise. (O’Sullivan 1879: caption for Plate 59) O’Sullivan was the first but certainly not the last observer to comment on the high quality of the stone masonry at Salmon (Figure 1.2). The fine masonry and construction techniques used at Salmon indicate the presence of Chacoan architects and builders as advisors, if not actual colonists. O’Sullivan touched upon the “mystery” that would confound explorers and archaeologists for decades to come when he wrote that “the present tribes know nothing, not even in tradition.” The Native Americans encountered by all explorers to the greater San Juan region after abandonment of the area by Puebloan peoples at ad 1300 were culturally distinct from the builders of the great pueblos. Although groups in the area—the Navajo, Jicarilla Apache, and Ute (of various bands)—had oral traditions about the abandoned sites, they were not descended from the builders. Thus, O’Sullivan and other early explorers obtained only limited information from Native American scouts belonging to these tribes. A common element, however, of early accounts is the “Montezuma” or “Aztec” legend. Briefly, the legend attributes ruined pueblo sites across the northern Southwest to the Aztecs of Central Mexico during the course of their mythological wanderings. In fact, this legend resulted in the inaccurate name given to Aztec Ruins in the late nineteenth century. The historic record indicates that most early explorers heard some variation of this legend from Native American guides. Lt. Rogers Birnie was a member of the same 1874 Wheeler Survey during which O’Sullivan took historic photos of Salmon. Birnie apparently

In the northwestern part of which Territory these remains were discovered in the expedition of 1874. They are situated on the north bank of the San Juan River, about 15 miles west of the Cañon Largo. The ruin is about 350 feet square, and is built of stone in its natural shape, joined by a mud cement. Observe the regularity of the courses of masonry, and the resemblance of the design to the English bond of our present style of architecture, the large blocks of stone corresponding to the stretchers which, in our structures, retain the walls longitudinally. The measuring scale of 36 inches, to be seen in one angle of the room, is a criterion by which to estimate its dimensions. This ruin is characteristic of an ancient people and civilization of which the present tribes know nothing, not even in tradition. There is a belief that the Aztec race 8

Setting the Stage

Figure 1.2.  Photograph of Salmon Pueblo taken by Timothy O’Sullivan during the Wheeler Survey in October 1874. Note the fine Chacoan masonry in the wall and the collapsed roof in the foreground.

observed Salmon Pueblo, but he wrote nothing specific about the site (identified as “Pueblo San Juan” by O’Sullivan) for the report. Instead, Birnie contributed a sizeable narrative on what came to be known as Aztec Ruins, quoted here in part. Although his report is not the first on Aztec, it is very detailed, and he seems to have captured the essence of the Pueblo well. Birnie did not identify the pueblo he was describing by name, but there can be no doubt that his description is of Aztec Ruins. Birnie describes both the East and West ruins at Aztec, and he found entry into a well-preserved room in the East Ruin. He provided a description of a typical Pueblo roof construction with two or three big pine or spruce beams (vigas) laid first across a room, followed by a series of smallerdiameter­pieces (latillas), then a layer of split wood and bark, and finally earth (dirt and mud). Birnie noted numerous inscriptions in the plaster and on the rock walls of the room—the first observation of what was revealed to be many inscriptions left

over a large period of time at Aztec. Birnie also provided the first brief description of the unusual circular (tri-wall) structure located directly between the East and West ruins. Curiously, Birnie references one of O’Sullivan’s photographs of Salmon (cited as Plate IX in the report) to illustrate his description of the height of the walls at Aztec. It is unclear why Birnie confused the two sites, as his description of Aztec is correct to the level of minute details about the layout of the site and the East and West ruins. O’Sullivan apparently did not take any photographs of Aztec, and it is possible that Birnie did not visit Salmon. Thus, he might have assumed that the photograph in question was taken at Aztec. Like most of the early explorers, Birnie correctly surmised that Aztec was a site of considerable antiquity: The most extensive ruins met with were on the right bank of the Las Animas River, about twelve miles above its junction with the San 9

Paul F. Reed Juan. I had been previously informed of this, my informant stating that he had counted 517 rooms in one pueblo. On visiting the ruins we found what had once been, apparently, quite a town, with two main buildings and numerous small ones about them. One of the main buildings, situated nearest the river, extended to and was built into a bluff separated a few hundred yards from the river by a flat. The plan was rectangular with a small court on the south side, the court flanked on either side by two circular rooms or towers at the corners of the building; two more of these rooms at the other corners, and three through the center and parallel to the longer side of the building; the walls supporting the towers on either side of the court were square-cornered but had re-entrant angles. In many places along the San Juan River, pieces of old crockery were observed and remains of several small stone houses. In one of these I found a very fine specimen of a stone hammer, oval and of natural shape, with the ordinary groove cut about it for attaching the handle. A number of important ruins were also observed along the Cañon de Chaco. None of these so minutely described by Lieut. Simpson in 1849 were visited by us, as we did follow his route only perhaps a very short distance. The Navajo Indians ascribed some of the figures and signs seen in the lower room of the ruins to Apaches and Comanches; but their explanations were very vague, principally from the difficulty of understanding them. (Birnie 1875)

region was found. It resembled, in the arrangement of its houses and the “castle” in the center, the city and ruins discovered on the Animas [Aztec Ruins]. The river has washed away a portion of the stone building, and on its banks a number of interesting facts were observed. Fire seems to have destroyed at least part of the building. Gradually the river deposited about 10 feet of sand and silt in the court-yard and in the northern chambers, which it probably reached through the windows or doors. At that depth below the surface a layer of what at first appeared to be charcoal was observed, 2 to 5 inches in thickness. Upon examination, however, this proved to be an Indian corn, still unhusked, but completely charred. Probably the chamber thus cut by the river, which exposes its section, was used as a granary. Beyond that, along the same vertical bank of sand, innumerable fragments of pottery, bones of deer, of rabbits, and what appeared to be sheep, were found. Had it been possible to spare more time, I am confident that excavations at that locality would have developed many interesting facts. The time that it must have taken the river to wash away one-third of the building, which probably was not erected immediately upon its banks, must have been considerable, but in spite of such evidence of this, it has been impossible for us to assign, with any semblance of correctness, any definite age to these and other ruins. An estimate, little better than a guess, may claim for them an age, as ruins, not much exceeding 300 years. (Endlich 1877:179)

Geologist Frederic Endlich made the next recorded visit to the Middle San Juan region in 1875, as part of the Hayden Survey (officially known as the United States Geological and Geographical Survey of the Territories; Endlich 1877). Endlich traveled first to Aztec Ruins along the Animas River and provided a good description of the remains. Next he visited the site on the San Juan River that came to be known as Salmon Ruin and wrote the following description:

A number of Endlich’s observations bear comment. He described the “castle” in the center of the Pueblo as similar to one at Aztec Ruins. These features are tower, or elevated, kivas built directly into the rooms of the pueblos. Evidence of the devastating fire that occurred at the abandonment of Salmon was obvious to Endlich, as was the in-filling of the site with river deposits. Endlich’s discovery of burned, unhusked corn presaged the incredible amounts of preserved perishable artifacts (corn, basketry, sandals, cordage, and textiles) recovered during excavation of Salmon Ruin in the 1970s (see

South of station 54 on the San Juan another ruined town of the ancient inhabitants of that 10

Setting the Stage Webster, this volume). Given the absence of domestic sheep in the aboriginal Southwest, the large animal bones he described as being from sheep were probably those of deer, antelope, or elk. Finally, Endlich’s “guess” of the site’s antiquity—300 years old, or dating to about ad 1575—was too young. Undoubtedly, the great preservation of materials at Salmon, even in surface contexts, and the presence of burned corn and other materials at the site led to the underestimate of the site’s true age. Following Endlich’s visit to the region, the next historical reference we have to Salmon Ruins is William Henry Jackson’s 1877 map, prepared for the Hayden Survey. Jackson’s map shows a large ruin along the San Juan River, in the correct position for Salmon ( Jackson 1878). It is not clear if Jackson actually visited Salmon, but he was aware of the site because of Endlich’s visit in 1875. Also on the Hayden Survey in 1875 was William Henry Holmes, who visited the site on the La Plata River that would come to bear his name (the ­Holmes Group), wrote a short description, and drew a map (Holmes 1878). Holmes (1878:388) adeptly perceived one of the primary differences between large sites on the La Plata River and many of their counterparts elsewhere in the Middle San Juan region, noting “the buildings have been isolated and, in a measure, independent of each other, differing in this respect from most of the groups of ruins further south and west.” As part of the Powell Survey (the U.S. Geographical and Geological Survey of the Rocky Mountain Region) of the late 1870s and early 1880s, Lewis Henry Morgan undertook a study of the houses and lives of the people he described as American aborigines. Morgan was an attorney with considerable interest in the anthropological study of American Indians. In his historical synthesis of human society, Ancient Society (1881a), he laid out a unilineal theory of human sociocultural evolution through three broad stages: savagery, barbarism, and civilization. The Pueblo Village Indians were classified by Morgan into his Middle Barbarism stage. Although his sequence constituted a significant contribution to the anthropology of the late nineteenth century, archaeologists of ­today reject the notion that all societies pass through a prede-

termined set of cultural stages. Today the individual development of the numerous, distinct human societies across time and space is viewed as a very complicated process. In his separate work on house types (Morgan 1881b), as excerpted below, Morgan devoted considerable time to the Iroquois Confederacy of the eastern United States and to the Aztecs of “Old Mexico.” He also committed two chapters in his monograph to the Pueblo Indians of the southwestern United States. A careful reading of his work suggests that Morgan never visited Chaco Canyon. His descriptions closely follow those of W. H. Jackson and Lt. James Simpson, who are quoted extensively, and the report reproduces maps from those two earlier reports. In contrast, Morgan is explicit regarding his visit to Aztec Ruins on July 22, 1878. Anticipating later conclusions about Chaco Canyon, Morgan suggested that “Chaco must have possessed remarkable advantages for subsistence.” He noted the period of highest stream flow in Chaco Wash occurred in July (based on Lt. Simpson’s trip) and inferred an abundance of water for agriculture during the critical summer months. More recent research, such as that of archaeologist Gwinn Vivian, has documented extensive systems in Chaco Canyon for the collection and transport of runoff water from summer rainstorms. Further, Chaco Canyon has recently been assessed by Vivian and other archaeologists as a superior location within the San Juan Basin for agricultural pursuits. Aztec Ruins was the primary focus of Morgan’s work among the ruins of the Pueblo Indians, and he made a number of astute observations. He identified five stories along Aztec’s back wall, and four stories running down each wing. Later explorers would document no more than three stories at Aztec. Earlier explorers, such as Newberry and Birnie, did not specifically describe the number of stories. Morgan explicitly documented the use of Aztec Ruins as a stone quarry by local residents, and a settler told him that perhaps one-quarter of the rock from both the East and West pueblos at Aztec had been removed. Certainly, Aztec Ruins had seen significant impact from collection of building stone and random vandalism by the late 1800s. It is possible that these impacts had removed upper stories 11

Paul F. Reed from Aztec’s back wall, the location in which Morgan had commented on extensive rubble piles, and where he indicated the fifth story was located. In assessing the origin of the Pueblo structures of the San Juan, Morgan developed a broad interpretation, well beyond that of other researchers of his day. In the Pueblo villages of the ancient San Juan, Morgan found the roots of all village life in North America, and perhaps of South America as well. He believed that the builders of the pueblos had migrated east, founding the societies of the Mound Builders, and south, to build the great Aztec and Mayan cities and civilizations of Mesoamerica. Morgan believed that the ancient Pueblo Indians of the American Southwest were the originators of the advanced sedentary culture (Morgan’s “Middle Status of barbarism”) that eventually spread to most of the New World. Archaeologists now view this idea as naive and overly simplistic since it is clear that societies developed independently in many areas, although ideas and people did circulate between them. Morgan described Aztec as

of these were carried up two stories, the third row three stories, and so on to the number of five stories for the main building and four for each wing. The external wall rose forty or fifty feet where the structure was five stories high and but ten feet on the court side, including a low parapet wall, where the structures was but one story high. Those familiar with the remains of Indian Pueblos in ruins will recognize at once the resemblance between this pueblo and the stone pueblos in ruins on the Rio Chaco, in New Mexico, about sixty miles distant from these ruins, particularly the one called Hungo Pavie, so fully described by General J. H. Simpson. There is one particular in which the masonry agrees, viz., in the use of course of thin stones, about half an inch in thickness, sometimes three together, and sometimes five or six. There courses are carried along the wall from one side to other, but often broken in upon. The effect is quite pretty. These stones measure six inches in length by one-half an inch in thickness. General Simpson found the same course of thin stones, and even thinner, in the Chaco ruins, and comments upon the pleasing effect they produced. I wish to call attention again to the San Juan district, to its numerous ruins and to its importance as an early seat of Village Indian life. These ruins and those of similar character in the valley of the Chaco, together with the numerous remains of structures of sandstone, of cobblestone, and adobe in the San Juan Valley, in the Pine River Valley, in the La Plata Valley, in the Animas River Valley, in the Montezuma Valley, on the Hovenweep, and on the Rio Dolores, suggest the probability that the remarkable area within the drainage of the San Juan River and its tributaries has held a prominent place in the first and most ancient development of Village Indian life in America. The evidence of Indian occupation and cultivation throughout the greater part of this area is sufficient to suggest the hypothesis that the Indian here first attained to the condition of Middle Status of barbarism, and sent forth the migrating band who carried this advanced culture to the Mis-

one of four [pueblos] situated within the extent of one mile on the west side of the Animas River, in New Mexico, about twelve miles above its mouth. Besides these four, there are five other smaller ruins of inferior structures within the same area. This pueblo [Aztec] was five or perhaps six stories high, consisting of a main building three hundred and sixty-eight feet long, and two wings two hundred and seventy feet long, measured along the external wall on the right and left sides, and one hundred and ninety-nine feet measured along the inside from the end back to the main building. A fourth structure crosses from the end of the wing to the end of the other, thus inclosing an open court. It was of the width of one and perhaps two rows of apartments, and slightly convex outward, which enlarged somewhat the size of the court. The main building and the wings were built in the so-called terrace form; that is to say, the first row of apartments in the main building and in each wing on the court size were but one story high. The second row back 12

Setting the Stage sissippi Valley, to Mexico, and Central America, and not unlikely to South America as well. (Morgan 1881b:154–197 passim)

interest in archaeology, toured the San Juan drainage in the Four Corners area. He visited the two largest pueblos in the San Juan, Aztec and Salmon, and made some astute observations at both locales. Prudden made note of some damage to the site from “random digging” and the construction of the Salmon irrigation ditch. He believed, however, that the site had great potential if it could be “speedily” excavated. Full excavation of Salmon Pueblo to realize the potential Prudden observed finally began in the early 1970s under the direction of Cynthia Irwin-Williams. Prudden’s observations on Aztec Ruins were also quite insightful. He noted that while the site had been vandalized, it was nevertheless newly protected by an owner (Mr. Koontz) who wanted to have the site scientifically studied. Such a study began in 1916, when Earl Morris and his crews began work for the American Museum of Natural History.

The next documented visit to Salmon Ruins came in 1897, with a party led by Warren Moorehead (Moorehead 1906). Moorehead was touring the Southwest on a collecting expedition for the Phillips Academy of Andover, Massachusetts, and he wrote the following account: In April, 1897, the writer left Farmington, New Mexico, with nine men, a large wagon, and five horses, bound for the Chaco Group ruins, seventy miles south.... The party returned from the Chaco Group to Farmington, and the [Chaco] specimens were packed ready for shipment to Mr. Peabody. We visited the Salmon ruins on the San Juan river, about 20 miles from Farmington, finding there two small buildings that had been destroyed by fire. Mr. [George] Salmon did not wish us to carry on explorations for greater length of time than three days, and we were unable to secure more than a few specimens. The site is interesting and merits thorough exploration. Such whole pottery as we obtained from this San Juan pueblo is as finely made as any found in the Southwest. Apparently, the pueblo was attacked, sacked, and burned while it was still occupied. By whom, may be determined after careful exploration. (Moorehead 1906:33, 53)

The only large ruin in the entire San Juan Valley stands upon a low bench at the edge of the alluvial bottom a few miles below Bloomfield. This is called locally “Solomon’s Ruin” after the name of the owner of the land on which it is situated. It is built largely of dressed stone, the walls in some places resting upon a foundation of small boulders. It was several stories in height, and contained many rooms, but is now so fallen and covered with sand and earth that the plan can be only partially made out. This ruin measures about five hundred feet along the back and is of the communal pueblo type like the ruin at Aztec on the Animas and the great pueblos of the Chaco. Recently considerable random digging has been done in the search for pottery, and water from an irrigating ditch has been turned into the ruin, undermining it in several places. Thus the existence of several rooms has been revealed whose walls beneath the covering of fallen stones and soil appear from without to be largely intact, the well preserved timbers above them being still in place. In spite of the vandalism which had its way with this ruin, there appear to be still promise of interesting results

Moorehead’s account is interesting for two reasons. First, he noted finding two small buildings destroyed by fire. Although Salmon Pueblo was burned at the close of its occupation, the description of two small buildings seems incongruous given the size of the overall site. Second, Moorehead was apparently frustrated by not being allowed to dig for more than three days. His account thus corroborates Salmon family oral history that highlights the determination of George Salmon to preserve the site and not allow it to be plundered for specimens. Just after the turn of the twentieth century, T. Mitchell Prudden, a medical doctor with great 13

Paul F. Reed should a proper investigation of what remains be speedily undertaken. Farther up, near the village of Aztec, on a low gravel bench west of the river, lies the group of large pueblos called the “Aztec Ruin.” The largest and best preserved of these was over three hundred and fifty feet long at the back and several stories in height with a court facing eastward. Near by are several large stone and earth heaps, indicating older sites. A large mound near the edge of the low bench bordering the valley bottom gives superficial evidence of many burials. It is said that in early days this ruin was used as a stone quarry by neighboring settlers. It is now on private property, and the owner, Mr. Koontz, wisely appreciative of the importance of systematic study of these relics of the elder folk, has guarded them from the onslaught of the vandals, so that here one of the most promising of the great old pueblos lies waiting for trained and authorized explorers. A small opening has been made in one corner of the ruin, through several rooms may be entered in succession. These are practically intact, with the ceiling timbers in place and well preserved. The exact size and form of this ruin are not evident in its present condition. Within sight of this ruin, in valley bottom, several small sites may be located. (Prudden 1903:1–65, passim)

and cultural context. He identified Aztec as a member of the group of Chaco Canyon sites, but commented on differences in layout and ground plan between Pueblo Bonito and Aztec. He quite correctly observed that Aztec more closely resembles Chetro Ketl. The basic E shape of Chetro Ketl— with an encircling row of rooms closing the open part of the E, forming then a D shape—is the standard Chacoan ground plan for all sites built between ad 1000 and 1100. Ground plan aside, the comparison Morris made between Aztec and Pueblo Bonito is apropos. He noted similarities in massive construction, the sheer rise of the back walls, and the open courts or plazas in front, which held great kivas. These characteristics, along with distinctive and specialized construction and use of masonry, set Chacoanconstructed sites apart from all others of contemporary age. At the time his first report on Aztec was finished in 1919, Morris had spent only a few years professionally excavating sites. His Aztec reports, this first one in particular, reflect this relative inexperience with the broad range of Anasazi-Pueblo culture. In particular, Morris’s emphasis on ­simple description of artifacts belies the insights he would later bring to an understanding of changes in Puebloan culture. By 1939, Morris had completed a large report on excavations in the La Plata Valley (west of Aztec but part of the Middle San Juan area), and he had much greater insight into the regional archaeology. Nevertheless, his 1919 report on Aztec is quoted below.

Hearing the name given to the site by local ­people, Prudden interpreted it as “Solomon.” His misat­ tribution stems from the difficulty of the name Salmon and the propensity to pronounce it like the name of the fish. The Salmon family name (and thus the name of Salmon Pueblo) is correctly pronounced with the l included, thus “sal mən” is correct. Prudden’s description is largely accurate, and his concern about impact to the site through random digging and water damage is well noted. Fortunately, damage to the site was largely curtailed in the years after Prudden’s visit, leaving the site in an intact state to await excavation by Cynthia IrwinWilliams and her compadres in the early 1970s. In the conclusion to his first report on Aztec Ruins, excerpted below, Earl Morris attempted to put the West Pueblo at Aztec into a chronological

Inasmuch as the exploration is only partially completed, ultimate conclusions cannot be drawn at the present time, but a few definite statements may be made without too great risk of error. Architecturally, the Aztec Ruin must be classed with Pueblo Bonito and the other members of the Chaco Canyon group. The similarity is close throughout. To be sure, there is a marked difference between the outline of the groundplan of the Aztec Ruin and that of Pueblo Bonito, but this difference is no greater than that observable between Pueblo Bonito and Pueblo Chettrokettle, a compo14

Setting the Stage

Figure 1.3.  Photograph of Salmon Pueblo taken by Timothy O’Sullivan during the Wheeler Survey in October 1874. The prominent wooden beams extending from the roomblock in the top of the ­photograph were probably sampled by Jeancon in 1923 during the First Beam Expedition.

nent and integral member of the Chaco group. If the preservation of Pueblo Chettrokettle had been such that it merited exploitation as the outstanding member of the Chaco group, and if it had in consequence become fixed in our minds as such, the similarity between it and Aztec Ruin would have been readily apparent to the most casual observer, for the groundplans of the two are practically identical, except that Chettrokettle covers a somewhat more ­extended area. Although the main building of Pueblo Bonito is arc-shaped, while that at Aztec forms three sides of a rectangle, the basic concept of the structures is the same. Each rises with sheer outer walls enclosing an open court which was protected from invasion on the south or sunward side by a low one-storied tier of rooms. (Morris 1919:104–105)

son 1935). These expeditions were dispatched all over the Southwest to collect the wood necessary to build University of Arizona astronomer A. E. Douglass’s data set for dendrochronology. According to his field notes, Jeancon collected ­samples from a northeast corner room at “Solomon” (Salmon). I have inferred that the room sampled is probably the same third-story structure shown in one of O’Sullivan’s 1874 photographs with a long wooden beam visible (Figure 1.3). Documented visits to Salmon Ruins by archaeologists apparently declined after the 1920s. It seems unlikely that the site was completely ignored or overlooked, yet little or no mention is made of it in publications on the Southwest and Anasazi archaeology from this time period. Most surprising, perhaps, is the lack of mention of Salmon by archaeologist Earl Morris, who covered an incredible amount of territory in the San Juan and across the Four Corners. Nowhere in any of his formal writing or journals did Morris write about Salmon Ruins. Part of the explanation perhaps lies in Morris’s evolving interests though time. Although he

Jean Jeancon came to Salmon in 1923 on the National Geographic Society–sponsored First Beam Expedition to collect tree-ring samples (Peter15

Paul F. Reed this brief review, I want to discuss more recent archaeological research over the last 30 years in the region. Rather than exhaustively noting every project, I ­focus on a few large projects of significance for the Chacoan and post-Chacoan occupations in the Middle San Juan region. In 1970, Cynthia Irwin-Williams initiated the San Juan Valley Archaeological Program (SJVAP), primarily focusing on intensive excavations at Salmon Pueblo (Irwin-Williams 1972; ­Irwin-Williams­and Shelley 1980; P. Reed 2006a). In addition, the project included an ambitious schedule of survey and testing of sites in the three valleys of the Middle San Juan region: the Animas, La Plata, and San Juan. Over five field seasons, more than 200 sites were identified and recorded, and testing was done at roughly a dozen locales. Although some progress reports were completed (e.g., Irwin-Williams 1976), the results were largely unreported. Nevertheless, two projects were completed using SJVAP survey and testing data. In her master’s thesis using SJVAP ceramic data, Whalley (1980) concluded that Chacoan ceramics circulated among great houses and small sites in a similar fashion. McKenna (1976) also studied ceramic assemblages, analyzing data from eight sites tested during the SJVAP 1974 field season. In addition to the Salmon-SJVAP project in the 1970s, several large projects of importance have been completed. Two large survey projects fall into the same basic category: the Anasazi Community study (Marshall et al. 1979) and the Outlier Survey (Powers et al. 1983). These two projects, sponsored by different agencies, had essentially the same aim: to identify large Puebloan communities across the San Juan Basin linked to Chaco Canyon. The Outlier Survey focused primarily on the area north of Chaco, whereas the Anasazi Community project explored sites mostly to the south. Both surveys compiled lists of “big” sites, potential Chacoan outliers, in the Middle San Juan region. Together, these projects greatly expanded the known sample of Chacoan outliers and added to the momentum of the Salmon Project and the National Park Service’s Chaco Project ( Judge 1984), bringing about a new era of Chacoan studies. In the 1980s, the Middle San Juan saw another

had worked on large Pueblo sites such as Aztec and the La Plata great houses (Morris sites 39 and 41), later in his career Morris was primarily interested in the early Basketmaker periods. Furthermore, Morris primarily focused on the Animas and La Plata drainages, and the San Juan River in the immediate vicinity of the town of Farmington. Other factors in the site’s relative anonymity during the middle decades of the twentieth century include the thick vegetation that came to cover the site because of the irrigation-based rise of the ­water table on the Salmon property and the collapse of the standing walls photographed by O’Sullivan. These factors, along with its private ownership, combined to reduce the site’s visibility. In the early 1960s, archaeologist Emma Lou Davis, undertaking a doctorate at UCLA, visited Salmon (Davis 1964). Davis was tracking Mesa Verde migrations across the northern Southwest, and she conducted surveys in several locales. Davis’s recording of the Salmon site was perhaps the first formal documentation of the pueblo. She was so impressed with Salmon and the San Juan and its associated valleys that she speculated on their importance for Pueblo-Anasazi culture: It is possible that the San Juan valleys were the regional center of Pueblo III culture climax, to which both Chaco and Mesa Verde were peripheral. This river basin was well situated to be both a recipient and a clearing ground for diffusion and larger population shifts from west, north, and south. In addition, it contained miles of bottom land to provide an economic base on which the beginnings of urbanism and high culture could flourish. (Davis 1964:221–222) In many ways, Davis anticipated the research that Cynthia Irwin-Williams would conduct a decade later—in particular, her emphasis on what we now call the Middle San Juan region.

A Brief Look at Recent Research in the Middle San Juan Region As the foregoing discussion illustrates, the Middle San Juan region has been the focus of exploration and research for more than 130 years. Beyond 16

Setting the Stage infusion of research, with several survey and reconnaissance projects that contributed to our understanding of the region’s archaeology. Dykeman and Langenfeld (1987) studied several locales in the La Plata Valley and provided a preliminary overview of its complicated archaeology. This work followed a century of research by early archaeologists such as William Henry Holmes, Deric Nusbaum, and Earl Morris. In the late 1980s, Stein and McKenna (1988) completed a similar project in the Animas Valley, focusing on the immediate vicinity of Aztec Ruins. Stein and McKenna’s survey and reconnaissance identified many great houses, great kivas, and associated small sites. The late 1980s brought two excavation projects to the Middle San Juan. The University of New Mexico’s Office of Contract Archeology worked on the B-Square Ranch along the San Juan River south of Farmington (Hogan and Sebastian 1991), excavating a small pueblo associated with the Jaquez great house and community (see Wheelbarger, this volume). The site produced good quantities of data revealing two occupations that overlapped with the Chacoan and San Juan periods at Salmon and Aztec. In the La Plata Valley, the Museum of New Mexico’s Office of Archaeological Studies undertook extensive excavations at a number of sites along New Mexico Highway 170 (the La Plata Highway). To date, one volume of eight has been produced for this important project (Martin et al. 2001), although several short papers and articles have been written (e.g., Toll and Wilson 2000; see also Toll, this volume). A summary of Chacoan and post-Chacoan great houses in the Middle San Juan (or Totah) region by Toll and McKenna (1992) provided a good update on our understanding of the region as of the late 1980s. Twenty years later, as the chapters in this volume attest, we have advanced our knowledge in several areas but still have many questions. Much of the other recent work in the Middle San Juan region, including the Center for Desert Archaeology’s Salmon Research Initiative and Aztec Ruins’ continuing research, is discussed in the pages that follow. One additional study of importance is Glowacki’s (2006) recent dissertation using INAA (instrumental neutron activation

analysis) to study ceramic distributions and population. Although Glowacki’s focus was primarily on the Northern San Juan region, she included several Middle San Juan sites. Her work confirmed the ­local approach to ceramic production and exchange that several earlier researchers had noted (e.g., L. Reed 2006c, this volume; Shepard 1939).

Chronological Issues in the Middle San Juan Region Salmon Pueblo Chronology

Chronological refinement was one of the primary goals of the Center’s Salmon Research Initiative and we made important modifications to the Salmon sequence (P. Reed 2006b). More than 500 pieces of wood were assessed for tree-ring dating, and 200 new samples were analyzed (P. Reed 2006b; Windes and Bacha 2006, this volume). Jeff Eighmy (2006) was engaged to reassess Salmon’s archaeomagnetic data. These new data and interpretations have changed our interpretation of Salmon’s chronology and history. In the 1970s, Irwin-Williams and her staff developed the Salmon Ruins chronology using three dating techniques: dendrochronology, archaeomagnetism, and ceramic cross-dating. The Salmon chronology (presented in the 1980 report by Rex Adams), defined three periods of occupation: Primary (ad 1088–1130), Intermediate (ad 1130–1185), and Secondary (ad 1185–late 1200s; see Table 1.2). Tree-ring dates from the site supported the initiation of the Primary or Chacoan period in the late 1080s and showed site occupation into the mid-tolate 1200s; the last dated event at Salmon was the reroofing and probable reconstruction of the Great Kiva in the spring and summer of 1263 (see Windes and Bacha, this volume, for greater discussion of tree-ring dates). Original archaeomagnetic samples collected and processed by Robert DuBois in the early days of the technique provided data for the inferred Intermediate period, which was bracketed by archaeomagnetic date means at 1130 and 1185. Ceramics were also used to support the concept of the Intermediate occupation, particularly the type McElmo Black-on-white, which was believed by Salmon researchers to be a prime indicator of mid1100s use of Salmon. Lastly, the Secondary period

17

Paul F. Reed Table 1.2.  Chronological Sequences used in the Middle San Juan Region Pecos Classification Pueblo II 900–1100

Pueblo III 1100–1300

Original Salmon Sequence

Center’s New Salmon and Middle San Juan Chronology

Chaco Project Sequence Adapted for Aztec Ruins

—

—

Primary (Chacoan) 1088–1130 Intermediate 1130–1185 Secondary (Mesa Verdean) 1185–1300

Chacoan 1060–1125 Early San Juan 1125–1190 Late San Juan 1190–1300

Classic Bonito 1040–1085 Late Bonito 1085–1140 McElmo 1140–1200 Mesa Verde 1200–1300

was dated from 1085 to the late 1200s using a handful of tree-ring dates and ceramic evidence. Part of the new research program involved reassessing and modifying the Salmon chronology. First I examined the original tree-ring data and subsequent sampling that Tom Windes and his associates undertook (Windes and Bacha 2006, this volume), which largely confirmed Salmon’s Primary or Chacoan sequence as defined by Adams and ­Irwin-Williams (R. Adams 1980). However, I refined the original sequence, concluding that Salmon was largely constructed in two phases: (1) during the first episode, the wood used had been cut between 1086 and 1090, culminating in construction of at least the first story (and perhaps rooms on the second and third floors) of the pueblo in the spring and summer of 1090; (2) during a second construction episode, the wood used had been cut between 1091 and 1094. The second construction phase involved rooms spread across the pueblo and probably focused on completion of Salmon’s second and third stories. Another construction episode occurred at 1105–1106, involving repairs at rooms scattered across the pueblo. The final Chacoan construction event at 1118 (with beams cut from 1116 to 1118) was the repair of the east wall of Room 62W. Based on the tree-ring dates from Room 62W and other data, I inferred that the Chacoan period at Salmon ended in the late 1110s or early 1120s, prior to the 1130 termination point favored by Adams and Irwin-Williams. As noted above, Salmon’s Intermediate ­period

Generic Sequence Chacoan 1050–1130

Post-Chacoan 1130–1300

was defined largely on the basis of archaeomagnetic dates. Jeff Eighmy (2006) reassessed that data using Robert DuBois’s original data sheets and correspondence from the 1970s. Eighmy replotted the result on the current version of the Southwest archaeomagnetic master curve (SWCV2000; Lengyel and Eighmy 2002). Eighmy’s (2006) reanalysis identified several problems. First, he was unable to confirm the date ranges and the level of precision DuBois obtained for several samples. DuBois dated five samples to Salmon’s Chacoan occupation; Eighmy’s work could not confirm that these samples dated to that interval. For several of the samples, Eighmy’s reanalysis suggested that the hearth features in question dated to the Secondary period, not the Chacoan. Second, for the six samples that DuBois assigned to the Secondary occupation with precise ranges (no larger than 80 years), Eighmy’s analysis produced much larger ­error ranges (for example, 975–1400), so large as to be unusable for chronological refinement. Finally, Eighmy plotted several DuBois samples ­using the 1975 DuBois curve and found that the dates DuBois reported could not be derived from the curve. In short, Eighmy’s reassessment casts serious doubt on the original Salmon archaeomagnetic date ranges provided by DuBois.2 These findings indicate that the Intermediate period, as defined using archaeomagnetic data, is not well supported. Despite these problems, I am not suggesting that Salmon was abandoned or unoccupied after the end of the Chacoan period in the 18

Setting the Stage 1120s. In contrast, data from several rooms indicate that the site was occupied throughout the mid-tolate 1100s. To emphasize the continuity that our new research revealed in the post-Chacoan period, however, I redesignated the mid-1100s as the Early San Juan period (1125–1190). Use of the term Intermediate suggests that a distinct group of people moved into Salmon Pueblo at the end of the Chacoan period, a finding not supported by the data. Instead, I have inferred that the original, local San Juan residents recruited by the Chacoans to be part of the original Salmon residential core remained in the pueblo at the end of the Chacoan occupation. This group (or groups) occupied Salmon throughout the remainder of the 1100s. In the late 1100s and early 1200s, the population began to increase, peaking in the Late San Juan period after ad 1240 and continuing through site abandonment in the 1280s. This later period at Salmon was designated the Secondary or Mesa Verde period by IrwinWilliams­and her colleagues. I have chosen to drop the “Mesa Verde” name in favor of “Late San Juan” for two reasons. First, because the post-1190 period at Salmon represents continuity from the Early San Juan period, it is desirable to links these periods by name. Second, new research at Salmon has revealed little in the way of direct connection to the Mesa Verde area. It is true that a limited quantity of ceramics (no more than 10 percent; L. Reed, this volume) was imported from the north, and that the basic architectural characteristics at Salmon certainly reflect patterns widespread in the 1200s across the entire northern Pueblo world. Nevertheless, in our new research, we did not find evidence supporting the inferred Mesa Verde migration proposed by ­Irwin-Williams during the original Salmon Project. In summary, my revised Salmon chronology suggests the following occupational history. Construction and occupation of Salmon commenced at ad 1090, with four construction episodes over the subsequent 25 years (see Table 1.2). Chacoan occupation continued into the 1120s, when changes to the site’s architecture and layout began to occur. Subsequently, we see evidence of post-Chacoan occupation by local San Juan Valley inhabitants (re-

cruited as part of the original Chacoan construction) at a modest level through the mid and late 1100s. In the late 1100s or early 1200s, occupation and usage of Salmon increased as local population in the Middle San Juan boomed and outside migrants arrived. By 1240, the site probably reached its peak of occupation, with perhaps 300 residents. Occupation continued with a major construction project, the reroofing (and remodeling) of the Great Kiva at 1263. Finally, a catastrophic fire ended Salmon’s use-life in the 1280s or 1290s.

Regional Chronology

With an understanding of Salmon’s revised chronology, we can turn to issues related to the larger Middle San Juan region. Table 1.2 shows several different chronological sequences that archaeologists use in the region, along with the basic Pecos Classification (the Pueblo II and III periods). After the original Salmon sequence and the revised Center chronology (discussed above), the next sequence shown in Table 1.2 comes from the Chaco Project (see Lekson 2006a) and is presented as used by Gary Brown and his associates (this volume) for the Aztec Community. The last “Generic” sequence simply indicates two periods: Chacoan and post-Chacoan. An initial assessment of the varying chronological schemes might suggest chaos and difficulty in finding common terminological ground. It is true that the three main sequences crosscut the Pecos Classification, and each identifies unique beginning and ending dates. Nevertheless, a closer examination reveals general and specific (in some cases) agreement between the chronologies. Few contemporary archaeologists believe that A. V. Kidder’s 1927 Pecos Classification adequately divides the ancient Puebloan sequence; in particular, Kidder’s ad 1100 transition between Pueblo II and III is not supported by the northern Puebloan archaeological record. Rather, many Chacoan great houses and other sites show continuous occupation after 1100. Thus, the terminal dates for the Chacoan period are identified as 1125, 1130, and 1140 in the sequences shown in Table 1.2. Similar variation is seen in the starting dates given to the Chacoan period in the Middle San Juan: 1040, 1050, and 1060. 19

Paul F. Reed The ­Center’s chronology, for example, uses 1060 as the beginning date because there is little or no evidence of Chacoan presence in the Middle San Juan prior to this time. In contrast, in the Aztec chronology, the beginning date of 1040 for the Classic Bonito period derives from Chaco Canyon and is not specific to the Middle San Juan region. In his concluding comments to this volume, Gwinn Vivian suggests that further progress in addressing regional research issues may be hindered unless a single chronological sequence is embraced by archaeologists working in the area. Clearly, the discussion above reveals considerable variation in phase names and starting/ending dates for periods of interest in the Middle San Juan. I would offer several points for consideration. First, many other geographic areas studied by archaeologists suffer from this same problem: multiple, competing chronologies or phase sequences. In fact, Chaco Canyon, the source of the Aztec Ruins chronology, falls into this category, with competing sequences proposed by Vivian (1990) and the Chaco Project (Lekson 2006a), among others. The many complexities of Chacoan research notwithstanding, I think it is fair to say that progress has not been slowed by variation in chronological sequences. Second, whereas many of the sites in the Middle San Juan region follow the same general pattern of occupation—​Chacoan initiation in the late 1000s, transition to local groups in the mid-1100s, and renewed occupation in the 1200s—each site nonetheless has a unique history. Thus, Salmon’s transition to a ­local, post-Chacoan focus in the 1120s was not duplicated at the Aztec pueblos. Rather, Brown and his associates (this volume) document the transition to “McElmo” times at Aztec after 1140. Clearly, we can bracket the transition to post-Chacoan phenomena at Middle San Juan sites in the interval from 1125 to 1140; choosing either the Salmon- or Aztec-specific sequence for the entire region, however, would conflate variation. Third, the Center’s work at Salmon and in the Middle San Juan has been organized from the beginning as a coalition of partners; not as a ­single, integrated project. Considerable prior work has been completed not only at Salmon but also at Aztec and other regional sites over the last 100 years. Thus, overlaying a single chronology to be fol-

lowed by archaeologists working at several different sites with distinct histories and for a variety of institutions, while perhaps desirable, does not seem practical. To address Vivian’s concerns about terminology, I do nevertheless encourage Middle San Juan archaeologists to reference the Center’s sequence and identify specifically (by reference to years ad) the periods being investigated.

The Role of the Middle San Juan in the Ancient Puebloan Southwest The next issue of concern is the role of sites such as Salmon, Aztec, and other Middle San Juan communities in the ancient Puebloan world. Historically overshadowed by Chaco Canyon to the south and Mesa Verde to the north, the ancient Puebloan communities at Aztec and Salmon have rarely been considered in their own right. Rather, the settlement histories of these sites (and the surrounding Middle San Juan region) are usually interpreted by reference to Chaco Canyon and Mesa Verde (e.g., Morris 1919a; Powers et al. 1983; but see McKenna and Toll 2001). Because of their spectacular ruins, these latter areas have understandably received the lion’s share of attention over the last century. As a result, however, interpretations of cultural developments across the northern San Juan Basin have focused on these centers while overlooking the intervening Middle San Juan region. Nevertheless, research over the last 30 years has indicated that Puebloan developments at Aztec and Salmon, and across the Middle San Juan region, have a unique trajectory, linked to, but also independent of, Chaco and Mesa Verde (Hogan and Sebastian 1991; Irwin‑Williams and Shelly 1980; McKenna and Toll 2001; P. Reed 2004a; Toll, this volume; Wheelbarger, this volume, 2003). Furthermore, it is clear that understanding the intervening ground of the Middle San Juan is critical to the construction of accurate models of Chacoan and Puebloan history from ad 1050 to 1300. Research indicates that between ad 1000 and 1130, Chaco Canyon functioned as the political, social, economic, and ritual center of the Pueblo world (Cameron and Toll 2001; Lekson 2006; Mills 2002; Sebastian 1992; Vivian 1990). The recently concluded Chaco Capstone Conferences 20

Setting the Stage have compiled a great deal of data, and the published volumes will tell much of the story of Chaco Canyon (see Cordell et al. 2001; Lekson 2006). Views of so‑called Chacoan outliers (including Salmon and Aztec) have evolved in recent years, and, currently, many Chacoan archaeologists see little evidence for an overarching Chacoan “system” (e.g., Cameron 2002; Durand and Durand 2000; Jalbert and Cameron 2000; Kantner and Mahoney 2000; Reed 2004a; Van Dyke 1999). The Middle San Juan region figures strongly in both of these histories—of Chaco Canyon and its outliers. Chacoan sites were established at Salmon and Aztec in the late eleventh and early twelfth centuries, as the Chacoans expanded to the north into the Middle San Juan. Other communities in the region apparently emulated the distinctive Chacoan style of architecture. The decline of Chacoan political influence after ad 1130 led to the rise of new centers across the greater San Juan landscape, including sites in the greater Mesa Verde region (Varien et al., this volume) and in the Cibola region to the south (Kintigh 1994). As part of this process, regional centers emerged in the Middle San Juan region at Salmon and Aztec (Brown et al. this volume; Irwin-Williams, this volume; Lekson 1999; McKenna and Toll 2001; Reed 2004a).

this volume; Washburn, this volume; Webster, this volume) but evidence supporting a migration of people from the Mesa Verde area to “reoccupy” the sites is lacking (L. Reed, this volume; Washburn, this volume). In fact, the notion of reoccupation is questionable because both Aztec and Salmon show evidence of continuous occupation during the suggested hiatuses of the mid-1100s. I have discussed Salmon’s sequence above, illustrating limited but continuous occupation throughout the 1100s. In Chapter 12, Brown and his colleagues reevaluate Aztec’s occupational history and conclude that Morris’s inferred hiatus did not occur. Rather, Aztec West was continuously occupied throughout the 1100s. Moving to the “how” question, we can explore the mechanisms by which Chacoans came to the Middle San Juan to build the sites of Aztec and Salmon. As discussed, Morris and Irwin-Williams both proposed colonization models, with Chacoans moving north into the Middle San Juan, perhaps as though it was foreign territory. This idea has been recently favored by archaeologists—some suggesting colonization (Lipe 2006; Reed 2002), and others preferring to talk about Chacoan migration (Reed, Webster, Clark et al. 2005). In contrast, other archaeologists (e.g., Toll, this volume) criticize the idea of colonization, saying that the Salmon, Aztec, and the area was previously part of an expanded San Juan– Middle San Juan: The When, Chacoan­ territory (see also Judd 1954, 1964). the How and the Why Along this line of reasoning, Cameron (2005) has When, how and why were Salmon and Aztec estab- recently proposed that the traditional Chaco and lished? What roles did they play in the waning years Mesa Verde “cultures,” as defined by Kidder (1924), of the Chaco world? Archaeologists have offered were not separate, distinct sociocultural groups. answers to these questions since the time of Earl Rather, they represent time-specific adaptations of Morris in the early 1900s. Morris (1921) believed essentially the same, large group of related people. that Aztec was established as a Chacoan colony (see During the Chacoan period, this related populaBrown et al., this volume). Cynthia Irwin-Williams­ tion built large great houses, a road network, traded (this volume) also proposed that Salmon was built extensively for goods, and distributed the traits deby Chacoan colonists. Both archaeologists invoked scribed as Chacoan across a large area of the greater abandonment of the sites after the end of the Cha- San Juan Basin. Later, in the post-Chacoan­period, coan occupations in the mid-twelfth century, and the “Mesa Verde” approach (involving large, agreoccupation by Mesa Verdean peoples in the late gregated pueblos built through accretion of room twelfth and early thirteenth centuries. Recent work units over time) to pueblo living evolved, and was has supported the idea that Salmon and Aztec were also spread across a large portion of the northern built by people from Chaco Canyon (see Baker, this Southwest in the late 1100s and throughout the volume; Brown et al., this volume; Reed 2004a, 1200s. 21

Paul F. Reed In this view, then, Chacoan expansion into the Middle San Juan in the late eleventh and early twelfth centuries was movement into ancestral territory. If this hypothesis is correct, then this process would have been less difficult than movement into “foreign” territory. Whether or not the Chacoans were returning to ancestral territory or moving into new lands in the Middle San Juan, it seems clear that they needed the cooperation of local people to succeed. Thus, I propose (following Irwin-Williams) that Salmon and Aztec were founded by Chacoan migrants who had the explicit cooperation of (and perhaps permission from) local families who were incorporated into the sites as residents. These sites, then, were built and inhabited by a mix of cultural groups from Chaco and the local San Juan area (see Van Dyke, this volume, for a similar view). This view of a combined population for Salmon (and extended to Aztec) has also been proposed for Chaco Canyon. Evidence for a multicultural and multiethnic population for Chaco Canyon has emerged from recent work in the Canyon. Akins’s (1986) research has shown the existence of two distinct groups in the Pueblo Bonito burial population. Schillaci’s (2003) work, although preliminary and based on a small sample, reveals connections between Chacoan groups and others across the Southwest. Human remains from the western part of Pueblo Bonito showed a close relationship with ancient Rio Grande populations, whereas burials from the northern portion were more closely matched with a Basketmaker group from Grand Gulch, Utah. Remains from several small house sites in Chaco revealed a relationship with a group from the Village of the Great Kivas, to the south near Zuni Pueblo. Akins’s work also suggests that status differences were present among the Chacoan populations, with some residents (possibly elites) taller and healthier than others (perhaps com­moners). These findings indicate that considering the “Chacoans” as a homogeneous group is problematic and that at least two or more distinct bio­logical and social groups were present in Chaco Canyon. Initial work on skeletal remains from ­Aztec Ruins (Schillaci and Stojanowski 2002) also indicates a multiethnic population. Finally, let us consider the “whys” of Chacoan

expansion into the Middle San Juan. Previous work by a multitude of scholars has considered this question, which must be linked, ultimately, to the “big” why question of Chacoan archaeology: Why did the Chacoan system collapse? Leading scholars have offered many opinions on this question (e.g., Judge 1989; Lekson 1999; Sebastian 1992; Vivian 1990). For now, I think we can defer discussion of this larger question because northward movement to the Middle San Juan region preceded the beginning of the end for Chaco by at least several decades. But what factors led some Chacoans to think about migrating to the Middle San Juan in the late 1080s? One possible factor was a drought that began in the early 1080s and lasted until almost 1100 (Reed 2004a; Southwest Paleoclimate Project 1996). Although not the first nor even a particularly severe drought, the downturn of the 1080s undoubtedly had some effect on maize and other crop production in Chaco Canyon. Furthermore, by the 1080s, construction of great houses in the Canyon had peaked, with all but the so-called McElmo structures (e.g., New Alto, Kin Kletso) built. The Chacoan system was perhaps at overload by the 1080s and needed additional areas into which to expand (Vivian 1990). Some research, however, suggests that the northward expansion of Chacoan society may have been underway before the 1080s. Chimney Rock Pueblo was built by the Chacoans as a ritual site connected to the observation of the moon at standstill (Malville 2005; Malville and Matlock 1993). Although poorly dated, with only two tree-ring cutting dates at 1075 and 1093, Chimney Rock may have been constructed in the 1070s or in the 1090s. Other northern outliers also have early dates, such as the Wallace site in southwest Colorado (with scattered dates in the 1060s) and Morris 41 (with several scattered dates in the late 900s and early 1000s). Several sites may predate 1090, but none have clear evidence (e.g., clusters of tree-ring cutting dates) of construction prior to this time. In contrast, construction at Salmon Pueblo around 1090 clearly represents the first and most intensive effort by the Chacoans to build in the Middle or Northern San Juan. With Chaco Canyon perhaps at peak capac22

Setting the Stage ity in the 1080s, expansion northward to Salmon, with the intent of establishing a residential population in the Middle San Juan, seems probable. Salmon Pueblo was built in a local area without a significant existing population (Reed 2004a). Settlement data from the area around Salmon reveal a maximum of 10 small pueblo sites (with perhaps 60 rooms) within a 6 × 6 km area surrounding the pueblo (Reed 2002). Certainly, this sample may not include all sites occupied during Salmon’s tenure. Nevertheless, the data in hand do indicate that Salmon was not the center of an extended community, with numerous outlying small sites (as we clearly see on the landscape around such Chacoan sites as Aztec and Pueblo Pintado). The landscape was certainly not empty, yet there was sufficient space for a group to come up from Chaco to construct a large pueblo. Furthermore, Salmon was built on the banks of the San Juan River not only to take advantage of this obvious water source, but also to utilize runoff from a side canyon (see IrwinWilliams, this volume). In short, the Middle San Juan, generally, and Salmon, specifically, appeared quite attractive to the migrating Chacoans, particularly with Chaco Canyon packed full of great and small houses and perhaps pushing its productive limits. Later, in the early 1100s, Aztec’s West Ruin was constructed, like Salmon, as a large Chacoan structure. Later still, Aztec’s East Ruin was built, beginning in the 1120s and 1130s. All of these construction events represent, in my view, Chacoan commitment to moving into the Middle San Juan region. Other authors in this volume (e.g., Van Dyke, L. Reed, Toll, Vivian, Cameron) offer differing views of relationships between the Middle San Juan and Chaco.

the history of studying great houses. Chaco’s much overlooked small houses certainly played important roles during ancient Puebloan history (see McKenna and Truell 1986), but these sites have taken a backseat to the big houses. Clearly, the great houses are one of the primary factors that set Chacoan society apart from other, contemporary Puebloan manifestations. As I discuss in Chapter 3 (this volume), views of Chacoan great houses have evolved considerably since the mid-1800s. Initially seen as the residences of thousands of people, current Chacoan “orthodoxy” identifies these structures as specialized administrative and ritual centers with extremely low resident populations. Pueblo Bonito, for example, is now thought to have had fewer than 100 fulltime residents at its peak (Bernardini 1999; Bustard 2003). Pueblo Alto is viewed by Toll (1985, this volume) and other Chacoan scholars as a specialized ritual site with minimal population in residence. The abundant midden at the site was interpreted by Toll (1985) as a ritual deposit, with thousands of ­ceramic vessels ritually “killed” in ceremonies. Other archaeologists have taken Toll to task for this interpretation, most notably Wills (2001). Current fieldwork by Wills (2005) in the mounds fronting Pueblo Bonito should help clarify the nature of their formation and the use of these features at other great houses. Data collected from Salmon Pueblo provides a sharp contrast with the orthodox Chacoan view of great houses serving primarily nonresidential functions. As I detail in Chapter 3, Chacoan floors and deposits at Salmon contain abundant evidence of residential usage. These data, together with its formal room suites and great size, point conclusively to Salmon’s use as a residential complex for a mixed Chacoan and local Middle San Juan–derived population. As discussed by Brown and his colleagues (this volume), Aztec represents a more complicated ­vision of great house function. Brown et al. discern more evidence of residence at Aztec in the McElmo and Mesa Verde phases (1140–1300) than during the founding Chacoan occupation (1105–1140). Nevertheless, Brown and his colleagues have not ruled out residential usage of Aztec during its

Changing Views of Chacoan Great Houses in the Middle San Juan With a good background and understanding of events in the Middle San Juan in the late eleventh and early twelfth centuries, we can explore an issue raised in several volume chapters: the function and use of Chacoan great houses. The history of research at Chaco Canyon and the sites identified as “outliers” can be described (at least partially) as 23

Paul F. Reed ­ hacoan occupation. One key difference between C Salmon and Aztec lies in the nature of room arrangements. Salmon contains formally designed room suites that I have interpreted (following ­Irwin-Williams) as residential family quarters. Aztec has arrangements of rooms, but not true suites similar to those at Salmon. I believe these and other clear differences between Salmon and Aztec relate primarily to the different goals and needs of the Chacoan populations that established the sites. Salmon was founded, in my view, by a group from Chaco who possessed marginal-to-average agricultural lands in the Canyon. When climatic conditions worsened in the 1080s (as discussed above), this group began to look for better opportunities elsewhere. Their search brought them to the San Juan River, and they built Salmon Pueblo in a relatively unoccupied area. Aztec, by contrast, seems to represent movement by families more central to Chaco’s leadership. If Steve Lekson (1999), Stein and McKenna (1988), and Van Dyke (this volume) are correct, Aztec was founded by Chacoan groups intent on re-creating the complex of great houses and ritual structures that formed “downtown” Chaco Canyon. Thus, Aztec West may have been constructed initially as an administrative-ritual center with a smaller number of planned residents compared to Salmon. The subsequent construction of Aztec East and outlying structures, such as the Hubbard site and other tri-wall ritual structures, completed the design of Aztec, creating a symmetrical pattern similar to downtown Chaco (Stein and McKenna 1988; Van Dyke, this volume). Clearly, we have just begun to fully explore these issues. Completion of the Center’s recent NSF project will shed considerable light on many issues (Reed, Webster, Clark et al. 2005).

Part 3 (Chapters 4–8) delves into the detailed material culture analyses undertaken in conjunction with the new Salmon Research Initiative and the original Salmon Project. In Chapter 4, Karen Adams summarizes the wealth of archaeobotan­ ical data from Salmon, exploring the similarities and differences between Salmon’s Primary (Chacoan) and Secondary occupations. Karl Reinhard ­tackles the expanding field of parasite pathoecology in Chapter 5, providing significant insight into the health of Salmon’s ancient inhabitants. In Chapter 6, Kathy Roler Durand and Steve Durand examine faunal exploitation in the Chacoan world via a study of Salmon’s extensive bone collection and comparison to those of other Chacoan great houses. Tom Windes and Eileen Bacha use Salmon’s extensive wood and tree-ring data to discuss chronology and construction aspects of the site in Chapter 7. Finally, in Chapter 8, Nancy Akins addresses the human remains from Salmon’s Tower Kiva, attempting to unravel the mystery of the complicated deposits in the kiva structure and how they relate to the occupation-ending fire that occurred in the late 1200s. Part 4 of the volume addresses material culture studies using Salmon’s assemblage as the starting point but extending the focus to the larger Middle San Juan region. In Chapter 9, Laurie Webster explores the wondrous array of perishable fiber artifacts from Aztec, Salmon, and Chaco Canyon sites, teasing out patterns across the region. In Chapter 10, Lori Stephens Reed addresses the complexities of local ceramic production at Salmon, Aztec, and across the Middle San Juan region. Part 5 expands examination of ancient Puebloan adaptations into the surrounding Middle San Juan region with chapters on other Chacoan outliers and communities. In Chapter 11, Linda Wheelbarger addresses five local communities on the south side of the San Juan River, west of Salmon. Gary Brown, Tom Windes, and Peter McKenna update our understanding of the Aztec Community and its central role in the Middle San Juan region in Chapter 12. Finally, in Chapter 13, Cathy Cameron takes us west down the San Juan River, with the Bluff great house community as her focus. The four chapters in Part 6 offer synthetic

Organization of the Volume The book is divided into seven parts, including this introduction as the first. Part 2 focuses on Salmon Pueblo: Larry Baker details aspects of Salmon’s architectural history and its connection to archaeoastronomy and ritual in Chapter 2, and in Chapter 3, I examine data related to Salmon’s use as a residential Chacoan site. 24

Setting the Stage views of the Middle San Juan region. Chapter 14 represents Cynthia Irwin-Williams’s final view of Salmon and its role in the Chacoan Phenomenon; Chapter 15, by Dorothy Washburn, builds on IrwinWilliams’s view using ceramic symmetry research to address the social and ethnic origins of Chacoan sites. In Chapter 16, Wolky Toll offers a synthetic view of relations between the La Plata Valley, the Middle San Juan (Totah), and Chaco Canyon. In Chapter 17, Ruth Van Dyke conceptualizes the Chacoan period of the Middle San Juan, with special attention to sacred and ritual connections. Part 7 of the volume offers two views from outside the Middle San Juan region. In Chapter 18, Mark Varien, Scott Ortman, Susan Ryan, and Kristin Kuckelman provide a unique view of Salmon and the Middle San Juan from their research base in the central Mesa Verde region. Finally, Gwinn Vivian offers his perspectives on Salmon and the Chacoan world in Chapter 19. Vivian’s contribution offers synthetic comments on the volume, drawing on his unparalleled experience in Chacoan archaeology over the last several decades. Finally, in Chapter 20, I provide closing comments, describe the current state of Middle San

Juan archaeology, and discuss areas that require additional research. As the chapters in this book attest, research in the Middle San Juan region is experiencing a resurgence. New ideas abound, and new classes of data are currently under study. It truly is an exciting time for archaeology in the region.

Notes 1. Portions of the section titled “The Middle San Juan Region in Historical Context” are reprinted, with permission, from “The Ruins of Salmon Pueblo in Historic Context,” in Thirty-Five Years of Archaeological Research at Salmon Ruin, New Mexico (Tucson, Arizona: Center for Desert Archaeology, and Bloomfield, New Mexico: Salmon Ruins Museum, 2006), edited by Paul F. Reed, and The Puebloan Society of Chaco Canyon, by Paul F. Reed, copyright © 2004 by Paul F. Reed, reproduced with permission of Greenwood Publishing Group, Inc., Westport, CT. 2. Robert Dubois was asked to comment on Eighmy’s reinterpretation of his 1970’s data. DuBois indicated that he was nearing completion on a publication that would summarize much of his archaeomagnetic work over the last several decades. He declined to share any new results that might relate to Salmon nor did he wish to respond to Eighmy’s reassessment.

25

Pa rt 2

Salmon Pueblo

2

Salmon Ruins Architecture and Development of a Chacoan Satellite on the San Juan River Larry L. Baker Current research on the “Chaco Phenomenon” (­Irwin-Williams 1972) is being undertaken at a time when the dynamics of these studies are greater than ever before. This is hardly surprising given the extensive body of research that has advanced our understanding of nearly every aspect of Chacoan architecture and material culture. Nevertheless, the breadth and diversity of these scholarly works, many by authors contributing to this volume, are staggering to say the least. Recent studies, for example, have begun to redefine the Chacoan regional system (Durand and Durand 2000; Kendrick and Judge 2000; Van Dyke 2000). In this regard, emphasis has been placed on the outlying communities compared to sites in the canyon and “downtown Chaco” (Lekson 1984). The current interest in outlying communities stems, in part, from these communities being understudied. In addition, researchers are being encouraged to think beyond the “Chaco Canyon box” and beyond a San Juan Basin–centric focus as well. Steve Lekson (2000:​163), for example, tells us, “Don’t just think ‘great,’ think BIG.” We are fortunate to have scholars who have broadened our understanding of Chacoan prehistory through study of outlying Chacoan communities, whether these sites are seen as the result of independent developments, local emulation, or Chacoan migration. Lekson’s Chaco Meridian treatise is a case in point (Lekson 1999). As archaeologists and scientists, we need to strive for this broader perspective, whether the individual ap-

proach tends toward new archaeology, processual, or post-processual. To work toward these greater ends, however, the parts of the system must first be better understood. Lekson (2000:163) has proposed that “the best way to understand that pattern is not to fragment this great big problem into tiny analytical units,” but in the context of the Salmon research initiative, it seems essential that the individual “analytical units” be understood in detail. In 1990, Vivian wrote: At some point during the Classic Bonito phase, buildings with an architecture reminiscent of the canyon great houses appeared as a scattered constellation on the fringes of the Chacoan world. Interpreting this event’s timing, extent, and form is central to understanding late ­Chacoan development. Classifying the timing of great house construction in outlier communities is particularly crucial. (1990:​ 328–329) The parts or components must be explained before the system can be grasped, no matter how it is defined. In this regard, Cordell and Judge (2001:9) offered this perspective a decade after Vivian’s comments: “our most immediate task for the future is to continue to describe with chronological precision the interactions within and between Chaco and the outlier communities.” It was in this vein that the current study on Salmon Ruins was developed. 29

L arry L. Baker

Prior Chacoan Architectural Studies

structed in brief building episodes, and cutting dates from these sites, though limited in number, most likely reflect the total period of site construction within one or possibly two decades. (1990:239)

Considerable work has been devoted to understanding the construction sequences of the great houses of Chaco Canyon, such as Chetro Ketl (Hawley 1934; Lekson 1983), Pueblo Bonito ( Judd 1964; Lekson 1984) and Pueblo Alto (Windes 1987a), to name a few. Generally less detailed studies have been undertaken at the out­liers with some exceptions, for example Guadalupe Ruin (Pippin 1987; Terrel and Durand 1979), Aztec Ruin (Brown et al. 2002, this volume; McKenna and Toll 2001), Lowry Ruin (Roys 1936), and of course, Salmon Ruins (R. Adams 1980; P. Reed 2006b). In most of these cases (e.g., Lekson 1984), considerable understanding of the detailed construction sequences has been realized. Prior studies have not, however, approached Chacoan architecture from the perspective of a modern architect or general contractor. To shed additional light on the issues discussed, I adopt such an approach in this chapter. In Chaco Canyon, chronological data indicate that a relatively large resident labor force was employed to erect major sections of buildings over a short period of time. Construction materials were amassed and large portions of structures built as cohesive architectural units. It has been assumed that the construction of Salmon Ruins, located approximately 75 km north of Pueblo Bonito, was undertaken in a similar manner (Figure 2.1). The numerous tree-ring cutting dates from ad 1088 to 1090 mark a singular, massive wood-harvesting episode that is second to none, even in comparison to sites in Chaco Canyon. Vivian has summarized these inferences in relation to great house construction both in and outside of Chaco Canyon.

For Salmon, there is no doubt that major tree harvesting was undertaken from ad 1087/1088 to 1090 (R. Adams 1980; P. Reed 2006b; Windes and Bacha 2006). In order to muster this level of effort, however, a massive influx of people would have been necessary—not only to harvest trees, but also to acquire stone, prepare foundations, initiate masonry construction, and provide the logistical support for laborers, masons, architects, and others. Just providing food and shelter for the builders would have staggered and strained an infrastructure possibly already under stress due to climatic, environmental, and demographic problems in Chaco Canyon (Cordell 1984; Vivian 1990). The present architectural research at Salmon Ruins proposes a realistic approach to understanding the construction and development of this massive masonry pueblo.

Salmon Data: Tree-Ring Dates, T-Doorways, and Wall Abutments Excavations at Salmon Ruins provided limited numbers of tree-ring cutting dates prior to the massive clustering of dates between ad 1087 and 1090. Many of the wood samples that produced cutting dates were collected from rooms in the Tower Kiva area and the east and west sections of the main roomblocks. Nonetheless, there are some clues to earlier construction events in other areas of the building. A few reliable samples clustering within a five-year period came from several rooms in the pueblo’s east wing (Figure 2.2). These dates consist of (1) two samples from Room 122 dating to ad 1068 and 1070; (2) one sample from Room 119 dating to ad 1070; and (3) three samples from Rooms 121 and 124 dating to ad 1072. Rex Adams (1980:​ 215) suggested in the original Salmon report that these beams recovered from the east wing may have been taken from a small, previously constructed pueblo in the area. In contrast, I think these dated beams provide evidence of a previously overlooked, earlier component of Salmon.

The dendrochronological evidence from Chaco Canyon suggests similar, short-term massive construction episodes within and between great houses. According to Lekson (1984) six major construction programs in Chaco Canyon carried out between 1075 and 1115 were sequential, with a program being completed about every 7–10 years. Based on these data, smaller classic Bonito outlier great houses probably were con30

Salmon Ruins: Architecture and Development of a Chacoan Satellite

Figure 2.1.  Location of Salmon Ruins in northwest New Mexico.

Early tree-ring cutting dates from ad 1068 to 1072 come from an articulated set of rooms in the southern end of the east wing. The tree specimens are ponderosa pine and Douglas fir, which represent the classic Chacoan approach to tree harvesting (Durand et al. 1999; Windes and McKenna 2001). This evidence indicates that these beams were procured by Chacoan builders for construction and were probably not taken from an abandoned unit pueblo in the general vicinity of Salmon. Closer examination of the data suggests that the southern and central portions of the east wing may have been one of the building’s earliest components. One piece of evidence is a T-doorway positioned in the central portion of the wall between

Rooms 122 and 129. Ostensibly, this T-door is interior, but Room 122 was originally constructed with three T-doors: one facing east and two facing west toward the plaza (Figure 2.3). Room 129, immediately east of Room 122, was not an original plazafacing room and does not exhibit reliable tree-ring dates until the ad 1090 construction episode. The significance of T-doors in Chacoan great houses is undeniable, and archaeologists have been exploring these unique doorways for years, since Judd (1964) was perplexed by their presence at Pueblo Bonito. In his Chaco Meridian, Lekson (1999:175–177) devoted an entire appendix to T-doorways. Speaking of their “original ancient use,” he is quite specific: “T-doors were generally exterior, ground-floor 31

Figure 2.2.  Location of Rooms 119, 121, 122, and 124 (shaded) with early tree-ring dates in Salmon’s east wing.

Figure 2.3.  T-shaped doorways in Rooms 100 and 122 facing the plaza and also the roomblock interior.

Figure 2.4.  Abutments and continuous wall sections within the east wing and eastern portion of the main roomblock of Salmon Pueblo.

Salmon Ruins: Architecture and Development of a Chacoan Satellite e­ ntries, open to the outside world. Not all T-doors were exterior but, at Chaco, most at least began as exterior doors; an informal analysis of other Anasazi T-doors suggest that they were predominantly exterior features.” At Salmon, examination of the original placement of T-doorways during any single construction episode suggests a similar pattern: these features were predominantly exterior and faced open spaces (such as the plaza) as initially built. Later room additions may have changed T-doorway positions as successive construction episodes enclosed earlier sections of the building (R. Adams 1980:226). This is indeed the case for Room 122 with regard to T-doorways facing east and west, respectively. All three may originally have been exterior T-doors. Another room (100W) at Salmon exhibits a similar pattern, as discussed below. Other construction evidence from Salmon’s east wing corroborates this line of investigation. Wall contacts (abutments) at room corners were examined in this area. The data used consisted of: (1) a 1975 site-wide abutment study ( J. K. Adams 1975) undertaken during the original Salmon project; and (2) a reexamination of the abutments during the present study. A pattern of continuous wall alignments in conjunction with bonded wall corners suggests that Rooms 121 and 124 were built as a unit (Figure 2.4). The west exterior walls of these rooms from the southwest corner of the east wing (Room 124) to probably the mid-wall section of the west exterior wall of Room 122 were built as a continuous wall section. The south exterior walls of the east wing of Rooms 123, 124, and 127 were also constructed as a continuous segment. Corner bonding was observed in the northeast and southeast corners of Room 121, the northwest and southwest corners (interior and exterior) of Room 124, and the southwest corner of Room 123. This evidence suggests that these two rooms (121 and 124) were built as a unit and that the east-to-west trending walls were extended to delineate later construction of rooms in this area. No wall bonding was observed in the northwest, northeast, or southeast corner of Room 123 or any wall corners of Room 127. Room 122, a single-story room, produced early tree-ring cutting dates and had the three T-doors

mentioned earlier. As a result, I grouped this room with Rooms 121 and 124 in the proposed early room suite at Salmon. However, this grouping could not be confirmed because Room 122 was not excavated to the point that preserved wall corners could be examined to determine corner wall junctures. Room 119 also has an early cutting date. The north and south walls of this room appear to be constructed as continuous wall sections, and the southwest corner is bonded. Corners are abutted in the southeast and northwest corners of Room 119, as well as in the southeast first-story corner of Room 129. These findings suggest that the east-west wall of Room 122 and 129 was built as a continuous wall segment. Unfortunately, other corners could not be observed. Further evidence of early construction and occupation in the east wing is suggested by a series of reliable dates at ad 1094, 1105, and 1106 from Rooms 121 and 122. In the original construction phase discussion, R. Adams (1980:232) concluded that “many of these dates occur in the south end of the east wing [and] it is suspected that this is the period during which these rooms were added.” Adams did allow that some “maintenance” or “repair” of the pueblo may have occurred during this time period (R. Adams 1980:222, 228). I would suggest that the latter is indeed the case, and that the east wing required repairs at 1094 and between 1105 and 1106 because of its early construction in the 1068–​ 1072 period. Another section of Salmon Pueblo contains a room that exhibits two T-shaped doorways (see Figure 2.3). Room 100 is a single-story room in the northeastern sector of the main roomblock that contains T-doors in its south and east walls. The room produced no pre-ad 1087 tree-ring dates that would corroborate an early construction episode similar to that of Room 122. A tree-ring date at ad 1106 might indicate roof repair and maintenance, but dates at 1089 and 1090 indicate that the room was built during the first construction episode at 1090. The rooms directly east of Room 100 were not excavated, and the lack of tree-ring dates precludes determination of the timing of construction. R. Adams (1980:225) indicated that singlestory rooms along the eastern main roomblock were completed during the first phase of ­construction, 35

Figure 2.5.  Astronomical observations and architectural reference points related to the summer and winter solstices, equinox, minor lunar standstills, and north-south cardinal alignment.

Salmon Ruins: Architecture and Development of a Chacoan Satellite Table 2.1.  Astronomical Observations and Architectural Reference Points (Positions) at Salmon Pueblo Observation

Bearing

Lunar Standstill Summer Solstice Summer Solstice Winter Solstice Winter Solstice Equinox Equinox N–S Alignment

65.8° 59° 59° 123° 122° 90° 90° 1°

Architectural Reference Position Back wall of the pueblo NE corner of pueblo to red sandstone mosaic in Room 59 SW corner of east wing to center of Great Kiva Center of Great Kiva to NW corner of pueblo Center of Tower Kiva to mosaic in Room 59 Center of Tower Kiva to Rooms 119 and 122 Center of Tower Kiva to Room 82 observatory Center of Great Kiva to center of Tower Kiva

possibly being discontinued at Room 100. Without corroborative dates, any suggestion of pre-ad 1090 construction for Room 100 remains speculative. Nevertheless, it is possible that the architectural design radiated out from the room, if indeed it was earlier in the construction sequence. The northeast, southeast, and southwest corners of the room are bonded. Rooms 122 and 100 are similar in terms of their general position in the pueblo, as well as size and type of adjoining rooms (see Figure 2.4). Both were originally designed to be large, single-story, plazafacing rooms. They are flanked on either side along the plaza-facing transect by other large rooms. In both locations, there is a small room articulated against one corner of the large room, located two rooms way from the exterior east wall of the pueblo. For Room 122, the room in this position is Room 119, which contains an early date at 1070. For Room 100, this room is Room 159, which was not excavated and produced no tree-ring samples.

that summer and winter solstice, equinox, and cardinal direction alignments were employed by the ancient architects to orient the site layout and ground plan similar to modern use of a blueprint. Key reference points in the pueblo were pinpointed via astronomical reference alignments. Table 2.1 summarizes the architectural reference points in relation to astronomical alignments (Baker and Mantonya 2002). Figure 2.5 illustrates these observations with respect to the architectural references and positions. The data suggest that the proposed earliest construction unit at Salmon was integrated into this astronomical design scheme. Alignments related to the summer solstice and equinox were employed to position and identify the rooms in the earliest dated component of the building. Table 2.1 describes several of these alignments, which include the following observations: (1) Summer Solstice: center of Great Kiva to southwest corner of Room 124; (2) Equinox: center of the Tower Kiva to Room 82 observatory to Room 122 and Room 119. If Room 100 was part of an early component in the pueblo or, at a minimum, a key element in the primary phase of construction, one would expect to find an astronomical alignment connecting this room to the other areas exhibiting these reference points (see Figure 2.5). Most obvious is the lunar standstill alignment from the center of the Tower Kiva to the center of Room 100. The summer solstice also provides an alignment from the observatory opening in the east wall of Room 82 to the Room 100 T-doorway. Furthermore, the northwest corner of Room 122 and the northeast corner

Astronomical Alignments and Architectural Relationships If the proposed early architectural units are to be further verified, we must explore how they were integrated into the overall design of the larger structure. Several studies have indicated that astronomical observations were employed to define the architectural layout of Salmon Pueblo (Baker and Mantonya 2002; Sofaer 1997; Sofaer et al. 1991). Sofaer found that the back wall of the pueblo was aligned to a minor lunar standstill. Baker and Mantonya (2002) expanded on Sofaer’s work, finding 37

L arry L. Baker of Room 100 are connected by a north-south cardinal alignment. Provided the data and evidence presented herein represent an accurate view of the architectural development of Salmon, the chronological and design-construction sequence should reflect the integration of the data. This is indeed the case, and the following discussion is offered.

procured, and the building foundation trenches ­required excavation. Logistics were crucial for ­supporting the number of people necessary for this massive undertaking. Elements of planning, distance from Chaco, the magnitude of the project, and the complexity of events at Chaco itself may have contributed to the approximately 20-year ­delay between initial work in the mid-ad 1060s and the major construction episode culminating at 1090. In this scenario, the planning process took many years. In Chacoan society, it may have evoked numerous cultural and societal issues related to subsistence, organization, and religion, which were well beyond the pragmatics of basic masonry construction.

Design

It has been proposed that lunar and solar observations and respective alignments guided the design of Salmon Pueblo (Baker and Mantonya 2002; ­Sofaer 1997; Sofaer et al. 1991). The correlation of such events to construction episodes, however, must be supported with tree-ring dates. Salmon’s back wall is aligned with a minor lunar standstill. In relation to the proposed early unit, the minor lunar standstill of ad 1066 is the proposed design target date. The winter and summer solstices, equinox events, and the respective observations related to design would have been conducted over the course of at least a year or two, or possibly longer, depending on weather conditions. If the weather conditions were ideal, the design of the ground plan would most likely have been complete by ad 1067. The site was then ready for construction.

Causal Implications for Construction of the Salmon Satellite Given the proposed design date of ad 1066, the construction of a major Chacoan structure (Salmon Pueblo) 75 km from its cultural center must have been related to internal factors in Chaco Canyon. Climatic and environmental causes are frequently cited in discussions of population movements in the greater San Juan Basin. Vivian (1990:478) notes that “precipitation was generally well above normal during the first decade of this period (ad 1020–​ 1120) and then oscillated downward until about 1050, when average [rainfall] increased for about twenty years.” This pattern set the stage for what happened in the next two decades and is connected to Chacoan long-range planning as a means to circumvent subsistence catastrophe. Vivian (1990:​ 222) also observed that “despite a continued favorable climate, resource shortages probably occurred in much of the Eastern Anasazi area when population exceeded production.” The reconstructed Palmer Drought Severity Index (PDSI) for Chaco Canyon (derived from Jeff Dean’s Southwest Paleoclimate Project [1996]) is particularly intriguing (Figure 2.6). The index shows a yearly value well below the mean in ad 1062 and a general trend in the 1060s toward reduced precipitation and drought. Community fissioning among the ancient Pueblos has been discussed in relation to popula-

Initial Construction

Tree-ring dates suggest initial construction in Room 122 at ad 1068–1070. Dates in other rooms of the proposed early roomblock range from ad 1070 to 1072. Thus, I would suggest that the construction of the four rooms occurred during a fiveyear time span (ad 1068–1072). Construction of Room 100 and adjacent rooms could have been undertaken during a similar time frame. Due to the absence of early dates in the room, however, this cannot be confirmed.

Planning

With the construction of a few rooms at the site, the stage was set for moving groups of builders to Salmon for the monumental construction of the largest Chacoan structure outside of Chaco Canyon (prior to the construction of Aztec East). Massive amounts of building stone and trees had to be 38

Salmon Ruins: Architecture and Development of a Chacoan Satellite

Figure 2.6.  Reconstructed Palmer Drought Severity Index (PDSI) values for Chaco Canyon, ad 1000–1100. Data provided by Jeff Dean from the Southwest Paleoclimate Project.

tion dispersion ( Johnson 1982:407; Vivian 1990:​ 222). Regarding the construction of Salmon Pueblo, this process cannot be ruled out as a prime influence. Alternately, both Johnson (1982) and Vivian (1990) propose the development of simultaneous hierarchies as a result of emigration. Vivian (1990:​ 484–​485) specifically addresses these issues related to the communities established in the northern San Juan Basin. This is particularly telling when he states, “Thus, the Chacoan case may represent a developing two-level (parent-colony) simultaneous hierarchy with each level organized as a rotating sequential hierarchy. If true, this process probably began in the 1060s when new ­Chacoan-San Juan communities were established at the edges of the Core” (Vivian 1990:​485). The data presented in the current Salmon study suggest that this process began in the 1060s, on or about the lunar standstill of 1066. As intriguing as this discussion is in relation to later Chacoan sociopolitical organization and the development of the Salmon community, it is beyond the scope of this chapter. In relation to 1060s construction both within and at the margins of the Chaco core, a series of sites exhibit construction during the same time as Salmon. Additions to existing sites aside, several great houses were initiated during this period. 39

These included Pueblo del Arroyo, Pueblo Pintado, and Hungo Pavi (Lekson 1984). The similarities in general form and site layout between Hungo Pavi and Salmon Ruins are uncanny (Figure 2.7). Because Hungo Pavi has not been excavated, it is not possible to make further comparisons. Nevertheless, Lekson (1984:152) concluded that “although Hungo Pavi may have been built in a series of discrete construction events, these cannot be defined from the visible ruin. I suspect that Hungo Pavi was built in a series of stages from 1060 to at least 1080.” At this time, we can only speculate about a relationship beyond the chronological construction sequences of Salmon and Hungo Pavi. Pueblo del Arroyo, located in the heart of “downtown Chaco” (Lekson 1984), has little in common architecturally with Salmon beyond the proposed correspondence in ad 1060s construction. In comparing the two sites, their forms and ground plans are substantially different. On the other hand, Pueblo Pintado is intriguing not only because of its 1060s construction but also its architectural similarities with Hungo Pavi (Vivian 1990:​ 286) and its location near the headwaters of Chaco Canyon at the margins of the Chaco core. It is possible that for Salmon and Pueblo Pintado, the timing of their construction and their ­emplacement

Figure 2.7.  Comparison of forms and ground plans of Salmon Pueblo and Hungo Pavi.

Salmon Ruins: Architecture and Development of a Chacoan Satellite beyond “downtown Chaco” served a similar function. Along these lines, Vivian (1990:481) has encapsulated their purpose quite well, suggesting that “the establishment of new communities, such as Pueblo Pintado in the 1060s, probably reflected a need to expand the arable land base as the productive capacity of communal farms for great houses in Chaco Canyon failed to meet the needs of a growing population.” The implications of this statement in relation to the construction of Salmon Ruins during this same time period are clear.

The proposed early architectural units at the Salmon Pueblo may have served as a logistical staging facility in advance of the major construction ­episode at ad 1090. The completed rooms could have housed architects, laborers, and masons and provided limited storage facilities for supplies. This approach to planning seems essential for under­ taking the massive construction project at Salmon, more than 75 kilometers from Chaco Canyon. In the absence of such an approach, we could imagine that the Great North Road would have likely been seen as the “Great Salmon Road” if it was used to maximize time and minimize distance for transportation of people and goods to the Salmon locality. This supposition is speculative, at best. However, it has been proposed that the Chacoan roads saved energy and allowed for more efficient transportation of people and materials (Morenon 1977). In the absence of an early architectural component at Salmon, the design, planning, and construction might have been more logistically difficult and taxing. Finally, the construction of the pueblo conforms to an interesting series of “planning principles” that have been proposed by scholars outside the archaeological community. As architects and planners, Dent and Coleman (1997:59–60) have identified planning processes they believe were incorporated into the creation of Chacoan architecture. In their final assessment, they concluded that “Chaco’s built forms exhibit the sense of design and order that came from both the singular vision of the architect and the perseverance over time of the planner. Consequently, we see these remains as evocative of a better way to plan and build” (Dent and Coleman 1997:60). This was indeed the case for the Salmon Pueblo. The architects and planners had the entire history of Chacoan architectural construction from ad 850 to 1060 to draw on as the pueblo was designed, planned, and built. Those involved in the project approached the undertaking with organized precision to achieve the creation of Salmon Pueblo so far from their cultural center at Chaco Canyon.

Summary and Conclusions In this chapter, I have proposed that the initial construction of Salmon Pueblo on the San Juan River began with one or two architectural components predating the massive construction process that began at ad 1090. Based on tree-ring cutting dates in the east wing of the E-shaped pueblo and relationships to archaeoastronomy, I believe that the design and layout of the site were undertaken in ad 1066–​1067 and that the early architectural unit was constructed from ad 1068 to 1072. This early unit was later incorporated into the overall ground plan of the pueblo. The architecture of the initial component and the entire site exhibit many of the “hallmarks” that have become synonymous with Chacoan masonry in and beyond Chaco Canyon. In the current study, select characteristics of Chacoan architectural construction (Marshall et al. 1979; Powers et al. 1983) include: (1) regular and planned site layout; (2) large room size; (3) coreveneer style wall construction; and (4) variability in wall facings. These criteria and others have become the standard for distinguishing and defining Chacoan outliers and in differentiating them from other prehistoric Puebloan sites across the San Juan Basin (Powers et al. 1983). It has been proposed that astronomical observations were employed to design site geometry (Sofaer 1997; Sofaer et al. 1991) and that the use of archaeoastronomy in Chacoan buildings might be considered another hallmark of Chacoan architecture (Baker and Mantonya 2002:12).

41

3

Salmon Pueblo as a Ritual and Residential Chacoan Great House Paul F. Reed

Salmon Pueblo was built on the north bank of the San Juan River, west of the present-day town of Bloomfield, New Mexico, around ad 1090. The site was constructed by probable Chacoan migrants, along with a contingent of local groups, with about 300 rooms spread over three stories. Two primary ritual structures were part of the pueblo’s design: an elevated, second-story (tower) kiva in the ­middle of the primary room block and a great kiva in the plaza. As discussed in Chapter 1 (this volume; see also P. Reed 2006b) Salmon’s Chacoan occupation lasted into the 1120s. After this time, Chacoan architecture at the site was transformed by local residents who were no longer constrained by the prior Chacoan design and floor plan. This transformation unfolded over more than a century, into the mid-1200s, and comprised a number of changes, including placement of kivas into rooms and into the plaza (more than 20 kivas had been added to the site by the 1200s) and division of large Chacoan rooms with the addition of cobble-and-mortar­ crosswalls. The site was occupied at a steady but low level throughout much of the 1100s. In the late 1100s or early 1200s, use of the site increased as a result of local population growth and in-migration­from the surrounding Middle San Juan area. Salmon probably reached it population peak around 1250; it was abandoned in the late 1280s, coincident with a pueblo-wide fire. My focus in this chapter is the first part of Salmon’s history: the Chacoan period from ad 1090 to 42

1120s. As detailed in a number of recent publications (e.g., Baker 2006, this volume; P. Reed 2006a; Windes and Bacha 2006, this volume), Salmon Pueblo was built in perhaps the shortest time of any Chacoan site. Based on the pattern of tree-ring cutting dates, most of this massive 300-room, threestory pueblo was erected in about five years (from ad 1090 to 1095), with some repairs and maintenance at 1106 and 1118. Tree-ring data do not support such rapid construction for any other Chacoan site. Salmon was unique in this regard and in several others, as I explore in this chapter.

Salmon as a Unique Chacoan Outlier Following Cynthia Irwin-Williams (this volume) and others, my research colleagues and I have hypothesized in our National Science Foundation– sponsored research that Salmon represented the Chacoans’ first large-scale attempt to move to and settle in the Northern and Middle San Juan regions (Reed, Webster, Clark et al. 2005). Other sites north of Chaco Canyon may have been built earlier​ —​Wallace, Lowry, and Chimney Rock in southwestern Colorado, for example—but no other sites of comparable size and scale were constructed prior to 1090. Indeed, at 300 rooms, Salmon was many times bigger than any of the great houses that may have predated it. Salmon’s construction was quickly followed by that of Aztec’s West Pueblo, built between 1105 and 1115. Nevertheless, the construc-

Salmon as a Ritual and Residential Great House tion of Salmon was a result of different concerns and needs than the construction of most or all of the other post-1050 great houses in Chaco Canyon and, perhaps, across the greater San Juan region. What were these differences? I noted the primary difference above: Salmon’s size supports the idea that the Chacoans built the pueblo as a residential outpost from which to expand into the Middle San Juan region. Most of the other early Chacoan outliers (excluding the Aztec community) across the Middle and Northern San Juan region were much smaller, with Lowry having perhaps 40 rooms; Chimney Rock, perhaps 55; and Wallace, about 70 (Powers et al. 1983). If Salmon was to be yet another Chacoan administrative center, ritual site, or elite residence, depending upon the interpretation of great houses, why was it built as a massive 300-room dwelling? Despite arguments to the contrary (see Toll, this volume), the data strongly suggest that residential usage was the primary, planned use of the massive Salmon building. Furthermore, Salmon was founded along the San Juan River in an area nearly lacking earlier or contemporaneous Puebloan settlement. An initial settlement study identified four small pueblo sites that may have been contemporaneous with Salmon in the 1 km surrounding area (Reed 2002). Expanding the area, 12 habitation sites containing 68 rooms and up to 10 kivas were identified in a large 6 × 6 km area around Salmon (Reed 2006d). Most of these sites were found during the 1970s survey of the Pleistocene terraces in the immediate vicinity of Salmon and other adjacent locales during the San Juan Valley Archaeological Program (the moniker Irwin-Williams assigned to the Salmon Project; Grove 1972). The survey did not entail 100 percent coverage because of the complexity of land ownership and lack of permission to survey some parcels. Nevertheless, much of the territory in the 1 km area around Salmon was surveyed. As a caveat, it is likely that flood deposits from the San Juan River, along with alternating cycles of erosion, may have concealed or removed other sites located on the floodplain below Salmon. We have no way of knowing how many such sites may have been present. With the data that are available, however, it is clear that Salmon was not the center of a large

community of surrounding small pueblos; rather, Salmon largely comprised the entire community. This finding contrasts sharply with other large, late Chacoan sites that were surrounded by abundant, contemporaneous small sites such as Aztec, Pueblo Pintado, and Newcomb, to name just a few (Marshall et al. 1979; Powers et al. 1983; Stein and McKenna 1988). Thus, I think Salmon was deliberately sited in a largely ­vacant landscape, at least at the ­local level, as a large-scale Chacoan residential and ritual complex. The founding of Salmon around 1090 represents a watershed in the history of the Middle San Juan region. The shift northward from Chaco Canyon has been attributed to various factors, including changing climatic conditions in the late 1000s. Chaco continued as one of the primary centers of ancient Puebloan life into the early 1100s and beyond. Nevertheless, the communities built in the Middle San Juan region, such as Aztec and Salmon, and those built farther north in the greater Mesa Verde region, indicate a change in the focus of activities and a broader geographic spread of Chacoan and post‑Chacoan culture by the early 1100s. Salmon and Aztec were deliberately built in fertile, alluvial valleys next to some of the largest rivers in the northern Southwest. Given the development of water management techniques in Chaco during the 1000s, it is not surprising that Chacoan movement northward focused on areas where these newly developed technologies could be implemented on a larger scale. Indeed, the available evidence indicates that both Salmon and Aztec produced large quantities of corn; in the case of Salmon, some of this corn may have been exported as ground meal (Shelley 2006). The Chacoan grinding rooms discussed below, each containing 6 to 12 milling bins, indicate large-scale production of corn meal, presumably some of which was exported in trade. Further, the area around Aztec has evidence of at least two ancient irrigation ditches, first documented by John Newberry (1876) during an 1859 expedition. Based on my research at Salmon over the last six years, I do not think the creation of a ­ritual-​ ceremonial­structure was the primary or sole reason for the construction of Salmon by the Chacoans. I think that Salmon was, in fact, a pueblo: a ­residential 43

Paul F. Reed

Brief Historical Context of Chacoan Great House Functional Studies

town on the San Juan River (Reed 2004a). The evidence to support this assertion comes in several parts. First, as discussed above, Salmon represents a unique great house in the Chacoan world because of its size and placement in the Middle San Juan region. Second, its floor plan and arrangement of room suites is rare among excavated Chacoan great houses. Salmon closely resembles Hungo Pavi in Chaco Canyon, but the latter site has not been excavated, so we lack data beyond the basic floor plan. Third, the excavated sample of rooms at Salmon contains many floor features (alone or in combination with others) that support residential usage. Fourth, analysis of faunal remains from numerous rooms provides strong evidence of a residential pattern (see Durand and Durand, 2006, this volume). Finally, patterns of trash disposal during the Chacoan era at Salmon indicate residential usage of rooms across much of the pueblo. In this chapter, then, I explore residential aspects of Salmon and address the idea that all Chacoan great houses were primarily non-residential. With the exception of Gwinn Vivian (1990, this volume) and perhaps a few others (Baker, this volume), most Chacoan archaeologists consider great houses, in the Canyon and outside, as predominantly non-residential­ structures. Toll (this volume), for example, highlights the “lack of evidence for habitation in the great houses.” Salmon has been put in this non-residential category— lumped, as it were, with most other Chacoan great houses. Salmon, however, has explicit evidence of residential use in its feature and artifact assemblages. Thus my goal in this chapter is to present the data that support residential usage of Salmon during the Chacoan period. Beyond the question of residential usage, Salmon Pueblo is well positioned to play an important role in answering questions of compelling interest for Chacoan and post-Chacoan archaeology. What was the primary purpose and function of Chacoan great houses? Second, what functions were most important for these structures? Finally, is there any reason to assume that great houses had the same functions everywhere in the Chacoan world?

Debate regarding the uses and purposes of the great houses in Chaco Canyon began nearly simultaneously with their rediscovery by Euro-Americans­in the 1840s (McNitt 1964; Simpson 1850). In truth, the earliest explorers and archaeologists had ­little reason to doubt that the Chacoan structures were used for residence. The sites were large, well built, and appeared very similar to the extant pueblos of the day (see historic documents in Reed 2004a:​ 114–​164). Rather than debate the function of the buildings, Chaco’s first few generations of archaeologists focused primarily on describing the architecture and exploring questions of ritual, population, social organization, and migration—wondering to where the people who built the huge, empty structures had migrated. Somewhat more sophisticated methodologies­ (and the resultant data) emerged in the early-to-​mid twentieth century with the work of Judd at Pueblo Bonito, and Hewett and his colleagues at Chetro Ketl (Hawley 1934; Hewett 1936; Judd 1954). The questions changed, as well, as the years of work began to paint a picture (albeit incomplete) of life in Chaco Canyon. A greater concern emerged about the effects of the environment on Chaco’s inhabitants, particularly regarding the impact of drought on subsistence. Gwinn Vivian’s work in Chaco Canyon was the first to combine an explicit scientific approach with a concern for understanding the primary function of great houses and small houses (Vivian 1970). Although prior work had addressed aspects of social organization, Vivian’s work was the first to attempt a comprehensive explanation for the rise of Chacoan society and social organization, and to relate it to other factors (for example, the use of an extensive system of canals and other facilities to collect and channel runoff from rainfall on the north side of Chaco) (Reed 2004a). Building on the earlier work of his father, Gordon Vivian, and that of Clyde Kluckhohn, Vivian proposed that two distinct ethnic-social groups lived in Chaco. A group described as Chacoan–San Juan (derived from the

44

Salmon as a Ritual and Residential Great House north) built and lived in the great houses, while another social group (identified as the Cibolans) built and occupied small houses. Vivian viewed the great houses as residential units that were large enough to accommodate big groups of people, following what he viewed as a long-standing pattern of aggregation among the San Juan Puebloans (Vivian 1989, 1990). The Chaco Project (dating from the early 1970s, with some reporting still underway) dramatically changed the way Chacoan great houses were viewed. Among many other projects, Pueblo Alto was a primary focus of the Chaco Project’s research (Windes 1987a). Work at Pueblo Alto resulted in a dramatically different interpretation of Chacoan site function and, by extension, the function of all Chacoan great houses (Toll 1985; Windes 1987a). Limited discovery of floor features, particularly hearths used for heating or cooking, led Chaco Project researchers to infer a largely non-residential­ function for Pueblo Alto. These data, combined with apparently unusual patterns of trash deposition in middens, interpreted as evidence of periodic feasting and ritual destruction of pottery, suggest a unique and special function for Pueblo Alto. The convergence of roads at the site and its location on the mesa above Pueblo Bonito led researchers to infer a gateway function for the site, without a sizeable residential population. Of late, studies have deconstructed Chacoan great houses in a variety of new and novel ways. I mention only a few here to set the stage for a consideration of Salmon’s attributes. Cooper (1995), Bustard (1996), and Bernardini (1999) completed studies attempting to understand the usage of space and function at Pueblo Bonito, specifically. Bernardini’s work, in particular, suggests a very low residential population for Pueblo Bonito of 72 individuals (at peak). More recent work has continued to focus on Pueblo Bonito, with the goal of understanding this “place of unprecedented power” (Neitzel 2003a:1). The dedicated Smithsonian volume on Pueblo Bonito brought together a number of scholars to explore multiple aspects of the Chacoan world (Neitzel 2003b): ritual (Farmer 2003); burials and

45

burial patterns (Akins 2003); labor (Metcalf 2003; see also Wills 2000); artifacts (Mathien 2003); architectural and construction issues (Marshall 2003; Windes 2003), and residential usage (Bustard 2003). The latter study by Bustard is particularly useful as we look at Salmon Pueblo’s unique history. Among her conclusions, Bustard (2003:82) reasons that “great houses...display wide-ranging spatial and functional diversity, making single-use models unpersuasive.” In my view, this single sentence goes to the heart of the issue: we should not reduce the complicated Chacoan great house structures by invoking simple, exclusive determinations of ritual, residential, administrative, palatial, storage, or other functions. Although my goal in this chapter is to explore the residential aspects of Salmon’s Chacoan occupation, in so doing, I acknowledge the complexity of the Chacoan adaptation at the site and make clear that the ritual usage of Salmon was as critical to its history as its residential functions (for discussions of Salmon’s ritual importance, see the chapters in this volume by Baker; Durand and Durand; and Van Dyke). However, I will focus primarily on Salmon’s residential aspects and leave greater discussion of its ritual implications for another paper.

Room Suites at Salmon Prior to delving into the feature database, it’s important to note an aspect of architecture that characterizes Salmon Pueblo and makes it unique in the Chacoan world: its obvious arrangement of room suites. Although other Chacoan sites exhibit a broadly similar pattern (for example, Aztec and Hungo Pavi), none are as clear. Steve Lekson (1986, 2007) has described early room suites for Pueblo Bonito and other Chacoan sites consisting of one large room and two connected smaller rooms; however, the building of such suites ceased prior to ad 1030 or 1050, when Bonito’s inferred function shifted away from residence to ritual and other concerns. Lekson seems to favor the change in function argument, noting “either the formal concept of suites was changing and becoming less patterned, or many of the later structures were not

Figure 3.1.  Map showing Chacoan (Primary) occupation at Salmon Pueblo.

Salmon as a Ritual and Residential Great House designed for the functions that were housed in the paired-room suites and probably continued in the linear suites” (Lekson 2007:​29–​30). On this basis, then, the construction of obvious room suites at Salmon at ad 1090 would signal use of the site as a primary residence. At Aztec’s West Ruin, Earl Morris’s work revealed a pattern of room suites in the north and east wings (see all Brown et al., this volume). Rooms similar to Salmon’s Type 2 rooms were arranged in suites and interconnected via doorways (Morris 1919a). In contrast, the west wing of Aztec’s West Ruin had much smaller room sizes and no suites comparable to those at Salmon. Furthermore, recent architectural work at Aztec’s East Ruin indicates it does not have rooms suites with a comparable pattern either (Gary Brown, pers. comm., 2005). With some variation, Salmon’s room suites consist of a large, square, single-story, plaza-facing room connected via a doorway with a smaller, narrow room behind, which, in turn, is also connected via a doorway to rooms behind (Figure 3.1). These suites are defined specifically for Salmon’s first story. With second and third stories across much of the pueblo, the suites likely included upperstory rooms and may have been as large as 8 to 10 rooms. Unfortunately, we have little data from upper floors at Salmon because of poor preservation and the collapse of these upper stories well before the site was excavated in the 1970s. An example of a typical room suite (from north to south in Salmon’s northeast wing) includes Room 100W (large, square residential room), 97W (a specialized milling room), Room 101W (with three hearths and residential usage), and Room 76W (back wall, probably storage room). Similar suites span the known portion of the pueblo. The middle portion of Salmon’s west wing was substantially modified by the later San Juan inhabitants, with original Chacoan walls removed and new walls built according to at least two different layout schemes. However, there is no reason to suspect that the west wing was substantially different from the rest of the pueblo. During excavations at Salmon, crews working on the west wing found evidence of the original Chacoan room walls and foundations,

indicating that the pattern seen elsewhere in the site was probably present in the west wing also. The large, square rooms did not usually include doorways providing access to adjacent square rooms, although exceptions occur (between Rooms 100 and 103, and 46 and 92). These latter examples are best interpreted as expanded room suites consisting of two parallel rows of ground floor (and probably upper floor) rooms. In summary, Salmon’s floor plan reveals a clear pattern of suites created to maximize the use of space, with one large, square residential room in front, backed by rooms of smaller length but comparable width (P. Reed 2006c). These suites represent logical, apartment-like units that were utilized by the families in residence at Salmon. Examination of the site map shows some variability within the basic pattern (see Figure 3.1), but despite this variability, it is clear that the suites were a planned aspect of the pueblo. Following Cynthia IrwinWilliams (this volume, 2006b), I infer that the residential suites were occupied by extended family groups. The interpretation of Salmon as a residential site, then, rests on the architectural suites as the foundation.

Room Features at Salmon Features have been identified in every excavated (and many unexcavated) rooms at Salmon Pueblo. The final count of features is 1,719. Of these, 44 percent (n = 763) date to the Chacoan occupation. A simple majority of the features are thus associated with the later, San Juan occupation of the pueblo. To make sense of the vast number of features, I employed a two-level feature typology (Table 3.1). The first level consists of seven large categories: ritual, residential, architectural, access, burial, construction, and unknown/undetermined. Specific features were then assigned to these larger categories. Hearths, for example, are considered residential features. Chaco age features are present in at least 85 of the 300 rooms at the site. Many of these are doorways, with more than 125 scattered throughout the pueblo. Many have not been uncovered because the rooms are unexcavated. Excluding doorways, Chacoan features are present in almost every room 47

Paul F. Reed

Figure 3.2.  Distribution of Chacoan floor hearths in rooms at Salmon Pueblo.

e­ xcavated. My primary concern here is with features that are indicative of residential usage. To explore evidence of residential use at Salmon, I draw on a sample of rooms with evidence of residential usage during the Chacoan and San Juan occupations. Detailed feature data have been compiled for these rooms, and I discuss these data herein.

Table 3.1.  Generic Categories Used in Salmon Feature Analysis

Residential and Storage Features

I have classified numerous features from Salmon into the residential category. Here, I consider three residential feature types: hearths, milling bins, and storage features. I also discuss burials, which I put into a separate category. Following many other archaeologists, I consider the presence of a hearth as the primary indication of residential function for a Chacoan room. The graph in Figure 3.2 shows the Chacoan rooms in the sample that have hearths. Nineteen different rooms have a total of 59 hearths, and 13 rooms have at least two hearths at different levels. These hearths vary from highly formalized, slab- and/or adobelined features, to shallow pits with heavy oxidation. As shown in Figure 3.3, the Chacoan rooms with hearths are distributed around the pueblo. Most of the big, square rooms that have been excavated at Salmon that I interpret as the primary residential rooms have hearths. Some, such as Room 86W,

Feature Type

Examples

Ritual Residential Architectural Access Burial Construction Unknown/   Undetermined

sipapus, foot drums hearths, storage pits, milling bins benches, pilasters, racks doors, vent windows burials vigas, latillas, postholes, posts pits

do not contain hearths. Such an absence, however, should not be considered definite because the intensive use of certain Chacoan rooms by later San Juan occupants may have removed all traces of the earlier occupation. Room 86 has over 100 documented features, but its only clearly Chacoan features were doorways. The larger, slab-lined hearths that were found in most of Salmon’s Type 1 rooms probably functioned as cooking and heating facilities. Smaller hearths without formal rock lining were used for heat and to provide light for rooms tucked into the back of the pueblo. Quite a few of the rooms located behind the square rooms also have hearths, including Rooms 31, 33, 97, 101, 123, 127, and 129. If hearths are a good

48

Figure 3.3.  Map of Salmon showing Chacoan rooms with hearths (shaded).

Paul F. Reed

Figure 3.4.  Distribution of Chacoan milling bins in rooms at Salmon Pueblo.

indication of “living” rooms, then several rooms not in the large, square category also served this purpose (for certain periods of time). In sum, the hearth record in Chacoan levels at Salmon is considerable and, in my view, supportive of residential use of the site. For comparison, 44 hearths dating to the San Juan occupation are present in the same set of rooms. Thus, using hearths as an indicator, there is more evidence of residential usage during the Chacoan occupation of these rooms than during the later San Juan occupation.

utilized flat, slab metates and were completely enclosed in slabs to hold the ground meal. Beyond these structural differences, the ­major finding of Shelley’s milling study was that Chacoans had fewer but larger grinding facilities than the later folks. Figure 3.4 shows the frequency of milling bins in Chacoan rooms at Salmon. Figure 3.5 shows the distribution of milling bins in Chacoan rooms. Four Chacoan rooms (82W, 93W, 97W, and 129W) had the majority of milling bins, a total of 28, and a mean of 6.75 features per room. In contrast, grinding features were more spread out during the Secondary occupation, with nine rooms containing at least one bin, a total of 14 bins, and a mean of 1.55 per room. The Chacoan mills were clustered east of the Tower Kiva and along the east wing. San Juan mills, conversely, were spread across the site, from the west wing, west and east of the Tower Kiva, and down the east wing. Shelley noted that this pattern of clustered milling was present at Aztec and other Chacoan sites, an interpretation that suggests a more centralized level of organization for the Chacoans. This conclusion is supported by other findings from the site, including its design and construction. Additional research conducted as part of this project has revealed more milling bins than

Milling Bins

Phil Shelley compiled initial data on Chacoan milling bins for the 1980 Salmon report and his 1983 dissertation (Shelley 1983, 2006); I have drawn upon this work, and have also added to it. Briefly, Shelley’s primary emphasis in discussing milling bins was contrasting the Primary and Secondary occupations. His research showed a difference in the placement and quantity of milling bins. Primary Chacoan milling bins, which Shelley termed “grist troughs,” consisted of trough metates set in place and separated by side-placed slabs. These features lacked rear slabs because the trough held the ground corn meal. The later, Secondary occupation (late twelfth and thirteenth centuries) milling bins 50

Figure 3.5.  Map of Salmon showing Chacoan rooms with milling bins (shaded).

Paul F. Reed

Figure 3.6.  Distribution of storage features in Chacoan rooms at Salmon Pueblo.

r­ eported by Shelley. Four additional Chacoan floors have one milling bin each: 92W, 100W, 121W, and 124W. These single unit facilities support the idea of family-level milling in the large, square rooms at Salmon. In contrast, the clustered milling facilities indicate production of corn meal on a much greater scale, perhaps in conjunction with ceremonial activities or for export to other Chacoan sites. Thus, the data indicate that both family and supra-family levels of corn milling occurred at Salmon Pueblo.

linked to these features were decentralized, at the discretion of family groups. The distribution also suggests that storage facilities were necessary across the pueblo. By comparison, there are 10 documented storage features in the same group of rooms for later San Juan floors. In addition, Rooms 30a and 31a contained substantial quantities of burned corn on two separate Secondary occupation floors. Storage features for the San Juan occupation are also spread across the pueblo without any significant clustering. Storage features, then, as currently understood for the site, occurred in similar frequencies in both occupations—a pattern that supports a residential function for the pueblo.

Storage Features

Clearly identifiable storage features are not particularly common in either occupation at Salmon Pueblo. Twenty-five storage features have been identified in the Chacoan sample, from 12 of the rooms. Figure 3.6 shows the frequency of these features, and Figure 3.7 shows their distribution across the site. These features include cists, bell-shaped pits, subfloor pits, and other pits with specific evidence of a storage function. Rooms with definite storage features are spread around the pueblo: in the main north block, east and west of the Tower Kiva, and in the southwest and southeast corners. The pattern of storage features is widespread across the site, unlike the clustering seen in milling bin locations. This suggests that the storage functions

Trash Deposits and Middens Salmon Ruins lacks a large, extramural trash midden or mound. Irwin-Williams suggested that the San Juan River may have removed or covered such a structure, if it did exist at the site. Because a channel of the river was once within 20 m of the site, this seems like a good possibility. In terms of trash deposition in and around the pueblo, numerous trash deposits of Chacoan age were uncovered during excavations at the site. In many cases, these features were not assigned 52

Figure 3.7.  Map of Salmon showing Chacoan rooms with storage features (shaded).

Paul F. Reed

Figure 3.8.  Distribution of trash deposits (> 30 cm) in Chacoan rooms and units at Salmon Pueblo.

s­ eparate feature numbers. Nevertheless, six such features and deposits were recorded (Table 3.2). Three of the rooms are residential-living loci (6W, 96W, and 100W), whereas the remaining three include two back storage rooms (83W and 84W) and one room of ritual importance (82W). For comparison, discrete trash pit features were recorded in three San Juan age rooms at Salmon: 31W, 121A, and 128A. Excavation of numerous units in Salmon’s plaza did not reveal extensive Chacoan age trash or midden deposits. Scattered, limited trash was found, but nothing substantial. During the later San Juan era, the plaza was used more extensively for trash deposition Excavations in rooms across the pueblo revealed 125 separate strata representing structured trash midden deposits of Chacoan age. Looking more carefully at trash deposits, we can identify the thickest and most discrete episodes of trash dumping by stratifying the sample. Taking a minimum of 30 cm of thickness or depth, the number of identified trash strata is 51 in 13 Chacoan rooms. Figure 3.8 shows the frequency of these strata. These vary from thin layers of several centimeters to more substantial (up to 0.50 m) deposits representing trash accumulations over years. Three rooms (91W,

Table 3.2.  Chacoan Age Trash Pit Features at Salmon Pueblo Room

Feature No.

006W 062W 082W 083W 084W 096W 100W

26 127 106 5002 none 5109 5032

100W, and 129W) containing numerous structured trash deposits were used as primary middens during the Chacoan occupation. Figure 3.9 shows the distribution of structured trash strata across the site. Chacoan age trash deposits are scattered and do not reveal any clustering. If these deposits are representative of residential use, the pattern reveals residence across the pueblo. By comparison, San Juan levels at the site contained 174 strata (greater than 30 cm in thickness) coded as structured trash. Given that the San Juan occupation lasted for more than 150 years, five times longer than the 30-year Chacoan era, the fact that about four times the number of trash strata were 54

Figure 3.9.  Map of Salmon showing Chacoan rooms with trash strata (> 30 cm) (shaded).

Figure 3.10.  Map of Salmon showing Chacoan rooms with burials (shaded).

Salmon as a Ritual and Residential Great House found in the San Juan levels is not surprising. In addition, the available evidence indicates that the Chacoans left Salmon in a planned fashion, even removing metates from milling bins.

would be expected given residential usage. Jeff Dean (1970:​168) commented on the dearth of burials at the late Pueblo III Kayenta sites of Betatakin and Keet Seel. Inscription House (also in Kayenta) produced a paltry 33 burials. Similarly, many late Pueblo III sites on Mesa Verde have limited burial assemblages. At Mug House, for example, Rohn (1971:​87) reported 46 burials, noting that “certainly, we did not find the remains of all or most of the people who lived and died at Mug House.” Rohn speculated that many burials may have been removed by nineteenth-century pot-hunters prior to any scientific work at the site. Certainly, rockshelter and cliff dwelling sites present different challenges for burying people, but if we believe the idea that “Pueblo people bury their dead where they live,” burials should have been found in greater quantities at all of these sites. The fact that many sites across the ancient Pueblo Southwest lack sufficient burials to reflect actual residential population suggests two things: (1) Pueblo people did not always or perhaps even frequently bury their dead where they lived; and (2) we have to look for another explanation for the absence of burials beyond simply reducing our population counts. To further explore Salmon’s burial population (in both the Chacoan and San Juan periods), I estimated the total number of people who had died at Salmon over the 200-year occupation. This work draws on a paper by Schlanger (1992) that estimated a death rate of 4 to 6 people per 100 every year in a pueblo dwelling. I used the low end, with an estimated death rate of 4 per 100. Table 3.3 illustrates the calculations for the three periods at Salmon (using a base population for each period based on the number of rooms occupied) and shows the actual number of burials recovered during excavations. In assessing the number of people who died and could have been buried at Salmon Pueblo, I think we can echo Rohn’s finding and state that excavations did not find most of the people who lived and died at Salmon over 200 years. The table reveals that a greater percentage of the San Juan residents who died at the site were buried at the site and subsequently recovered. Nevertheless, nearly 90 percent of the residents who died in all periods were not recovered during excavation.

Burials Four burials from confirmed Chacoan contexts have been excavated at Salmon. Figure 3.10 shows the distribution of these single burials, in Rooms 4W, 33W (the so-called Bow Priest), 100W, and 129W. In contrast, more than 80 San Juan burials are known. As potential evidence of residence, then, we see the most difference between the two occupations regarding burials. Although many archaeologists have pondered the lack of Chacoan burials for sites in Chaco Canyon and elsewhere (see Akins 2003), I do not think a good explanation is at hand. Salmon represents the largest dilemma in this regard. Based on the feature and room data presented here, Salmon Pueblo had a substantial residential population during its Chaco occupation. The discovery of only four burials is problematic and requires further exploration. Much has been made of the apparent lack of burials in Chaco Canyon sites. Some archaeologists, particularly Toll (this volume), would suggest that the relative lack of burials in great house sites is indicative of their non-residential usage. In this view, the Chacoans are believed to have buried their dead where they lived. Given the apparent lack of burials in great houses, proponents of this view conclude that the structures were not used residentially. This notion seems reasonable enough until we examine the burial data from a number of other sites, Chacoan and non-Chacoan alike. Indeed, the lack of burials to match inferred population levels is a problem across the ancient Puebloan Southwest. For example, in his work in the La Plata Valley, Earl Morris (1939:32) was astounded at the lack of burials in several sites with obvious and abundant evidence of habitation and residential usage, most notably Morris 33 and 39. Regarding Site 39, Morris noted that few intact burials were revealed; most were highly eroded, and many fewer were encountered than he had expected. Other southwestern sites in the Pueblo II– III interval also “lack” the number of burials that 57

Paul F. Reed Table 3.3.  Estimated Salmon Population and Deaths by Period

Period Chaco Early San Juan Late San Juan

Estimated Population (average year)

Duration of Occupation

Projected Deaths

Individual Burials Recovereda

150 75 250

35 65 70

210 195 700

4 — 95

Note: Burials from the San Juan period were not divided into early and late; all are listed under Late San Juan. Projected numbers of deaths are based on a death rate of 4/year/100 people. a

Includes all burials-human remains (complete, fragmentary, and single bone finds) from the excavated portion of Salmon Pueblo

Thus, ­despite clear differences in burial practices between the Chacoan and San Juan periods, we are still unable to account for the majority of people who died at Salmon Pueblo over its 200-year history. Clearly, we need to look for other answers to address the burial dilemma. Although I cannot offer a definite answer to the “missing burials” issue, I would suggest that multiple factors are at play for Chacoan and nonChacoan­sites in the late Puebloan Southwest. First and foremost, preservation is highly variable across the Colorado Plateau from site to site, and it seems clear that many burials have simply decayed and eroded beyond recognition over the centuries of time. In this regard, the data from cave and rockshelter settings are particularly telling because even in many of these bastions of preservation, archaeologists have been unable to identify the number of burials predicted by other residential data. Second, looting and pot-hunting have irreparably impacted numerous sites across the Southwest, forever removing burials from the archaeological record. Earl Morris, for example, described many looted burials in Chaco Canyon during an early visit in 1895. Finally, archaeological work conducted over the last 100 years in the Southwest reveals that ancient Pueblo people buried their deceased in a variety of settings that went well beyond their living spaces: plazas, middens, behind the back walls of pueblos, at the base of cliffs, and in niches spread across a variety of topographic settings. Given the many settings that were employed, along with the variable factors of preservation and looting, I think it is safe to assume that archaeologists may never encounter many of these “missing” burials. I also

think it is a mistake to rely on the supposed absence of burials in Chacoan and non-Chacoan settings to infer a small or nonexistent residential population for any given site.

Co-Occurrence of Residential Features at Salmon As a final step, I examined the co-occurrence of residential features in Chacoan rooms at Salmon Pueblo. Figure 3.11 plots frequencies of feature types by rooms for 20 rooms with Chacoan features. Six rooms (1W, 4W, 5W, 30W, 62W, and 127W) have a residential feature of only one type. Six rooms (6W, 33W, 101W, 102A/D, 121W, 123W) have occurrences of two feature types. Finally, eight of the rooms (82A-W, 92W, 93W, 96W, 97W, 100W, 124W, and 129W) have three of four residential feature types that co-occur. Figure 3.12 shows the map locations of the 14 rooms with at least two types of residential features. These data indicate a strong relationship between the different types of residential features and their co-occurrence in Chacoan rooms. The big, square room class, interpreted here as the primary locus of residential activities at the pueblo, in particular, shows strong relationships among the different residential feature types. Of the 14 rooms that show co-occurrence of at least two feature types, six of these are Type 1 rooms. In fact, only one excavated Type 1 room at Salmon (Room 56W) does not show evidence of residential usage. Clearly, ­Irwin-Williams and Adams’s original interpretation of these rooms as primary, residential areas during the Chacoan occupation at Salmon is supported. 58

Salmon as a Ritual and Residential Great House

Figure 3.11.  Distribution of co-occurring residential features in Chacoan rooms at Salmon.

Conclusions

year difference in occupational spans between the Research over the last six years has changed our groups. Milling bin features exhibit very differunderstanding of Salmon Pueblo and its place in ent patterns from the Chaco to San Juan compoChacoan and Puebloan history (P. Reed 2006a), nents, as Shelley’s (1983, 2006) work demonstrated. augmenting the work undertaken through the orig- In general, Chacoan milling bins occur in fewer inal Salmon Project and San Juan Valley Archaeo- rooms, and the rooms contain a greater number logical Program. Cynthia Irwin-Williams directed of bins, but single, family-type milling bin features these projects and began the involved process of were found on the floors of four Chacoan rooms. In understanding Salmon with her first visit to the site the rooms with concentrated milling bins, I think it in 1969. Her prescient understanding of the Chaco is reasonable to infer greater centralization of millphenomenon laid the foundation upon which all ing activities (perhaps for ritual purposes) for the of the work reported in this volume has built. Chacoans. Thus, regarding the use of Salmon as a The data presented in this chapter on site lay- Chaco residence, the milling bin data are equivocal, out and size, the nature of architectural suites, and supporting both a more centralized pattern of corn the exploration of room features at Salmon Pueblo grinding and family-level usage. during the Chacoan occupation indicate residenPatterns in the distribution and frequency tial use of the site. Salmon’s massive size and con- of identifiable storage features are similar for the struction in a largely vacant landscape point to the Chaco and San Juan occupations at Salmon. Clussite’s primary purpose for the Chacoan builders: tering or centralization of food storage facilities is use as a central, residential place in the Middle San not indicated for the Chaco population. In conJuan. trast, the data indicate that storage was necessary The suite of Chacoan floor features in Salmon for families located in all portions of the pueblo, rooms strongly indicates a residential pattern. supporting the notion of a residential site. The presHearth features are more common in Chaco strata ence of Chacoan trash deposits of varying quantithan in later San Juan levels, even given the 100-​ ties across the pueblo and concentrated in four 59

Figure 3.12.  Map of Salmon showing Chacoan rooms with co-occurring residential features.

Salmon as a Ritual and Residential Great House rooms used as “in-site dumpsters” also supports use of the site as a residence. In summary, Salmon was built as a residential Chacoan site around 1090 and was occupied by Chacoans until the 1120s. After the Chacoan leadership at Salmon ended, the pueblo began a transition to a local San Juan settlement. Cynthia Irwin‑Williams (this volume) believed that the drought that began around 1130 was a factor in the decline of Chacoan society, not just in the canyon but across the entire San Juan Basin. Certainly, the

drought played a role; however, changes at Salmon began in the 1110s and 1120s, prior to the onset of the drought. I have suggested that local conditions may have caused the Chacoans to leave Salmon and find their way to Aztec’s East Ruin in the 1120s (Reed 2006d). I think the Chacoans from Salmon, realizing that the location of the pueblo would not meet their needs, moved to Aztec and helped to build Aztec East, the symmetrical partner to Aztec West, in the 1120s and 1130s.

61

Pa rt 3

Material Culture Analyses: Salmon Pueblo

4

Subsistence and Plant Use During the Chacoan and Secondary Occupations at Salmon Ruin Karen R. Adams This chapter summarizes subsistence and plant use by Prehispanic occupants of Salmon Ruin, an outlying Chacoan great house community built along the San Juan River in northwest New Mexico beginning around ad 1090. The initial Chacoan occupation lasted for several decades until a significant drought within the region between ad 1130 and 1180 (Van West and Dean 2000) led to reduced population within the Salmon great house. By the early 1200s, the Secondary occupation had taken hold, with alterations to Salmon’s architecture and use of new pottery types. The pueblo was abandoned by this group in the late 1200s. The terminology for this chapter follows Irwin-Williams’s (1972, 2006a) original definitions of a Chacoan (Primary) occupation followed by remodeling during a Secondary occupation of Mesa Verde–affiliated­ ­people, with an intervening Intermediate period of minimal use between the two major occupations. The reader should note that this original chronology has been substantially revised during the new work at Salmon (see Reed, Chapter 1, this volume; P. Reed 2006b). The Salmon Ruin archaeobotanical program was initiated at the beginning of substantial excavations in 1972 to help assess the role of Salmon Pueblo in Chacoan society as outlined by IrwinWilliams (Chapter 14, this volume). In this model, Salmon Pueblo was built and overseen by religious personnel who were also central to the economic system. Irwin-Williams predicted that major Chacoan outliers such as Salmon would be surrounded 65

by extensive plots of farmland, highlighting the critical importance of agriculture in the diet. Following a synthesis of Salmon plant remains, I will summarize the highlights of this record in terms of the Irwin-Williams model. A number of professional and student researchers analyzed Salmon plant remains and produced a series of journal articles, books, master’s theses, and other manuscripts, all of which have served as the basis for this synthesis (Table 4.1). Early in the project, development of a standardized set of techniques and approaches for acquiring and analyzing plant remains provided guidance for excavators and analysts during the multiyear project (Bohrer and Adams 1977). Among the many methodological contributions, regularly occurring hexagonal, bullet-shaped items were eventually identified as termite fecal pellets (Adams 1984), revealing that termites infested roof timbers and fuel wood in prehistory, just as they do today. Many other contributions have been discussed in individual documents (Table 4.2). This synthesis incorporates archaeobotanical data from locations and strata within Salmon Pueblo that varied considerably in preservation. For example, Room 62W contained large quantities of organic remains from both the Chacoan and Secondary occupations, and it served as a trash dump and latrine during the latter portion of its use history. Alternatively, plaza samples represented the poorest preservation conditions at the site, with the highest levels of post-occupational

Karen R . Adams Table 4.1.  Publications and Master’s Theses Focusing on Plant Remains from Salmon Ruin Author(s)

Year

Title

Adams, Karen R.

1980b, 2006a 1980a, 2006b 2006c 2006d 1984

Conifers; Pines and Other Conifers from Salmon Pueblo

P P

2006a

Pollen, Parched Seeds and Prehistory; An Archaeobotanical Study of Room 93W at Salmon Pueblo Archaeobotanical Summary and Conclusions Native (Wild) Botanical Remains from Salmon Ruins Evidence of Wood‑Dwelling Termites in Archaeological Sites in the Southwestern United States Ceremonial and Medicinal Plants Ethnobotanical Techniques and Approaches at Salmon Ruin, New Mexico Introduction to Archaeobotanical Studies at Salmon Pueblo

2006b

Other Botanical Remains and Special Contexts

P

2006

Cultivated Plants from Salmon Pueblo

P

1979

A Study of Cucurbita Material from Salmon Ruin, New Mexico

MT

1976

A Preliminary Study of Wild Plant Remains Recovered by Flotation at Salmon Ruin, New Mexico Plant Remains Recovered by Flotation from Trash at Salmon Ruin, New Mexico; Plant Remains from Trash Deposits at Salmon Pueblo “Seeds” of Wild Grasses: A Major Food of Southwestern Indians Maize Variability and Cultural Selection at Salmon Ruin, New Mexico The Distribution Patterns and Fruit Productivity of Modern Juniperus Growing in the Salmon Ruin Area and the Archaeological Interpretation of Juniper Seeds and Cones Found in Salmon Ruin, New Mexico Utah Juniper ( Juniperus osteosperma) Cones and Seeds from Salmon Ruin, New Mexico

MT

Adams, Karen R. Adams, Karen R. Adams, Karen R. Adams, Karen R. Bohrer, Vorsila L. Bohrer, Vorsila L., and   Karen R. Adams Bohrer, Vorsila L., and   Karen R. Adams Bohrer, Vorsila L., and Karen R. Adams Bohrer, Vorsila L., and   John F. Doebley Burgess-Terrel,   Martha E. Doebley, John F.

2006 1977

Doebley, John F.

1981, 2006 1984 1983

Doebley, John F. Doebley, John, and   Vorsila L. Bohrer Lentz, David L.

Lentz, David L.

1979

1984, 2006

Status

P P P P P P

P P P MT

P

Note: P = publication; MT = master’s thesis.

contaminants (Doebley 1976), requiring caution in interpretation. Different researchers attempted to evaluate the effect of preservation on the remains being discussed (Adams 2006a; Bohrer and Adams 2006a) to better understand sample types or locations with similar preservation potentials. In the same vein, efforts were made to determine the role of rodents in the presence and use of plant parts during the Salmon occupations. Charred rodent feces in over half of trash flotation samples studied suggested that rodents occupied the Salmon community along with people and

were eating foods that people were discarding into middens (Doebley 1976). This was supported by the presence of rodent gnaw marks on piñon nutshells (Adams 2006c) and on some juniper seeds (Lentz 1979). The bulk of archaeobotanical effort focused on reproductive parts (fruits, seeds, etc.) and pollen grains, with less attention given to identifying wood utilized as fuels or tools. In this chapter, the term cheno-am is used as a hybrid category indicating seeds of either Chenopodium or Amaranthus, difficult to distinguish when burned and broken. 66

Berries provide 5,200–5,500 kilocalories per kilogram, provided by carbohydrates, fats, and proteins.

Revealed that dry-wood termites infested roof timbers and fuelwood in prehistory. Revealed the importance of diverse grasses to historic groups and, by ­inference, to Prehispanic groups as well. Principal components and discriminant analyses both segregated seeds of most species from each other, and seeds of subspecies as well.

Analysis of caloric content of juniper berries (cones), minus the indigestible stony seeds.

Identification of a widespread hexagonal unknown as termite fecal pellets.

Systematic classification of three ­domestic and one wild modern ­Cucurbita species through detailed study of seed traits.

Summary of extensive southwestern ethnographic record of grass grain use.

Identification of juniper seeds/ cones in site deposits based on wellsupported­data. Recognition of only Juniperus osteosperma present within 6 miles.

Seed size index values (length × width) calculated; overlap between Juniperus species determined. Confident identification of available juniper species within area as resources. Over half of this catchment area was covered in juniper woodland; piñon trees were not nearly as common as juniper trees.

Development of traits to distinguish modern seeds of Juniperus monosperma vs. Juniperus osteosperma. Transects within 6 miles of Salmon Ruin used to determine presence of juniper species in area. Calculation of woodland coverage based on a catchment study of juniper woodland available within 6 miles of Salmon Ruin.

All seeds/cones in Salmon Ruin ­identified as Juniperus osteosperma.

Result #3

Presence of modern stands supports morphological data for ancient use of only Juniperus osteosperma. Conclusion that juniper berries could The number of potential kilocalories produced per year by juniper berries = feed 52% of a population of 300 indi139,447,750 based on field estimates viduals for a year, eating ~2,430 kc/day. of berries (cones) /km2 in 1977, determination of caloric value, and factoring in of losses. Contribution to growing body of data on the energy contained within wild foods, useful in optimal foraging models (Barlow 2002); data on effort expended to gather and process the berries must still be gathered. Identification of an unknown reported Descriptions of many dry-wood in other locations in the American termite pellets and a key to their Southwest. identification. Ethnographic evidence that dependable grasses were often sought in times of food stress, with perennial grasses particularly important. Identification of all but a single com- Presence of burned rind and peduncles plete seed in Chacoan and secondary suggests roasting of mature fruit. Dried strata as Cucurbita pepo, although they strips of squash flesh indicate drying for storage. Same squash type used likely came from a different variety than the two modern ones examined through time at Salmon. (Connecticut Field and Small Sugar).

Result #2

Methodological Contribution Result #1

Table 4.2.  Methodological Contributions to the Ethnobotany of Salmon Ruin and Their Results

BurgessTerrel 1979

Doebley 1984, 2006

Adams 1984

Lentz 1979, 2006

Lentz 1979, 2006

Lentz 1979, 2006

Lentz 1979, 2006

References

Poorest conditions of preservation were those in which plant ­materials survived only by virtue of being parched or carbonized.

Bohrer and Adams 1977

Consistent sample acquisition and The training of hundreds of students, analysis throughout the long history of now practicing professionals, in the the Salmon excavation. routine acquisition of plant remains. Many are in governmental agencies that supervise archaeological excavations. Identification of noncharred piñon Development of criteria for recognizing parched piñon (Pinus edulis) shell nutshells in Salmon Ruin strata as from that brought in by other modes of cultural or non-cultural entry (e.g., rodents, etc.). Moderate conditions of preservation Identification of “best,” “moderate,” and Best conditions of preservation were “poorest” strata for preservation based indicated by the presence of uncarbon- included Yucca leaves (lacking fibers) ized Yucca fiber. and thin, fragile Cucurbita seeds (lackon the presence and condiction of ing interiors). plant materials.

Adams 2006a; Bohrer and Adams 2006a

Adams 2006c

Adams 2006c

Development of a thorough and organized way to identify the diversity of conifer parts (bark, berries, cone scales, needles, nuts, twigs, wood) recovered in Salmon Ruin strata.

Development of dichotomous keys for modern conifer parts (bark, needles, cones, cone scales, twigs, bark scales) of 11 species of southwestern conifers (Abies, Juniperus, Pinus, Picea, Pseudo­ tsuga). Development of standardized ethnobotanical techniques and approaches to Salmon Ruin plant remains early in excavation history.

Bohrer and Doebley 2006; Doebley and Bohrer 1983

Metate morphology suggested a focus on pop/flint types by Chacoans and on flour types by Secondary folks; other methods of preparation (parching, boiling) may have been in use.

Identification of the relative proportions of flour to flint/pop maize in Chacoan and Secondary occupations as similar, an unexpected outcome since 75% of the Chacoan metates were trough and 77% of the Secondary metates were slab.

Whole maize kernels from the Chacoan and Secondary occupations were examined and classified as to endosperm type.

References

Result #3

Result #2

Development of list of traits for classifying endosperm types using charred modern specimens of pop, flint, and flour maize.

Methodological Contribution Result #1

Table 4.2. (cont’d)  Methodological Contributions to the Ethnobotany of Salmon Ruin and Their Results

Subsistence and Plant Use

The Plant Evidence: Basis of Subsistence

For ease of reading, the use of the word “type” in the original literature cited here will not be carried over. However, the word “type” in conjunction with a taxon (e.g., Portulaca retusa type) informs the reader that an ancient plant part resembles the taxon named but also resembles other, similarlooking taxa. Other authors have independently analyzed and interpreted specific types of Salmon plant materials. Webster (2006) has described the manufacturing techniques and has identified construction materials in the perishable collection (basketry, matting, sandals, and other items) and has compared the Salmon perishables with those of other Chacoan great houses in Chaco Canyon and at Aztec Ruin (Webster, this volume, 2006). Windes and Bacha (this volume, 2006) focused on the chronological history, harvesting and preparation methods, and types of structural wood elements used to construct and repair Salmon Pueblo, in comparison to the same types of activities within Chaco Canyon and Aztec Ruin great houses. Windes and his colleagues reexamined some of the Populus species (presumed to be cottonwood) samples used for smaller construction elements at great houses and determined, based on anatomical details, that many of the specimens were aspen (also a Populus species) harvested at higher elevations (Tennessen et al. 2002). Because subsistence is tied closely to health issues, the research on parasites in human coprolites (Reinhard, this volume) as a general reflection of health is also relevant. Durand and Durand (this volume, 2006) have explored the contribution of fauna to the Salmon diet and have identified a shifting emphasis from hunting to food production through time. Although the use of lagomorphs (rabbits) remained fairly stable, use of artiodactyls (bighorn sheep, deer, elk, and pronghorn) was greater in the Chacoan period. Artiodactyl hunting declined in the Secondary period as focus on domestic turkeys increased significantly. Such a pattern appears common in the Northern San Juan region. Driver (2002) noted that by the Pueblo III period, families were eating less big game and were much more focused on raising turkeys for meat, relative to earlier Puebloan periods.

Cultivated plants were of major economic importance to both the Chacoan and Secondary occupants of the Salmon great house. Both groups grew maize, beans, and squash in their fields. They also gathered a wide range of wild plants available throughout the growing season, most of them as foods (Table 4.3). Crop failures or low productivity may explain the widespread recovery of cheno-am and Portulaca seeds, along with grass grains, cactus seeds, tansy mustard seeds, reproductive parts of sedge and other riparian taxa, and many other wild food resources. A comparison of the two occupations indicates that the Chacoan diet focused less on wild plants than the Secondary inhabitants. As a caveat, the Salmon archaeobotanical records provide no direct evidence of important resources such as leafy greens or other vegetative plants unlikely to survive in the archaeological record. It seems likely, therefore, that both occupations made use of these nearly undetectable plant resources.

Crops Maize

Zea mays cobs, cupules, ears, husks, kernels, pollen, stalks, and tassels were all recovered at Salmon Ruin. Prior studies of Salmon maize (Bohrer and Doebley 2006; Doebley and Bohrer 1983) provide a good overview of maize use. Carbonized kernels, cobs, and cupules were frequently recovered in flotation samples, indicating regular usage. Based on a study of 48 rooms, maize samples of greater than six cobs occurred in 22 units, primarily in trash and roof-fall. Although it appeared that a number of Secondary occupation second-story floors (firstfloor roofs) had associated maize when the roofs burned and collapsed, only one burned, first-story storage room (30W) contained maize cobs aligned parallel to the south wall in relatively horizontal positions (stacked for storage). A single stratigraphic unit of redeposited trash in Room 62W preserved stalks and tassels and empty tied husks, indicating that maize ears were once braided together by their attached husks. The presence of whole stalks suggests plants were sometimes cut and carried to the pueblo. It seems likely that tassels were abraded to 69

Table 4.3.  Plant Taxa Identified from Salmon Ruin Common Name Taxon

Part(s)

Use Category

Presence

seed seed fruit, leaf, seed achene, pollen seed achene all taxa, all parts pollen seed drupe

medicinal food food food? food food food ceremonial, food food, medicinal food

Rare Rare

cholla

Kallstroemia Cleome, C. serrulata Yucca baccata Eriogonum Shepherdia argentea Scirpus, S. acutus Cactaceae Typha Salvia reflexa Prunus serotina/ P. virginiana Cylindropuntia

pollen

food

clammy weed cocklebur common bean

Polanisia Xanthium saccharatum Phaseolus vulgaris

food food? food

cotton

Gossypium

crown beard Douglas fir dropseed grass fir or Douglas fir

Verbesina Pseudotsuga Sporobolus Abies concolor or ­Pseudotsuga Sphaeralcea, S. parviflora Chenopodium Chenopodium watsoni Chenopodium Cheno-am

seed involucre cotyledon, pod, bean (seed) fiber, cordage, textile fragment achene wood caryopsis bark

food building food building

Rare

pollen seed seed pollen embryo, seed

food food food food food

Rare

rind fragment all taxa, all parts bract pollen seed berry, nutlett

household need food food unknown food food

seed seed

food food

Rare

seed

food

Common

seed bark, seed, scale leaf, twig, wood

food food, building, fuel, medicinal/ ceremonial

Rare

Arizona poppy beeweed Broad-leaf yucca buckwheat buffalo berry bullrush cactus cattail chia chokecherry

globemallow goosefoot goosefoot goosefoot goosefoot  pigweed gourd grass grass tribe greasewood hackberry hawthorn hedgehog cactus horse purslane husk tomato inkweed juniper

Lagenaria Gramineae Paniceae Sarcobatus Celtis Crataegus saligna/ C. chrysocarpa Echinocereus, E. fendleri Trianthema portulacastrum Physalis, Physalis longifolia Suaeda, S. suffrutescens Juniperus

Comment

Common Rare

some large diameter Rare

clothing

clumped

clumped Widespread

Common clumped

Table 4.3. (cont’d)  Plant Taxa Identified from Salmon Ruin Common Name Taxon juniper (Utah)

Juniperus osteosperma

knotweed

Polygonum, P. sawatchense, P. persicaria Rhus aromatica Zea mays

lemonade berry maize

Mormon tea

Ephedra

pigweed pincushion cactus pine piñon

Amaranthus Mammillaria Pinus Pinus edulis

plantain ponderosa pine

Plantago Pinus ponderosa

prickly pear cactus purslane ragweed reedgrass

Platyopuntia Portulaca, P. retusa Ambrosia Phragmites

ricegrass

Stipa hymenoides

rush sagebrush screwbean sedge sedge family

Juncus, J. cooperi Artemisia Prosopis pubescens Carex Cyperaceae

spike rush spruce

Eleocharis montana Picea engelmanni, P. pungens Euphorbia glyptosperma Cucurbita Cucurbita pepo

spurge squash squash

stickleaf sunflower sunflower tansy mustard

Part(s)

Use Category

Presence

Comment

berry (cone), seed food, adornment, ritual achene food? Rare stone (seed) pollen, tassel

food food, ceremonial

cupule, cob, embryo, kernel, stalk stem

food

medicinal/ ceremonial seed food pollen food pollen unknown nut, nutshell (testa), food cone scale, twig seed food bark, needle fuel?, building, medicinal/ ceremonial? pollen food seed food pollen unknown stem fragment ceremonial (­cigarettes) bract, caryopsis, food floret achene food pollen unknown seed food achene food pollen food all taxa, all parts food achene food bark scale ceremonial

seed pollen fruit flesh, rind (exocarp), ­peduncle, seed, vine tendril Mentzelia albicaulis seed Compositae (long spine) pollen Helianthus pollen Descurainia, D. obtusa seed

Rare Rare

as Rhus trilobata parched, ­aggregates

clumped

Widespread clumped

as Oryzopsis hymenoides Rare clumped

Common

food, ceremony food food

Common

food food food food

Rare clumped Common

Karen R . Adams Table 4.3. (cont’d)  Plant Taxa Identified from Salmon Ruin Common Name Taxon tick clover umbrella wort unknown wild onion winged pigweed wolfberry yerba mansa yucca

Desmodium Allionia unknown Allium macropetalum Cycloloma atriplici­folium Lycium Anemopsis californicus Yucca sp.

Part(s)

Use Category

Presence

seed pollen variety of parts bulb scale seed seed seed leaf

unknown food? unknown food food food medicinal food, household needs

Rare Rare

Comment

Rare Rare

Note: Table was assembled from all resources listed in Table 4.1. It does not include taxa naturalized from other continents (e.g., Salsola, Sisymbrium altissimum) or assumed to have been incidentally carried into the pueblo by wind. In the “Presence” column: rare = in 1–20 of 64 locations (Adams 2006a); common = in 23–35 of 64 locations; widespread = in > 50 of 64 locations.

remove pollen for use in some rooms, including those adjacent to the Tower Kiva, where numerous examples of aggregated maize pollen and high concentrations of maize pollen were identified. Two human coprolites from the Secondary occupation have small cob segments embedded in them, suggesting cobs were eaten in times of food scarcity, similar to an ethnographically reported Pueblo practice (Hill 1938:45–46). The frequent recovery of cob fragments and cupules from fire pits indicates the use of leftover cobs as fuel.

1979). Prior to a detailed study of whole squash seeds (n = 84) associated with the Chacoan occupation and whole seeds (n = 168) from the Secondary occupation, Burgess-Terrell used principal components and discriminant analyses to develop a systematic method of separating three modern domesticates (Cucurbita pepo, C. moschata, and C. argyrosperma) as C. mixta, and one wild cucurbit (C. foetidissima) using detailed seed measurements. When applied to the Salmon seeds, the results suggested that all but a single seed of both occupations belonged to Cucurbita pepo. In addition, it Squash appears that more than one variety of squash was Cucurbita (squash) seeds, rind (exocarp) fragments, grown, and these varieties differed from the two dried flesh, peduncles, and vine tendrils were all modern varieties of Cucurbita pepo (Connecticut preserved in Salmon deposits (Bohrer and Doeb- field pumpkin and small sugar) included in this ley 2006; Burgess-Terrel 1979). Most of these were study. Seven Cucurbita pepo peduncles (stems) recovered as macrofossils, visible to excavators, and were also identified, some of them burned. virtually none were identified in flotation samples. Coupled with burned squash rind charred on In one study of trash units of moderate preserva- the exterior but not the interior, it seems that mation (Bohrer and Adams 2006b), the presence of ture squash fruits were occasionally roasted whole, squash seeds suggested a relatively high ranking as a a known historic preparation method for this spefood source, exceeding the frequency of piñon nut cies of squash (Castetter and Bell 1942:190). Other shells, juniper seeds, onion bulb scales, yucca seeds, uncharred rind fragments suggest other preparaand chokecherry seeds. Trash and roof units most tion methods. The ethnographic record also indioften retained squash evidence, along with four cates that although seeds of some squash varieties Chacoan floors and one Secondary floor. The bulk were eaten, those of Cucurbita pepo rarely were of evidence was recovered as seeds and uncharred because of their comparatively small stored food exocarp (rind) fragments. value (Castetter and Bell 1951:114). This finding Modern squash provided the basis for identify- helps explain the presence of squash seeds in the ing the squash recovered at Salmon (Burgess-Terrell­ Salmon trash record, where squash seeds were 72

Subsistence and Plant Use o­ ften removed from their fruit and discarded. Two clusters of squash seeds still clumped together were removed in bulk from the fruit. However, the presence of squash seeds in human coprolites (Bohrer and Adams 2006b) reveals that the seeds were eaten at least on occasion. Segments of curved, dried squash flesh from Room 62A/W reveal that the practice of drying squash has considerable time depth in the ancient Pueblo Southwest. When comparing squash seeds of both occupations within a well-preserved room (62W), the fruits were considered to be similar, suggesting continuity of this variety of squash through time.

lacking are the cotton seeds, bolls, and other parts of the plant that would indicate local cultivation. If people had access to cotton seeds, they likely would have parched some for consumption. Evidence for weaving cotton cloth appears to be very restricted in the San Juan Basin relative to the Kayenta, Hohokam, Sinagua, and Salado regions to the south (Webster, this volume). Webster reports that imported cotton fiber may have been woven in limited amounts in the Chacoan ritual core of the pueblo in Room 81W, where four weaving tools were found. However, she feels that most of the textile production focused on Yucca (yucca) baskets, sandals, and mats, and on turkey feather blankets and rabbit fur robes. Some cotton cloth may have been acquired in trade.

Beans Evidence of domesticated common beans (Phaseolus vulgaris) in 21 of 48 rooms belongs almost exclusively to the Secondary occupation (Bohrer and Adams 2006b). Three potential Chacoan strata with common beans contain evidence of contamination with Secondary occupation trash. Most of the bean evidence came from trash and roof strata, and two roofs (100W and 127W) apparently served as outdoor preparation areas for beans and other foods. Beans were typically recovered as burned cotyledons (seed halves), although others preserved whole or were even still in their pods. The presence of beans in six of seven Secondary occupation storerooms suggests beans may have ranked second to maize in importance of cultivated plants.

Amaranth Two instances (one Chacoan, one Secondary) of poorly preserved seed coats hint that a domesticated grain amaranth may have been in use at Salmon (Bohrer and Doebley 2006). However, this is not considered reliable evidence of cultivated amaranth.

Wild Plants as Foods

Many of the reproductive plant parts recovered in Salmon strata were likely used as foods (see Table 4.3). This suggests that wild plant gathering was a regular activity of considerable economic importance, and not merely a means for these agricultural people to survive times of food scarcity (Doebley 1976, 2006). A broad overview of wild plant use at Salmon was provided by a study of 216 flotation samples from 28 room floors and 36 trash units (Adams 2006a). Only floors that were well-preserved, undisturbed by rodents or roots, adequately sampled, and not overlain by trash were included in this analysis. All six Chacoan floors studied were unburned, as were 11 of 22 Secondary floors. Thirteen Chacoan trash strata and 23 Secondary trash strata were included. Evidence for subsistence use included condition (charred), nature of associated pollen evidence, context, and association with other known foods. Each taxon was also evaluated for the likelihood of incidental introduction into the Salmon

Gourd Clear evidence of Lagenaria use comes only from the Chacoan occupation (Bohrer and Doebley 2006). A single fragment of rind from Room 62W has the unique shape of a dipper gourd ­handle. Since no gourd seeds were recovered, this most likely represents a trade item. Cotton Gossypium fiber, cordage, and textile fragments were found in very limited amounts, indicating cotton was known to the Salmon inhabitants (Bohrer and Doebley 2006). Small pieces of cloth and cordage segments were preserved in Chacoan trash in Room 62W and in Secondary contexts. Notably 73

Karen R . Adams community. Taxa considered foods fell into one of three categories, with widespread and common taxa associated with both Chacoan and Secondary occupations, and rare taxa most often associated only with the Secondary occupation.

by Adams (2006a). A look at plant remains from three Room 93W floors (Adams 1980a, 2006b) was clearly able to associate widespread and common seeds associated with metates as indicating grinding (tansy mustard, pigweed, purslane, goosefoot, stickleaf ), and seeds recovered from within hearths as suggesting parching prior to consumption (tansy mustard, pigweed, goosefoot, stickleaf ). Both piñon nuts and juniper berries (cones) may have been common foods, although the evidence at Salmon Pueblo is contradictory. Few plants rank higher for energy return (1,200–1,700 kilocalories) per hour of labor than piñon nuts (Barlow 2002: Figure 5), and the digestible portion of juniper berries provides 5,200 to 5,500 kilocalories per kilogram, enough to suggest the presence of fat (Lentz 1979:​112). At Salmon, however, no evidence of either resource was preserved in 216 flotation samples from 64 separate floor and midden locations (Adams 2006a), nor in additional midden samples analyzed by Doebley (1976, 2006), unless the very best and worst conditions of preservation are included. In contradiction, parched ­piñon nuts or nutshell fragments were preserved as macrofossils in over 50 percent (25 of 48) of excavated rooms (Adams 2006c), and juniper berries (cones) were frequently recovered in conjunction with fragile squash seed evidence (Lentz 1979, 2006). More than 80 percent of 53 well-preserved strata in Room 62W contained either piñon or juniper evidence (Bohrer and Adams 2006b). The fact that they are routinely absent from floors may be due to their larger size, making them easy to retrieve when spilled. The fact that they are routinely absent from flotation samples within middens suggests that under relatively poor conditions of preservation, they degrade. It appears that both size and preservation conditions can notably influence the recovery of piñon and juniper remains.

Widespread Cheno-ams and Portulaca seeds recovered from 50 or more of the 64 locations studied (Adams 2006a) represent plants of disturbed habitats, such as fallow agricultural fields and trash heaps, which are able to produce quantities of both greens and seeds. Cheno-am and Portulaca seeds were also found in nearly all the trash samples examined from a separate study of 17 rooms and plaza test pits (Doebley 1976:103, 2006), and from three habitation floors within Room 93W (Adams 1980a, 2006b). Charred cheno-am seeds adhering to the interior of a large corrugated sherd in storage on a roof (Room 90W) suggests occasional storage of the seeds (Adams 2006a), and their presence inside human coprolites indicates consumption (Bohrer and Adams 2006b). The frequency of clumps of cheno-am pollen grains on floors could indicate that young Chenopodium or Amaranthus plants in the flowering stage were harvested as greens (Bohrer and Adams 2006b). Common In 23 to 35 of 64 locations studied by Adams (2006a), a number of taxa were commonly encountered. These included grass grains (Muhlenbergia, Stipa hymenoides [as Oryzopsis hymenoides], Paniceae/​Panicum, and Sporobolus), tansy mustard seeds (Descurainia), seeds of various cacti (Echinocereus, Opuntia, and Sclerocatus), spurge seeds, groundcherry seeds, and reproductive parts of a number of riparian resources (Cyperaceae, Carex, Eleocharis montana, Scirpus/Scirpus acutus). In his separate trash study, Doebley (1976, 2006) also noted that wild grasses (n = 12) were recovered more often than maize (n = 11). He pointed out that the Hopi have considered grasses among the most important of wild foods, with Stipa hymenoides and Sporobolus being chief in importance (Whiting 1939:18). Doebley’s study also identified the rest of the common resources recognized

Rare Reproductive parts of a long list of plants were recovered in from 1 to 20 of the 64 locations studied, and many of these were present only in deposits of the Secondary occupation (Adams 2006a). Evidence for the use of vegetative parts of plants is weak, but could have included young goose74

Subsistence and Plant Use foot, pigweed and purslane greens, squash flowers, young beeweed plants, roots/rhizomes of riparian plants such as sedge or cattail, other roots, wild grass leaves and yucca leaves as quids, juniper bark, prickly pear pads, and maize cobs (Bohrer and Adams 2006b).

gether prevent an accurate estimate of the role of juniper berries in the diet.

Spurge The presence of spurge (Euphorbia glyptosperma) seeds in conjunction with other foods in Room 93W and in trash deposits (Adams 1980a, 2006a; Doebley 1976) suggests it was a food in prehistory. They were also recovered within a milling bin in Secondary occupation Kiva 6A, but not in sediments from around the bin. These findings are surprising given the generally poisonous nature of spurge plants, yet are supported by its association with other known foods, the contexts in which the seeds were recovered, and limited ethnographic references (Adams 1980a, 2006a, 2006b). Bohrer (2006) has pulled together additional ethnographic literature to speculate that toasted spurge seeds were formerly used as a flavoring or symbolic flavoring added to dishes made of cornmeal for use in both everyday life and ceremony. Parching, roasting, and heating are all known detoxification methods used in processing some members of the spurge family and genus ( Johns 1988: Table 1).

Two Wild Plant Foods of Special Interest Juniper

An intensive study focused on the more than 1,600 juniper seeds and cones (berries) found in stratified trash, pits (including burials), roof-fall, and in other locations (Lentz 1979, 2006). Morpho­metric comparisons with modern juniper species reveal that they are all seeds of Juniperus osteosperma, the local juniper. Seeds and cones were both charred and uncharred, some with insect borer holes; some were likely carried in by rodents in prehistory (Doebley 1976:98) or recently. However, the majority of them were considered cultural, and interpreted as food, based on: (a) their presence in 76 stratified trash strata in association with other foods; (b) their recovery in burned storerooms of the Secondary occupation; and (c) presence in collapsed roof strata where they either had been drying on the rooftops or were being processed there, a common location for milling by the Secondary occupants. Seven charred juniper seeds on the bench in the Tower Kiva (64W) were associated with maize, several other cultigens, and grinding stones. Juniper was also used as a funerary item, for adornment, and to satisfy a medicinal or ceremonial need. In an effort to understand the potential contribution of juniper cones to the Salmon diet, Lentz (1979) made estimates of: (a) the available juniper woodland within 6 miles; (b) the caloric value of juniper cones, minus indigestible component, which ranged between 5,200–5,500 kcal/kg; and (c) the caloric requirements of a population of 300 individuals, assuming on average daily need of ~2,430 kcal/day. He determined that the juniper woodland could reasonably be expected to provide 52 percent of the required kcal/year for 300 occupants of the Salmon community. The fact that juniper remains do not occur routinely in flotation samples from floors and middens may be due to the size and preservation issues discussed above, which may to-

Special Use of Cholla at Salmon

The case for cholla bud use begins with evidence of charred cholla flower buds from the floor of the Tower Kiva (64W) and cholla pollen in trash strata. Presence of cholla pollen on all Chacoan and Secondary floors of Room 93W suggest use through time (Adams 1980a, 2006b). Evidence for cholla management is primarily based on pollen grains, associated with Chacoan occupation rooms (31W and 93W), that do not conform to the range in size of pollen of modern cholla plants growing in the area. Possible explanations are trade for a variety of cholla that grew elsewhere, or local cultivation of cholla plants incapable of growing in the area without human attention.

Comparison of Chacoan vs. Secondary Subsistence Crops

The presence of maize in 77 percent of flotation samples from Chacoan trash (10 of 13 strata) compared with 57 percent of the flotation samples from 75

Karen R . Adams Secondary occupation trash (13 of 23 strata) suggests greater agricultural reliance on maize during the Chacoan period, although a larger sample would be needed to demonstrate statistical significance (Bohrer and Doebley 2006; Doebley and Bohrer 1983). This difference takes on added importance because it is clear the Secondary occupants gathered a wider variety of wild plants, as discussed below. In addition to recognizing maize as an important subsistence resource, maize ear variation was studied to examine social behavior of the Salmon inhabitants (Bohrer and Doebley 2006; Doebley and Bohrer 1983). Multivariate statistical analyses of a number of cob and cupule traits for 12 samples of 22 to 30 specimens each revealed a number of interesting differences: 1. Maize of the Chacoan occupation at Salmon averaged 11.9 rows per ear, similar to some Pueblo II Chacoan sites (Cutler and Meyer 1965; Judd 1954), and greater than the average row number of Secondary occupation maize at Salmon (10.77 rows/ear) and other Pueblo III sites on Mesa Verde (~10 rows per ear) (Cutler 1966; Cutler and Meyer 1965). Differences in the shape and general appearance of Chacoan occupation maize cupules (narrow, open cupules) versus Secondary occupation cupules (broad, collapsed cupules) also suggest differences in maize varieties, with the Secondary maize being more like mais de ocho, an eightrowed, wide-cupule maize (Galinat and Gunnerson 1963; Galinat 1970) introduced into the northern Southwest sometime after ad 900. 2. Maize recovered from non-ceremonial rooms of the Chacoan occupation is dissimilar from that of non-ceremonial rooms of the Secondary occupation. When compared, maize in two trash samples from each occupation showed considerable dissimilarity in features; for example, Chacoan maize had more rows of kernels and narrower cupules than maize of the Secondary occupants. 3. Within the Secondary occupation, maize morphology appears to have changed over time. A comparison of maize within roof strata of early Secondary times showed certain cob features

(rachis diameter, rachis segment length, and cupule depth, as well as the ratio of rachis diameter to rachis segment length) differed significantly from maize of later Secondary times. 4. Early Secondary period maize from nonceremonial­ contexts in the northwest and southeast sectors of the pueblo are similar, revealing no evidence to suggest that separate lineages maintained distinct maize varieties; 5. Maize from a Chacoan ceremonial room (81W) had more rows of kernels and were generally larger ears than everyday Chacoan maize. 6. Maize from a Secondary ceremonial room (64W) also had more rows of kernels, but were generally smaller ears than everyday Secondary maize. 7. If the maize recovered from Chacoan (81W) and Secondary (64W) ceremonial rooms are ritual maize types, the differences between them are large. 8. A study of kernel endosperm texture (flour, flint, pop) revealed the proportion of maize types for both occupations to be similar: 27.3 percent flour and 72.7 percent pop/flint for Chacoan, and 36.7 percent flour and 63.3 percent pop/flint for Secondary. This supported expectations that the predominance of trough metates (75 percent) used by Chacoan occupants were better suited to crushing kernels with hard, translucent starch, but contradicted expectations that the predominance of slab metates (75 percent) used by the Secondary occupants would have been better suited for grinding soft, floury kernels. Post-contact practices such as parching/popping maize kernels prior to grinding, or preparing maize dishes via boiling, suggest that other factors may be operating between the two Salmon occupations relating to maize kernel preparation.

Wild Foods

Studies comparing plant remains of the Chacoan and Secondary occupations indicate an increased reliance on wild foods, including drought-resistant­ juniper berries, by the Secondary occupants (Adams 2006a; Doebley 1976, 2006; Lentz 1979, 2006). In Doebley’s study of flotation and pollen 76

Subsistence and Plant Use samples from 41 trash strata chosen to represent both spatial and temporal variation, the Chacoan occupants utilized 19 wild plants, compared to 28 used by the Secondary occupants. After samples from the best (Room 62W) and poorest (plaza) preserved locations were removed from consideration, the mean number of wild plant taxa (4.25) in flotation samples from trash (n = 14) of the Chacoan occupation was also lower than that (6.53) in trash samples (n = 17) of the Secondary occupation (Doebley 2006). A two-sided Wilcoxon rank sum test produced a 95 percent probability that the results were produced by a mechanism other than chance. Although the same trend was seen in the mean number of cultural plant taxa per pollen sample in trash, small sample size weakened the pollen results. Non-cultural explanations for these trends, such as differential preservation conditions and differential soil pH, were reviewed and dismissed as unlikely to have affected these values (Doebley 1976). Doebley considered cultural explanations more likely to account for these differences (1976). He suggested that it was reasonable that two groups of individuals with different backgrounds and lifestyles would have different preferences for foods. Such an explanation is difficult to test with this data set alone, and a regional examination of archaeobotanical records would be necessary. Another possibility that Doebley considered was that a decrease in agricultural productivity led to an increased reliance on wild foods by the Secondary occupants. In this case, one might expect that plants growing farther from the community would increase in presence (e.g., piñon nuts, broad-leaf yucca fruits), but this did not occur. Nor was there a strong preference by Secondary occupants for perennial over annual resources, similar to Hopi preferences when crops failed (Doebley 1976:88). However, an unfavorable climatic period between ad 1270 and 1300, including a long drought, lowered water tables, and stream entrenchment most likely did affect Salmon’s agricultural productivity in a negative way (Van West and Dean 2000). These climatic difficulties, coupled with increasing population pressure in the late 1200s, might explain why the Secondary occupants used a greater diversity of wild plant foods.

Lentz also tested the hypothesis that the Secondary occupants relied more on wild foods than did the Chacoans (1979:83–92). He examined the ratio of domesticated squash seeds to wild juniper seeds for 8 Chacoan strata (16.12) and 11 Secondary strata (3.59). A chi-square test result indicated that the difference is significant, at the 99 percent level, supporting Doebley’s suggestion of a heavier reliance on wild foods later in time. Adams’s large study (2006a) of 216 flotation samples from trash and well preserved floors of both occupations revealed that 15 plant taxa were recovered only in Secondary occupation strata, many of them likely foods. Finally, although it appears the Chacoans relied less on piñon nuts (in only 9 rooms) than the Secondary occupants (in 22 rooms), the original study (Adams 2006c) made no estimate of relative stratum preservation, which may have skewed these results. Elsewhere in the region, a study of coprolites at Mesa Verde revealed an increase in the amount of maize and the number of wild plants consumed during the ad 1200s (Stiger 1977, 1979). At Antelope House, use of squash, piñon nuts, beeweed, and purslane decreased in the same period (Fry and Hall 1975). Squash and piñon nuts would both be particularly susceptible to drought. A separate study of coprolites at Salmon (Reinhard, this ­volume) focuses on parasites and their relation to health.

Dietary Components in Human Coprolites

Direct evidence of food consumption preserved within human coprolites from Rooms 62W and 81W confirmed a number of Chacoan and Secondary occupation food choices (Bohrer and Adams 2006b). The following list must be considered incomplete, as it did not include identification of epidermal tissues that might reveal consumption of young leafy greens or underground parts such as roots and tubers. Chacoans consumed hawthorn fruit, partly roasted juniper twigs, and buffalo berry seeds. They may also have consumed beans, beeweed seeds, squash seeds, juniper bark, prickly pear seeds, and lemonade berry seeds. Secondary occupants also ate squash, beeweed, and lemonade 77

Karen R . Adams berry seeds, along with goosefoot, purslane, and groundcherry seeds. Coprolites also provide information on the general health of Salmon inhabitants (Reinhard, this volume, 2006). Salmon coprolites lacked giardia (Giardia lamblia), likely because the nearby water sources included the flowing San Juan River and other smaller drainages (Reinhard, this volume; Wilson et al. 2006). Low prevalence of pinworm infection among the Secondary occupation at Salmon may stem from people living in open ­second-story rooms with adequate air currents, and using specific rooms as latrines (e.g., 62W), which could slow or stop the spread of parasites among the population.

winter or early spring, stored foods (domesticates and wild foods) were most likely depleted. By this time, some plants were available as edible greens (Chenopodium, Descurainia), providing important nutrients but unlikely to be preserved in the archaeological record. Descurainia, Stipa hymenoides (as Oryzopsis hymenoides), and Mentzelia albicaulis would then be among the first to produce fruit or seeds. Fruit of Echinocereus fendleri would ripen in late spring, onion bulbs (Allium macropetalum) could be harvested, and cholla (Opuntia) flower buds would be ready to gather. By midsummer, following the start of the monsoons, young leaves and stems of Chenopodium, Amaranthus, Atriplex, and Cleome would have been in abundance, along with ripe fruit of Rhus aromatica (as Rhus trilobata). By August, the fruit of Physalis would ripen. By late August and early September, the fruit of sedges (Carex and Scirpus) and broad-leaf yucca (Yucca baccata) could be harvested, along with seeds of Cleome, Polanisia, Trianthema, Salvia, and Sporobolus. Fall ripening seeds of Chenopodium, Amaranthus, Cycloloma, and Portulaca offered additional resources. Once the fields of maize and other crops were harvested, groups from Salmon likely traveled to collect piñon nuts and juniper berries. Most plant collecting activities were complete by mid-November. When food reserves were insufficient through the winter, collection of cactus pads and joints, juniper berries still clinging to trees, tree bark, yucca leaves, and underground roots may have been necessary, along with consumption of maize cobs.

Evidence of Food Shortages

In addition to the foods mentioned above, the Secondary occupation coprolites also contained foods that can be considered starvation resources (Bohrer and Adams 2006b). These included juniper bark, broad-leaf yucca leaves, maize cob fragments, bones, and insects. Historic Tewa travelers baked and ate yucca leaves when other foods were not available (Robbins et al. 1916:50). Yucca leaves and juniper bark are generally recognized as famine foods in the ancient Southwest (Castetter 1935:32). Their real advantage is their availability throughout the year. Leftover maize cobs are also famine foods (Hill 1938:45–46). Limited evidence in the form of immature piñon (Pinus edulis) nuts and cone scales in strata of mixed cultural affiliation hints that Salmon folks occasionally brought immature cones to the pueblo and roasted them to open the cones to release their unripe nuts (Adams 2006c). This resource might have been important in the early fall if agricultural crops were meager and people were hungry.

Plants Used for Reasons Other than Subsistence Adornment

Some juniper seeds recovered from Salmon were pierced at both ends for use as beads. Six seeds with holes through both ends were found together in the roof-fall of the Tower Kiva, possibly suggesting a seed bead necklace (Lentz 1979, 2006).

Seasonality of Subsistence Resources The archaeobotanical record suggests that the Salmon inhabitants followed a seasonal round of plant gathering. However, because foods can be dried, parched, or prepared for long-term storage, the season(s) of gathering cannot be equated with season(s) of use (Adams and Bohrer 1998). By late

Construction Beams, Roofs, Ceilings, and Other Constructions

Salmon roof timbers identified by the Laboratory of Tree-Ring Research were originally discussed by Adams (2006c). However, additional analysis has 78

Subsistence and Plant Use expanded the number of construction beams dated and the types of trees brought to the pueblo for building (Windes and Bacha, this volume, 2006), providing the most complete view of the construction history and related topics (see also Baker, this volume). It seems clear that organized workforces traveled some distance to obtain construction beams, similar to modern Isleta efforts ( Jones 1931:​ 37), especially to stockpile beams for major building episodes. Information on building customs includes the sequence of layers used in roof construction, and possible rituals related to floor preparation. A well preserved first-story Chacoan roof segment (60A) revealed how it was constructed (Adams 2006c). Starting from the bottom layer: (a) five large vigas (up to 20 cm diameter) were oriented north-south and secured into wall sockets; (b) the first layer was topped by smaller 5–8 cm latillas spaced 5–8 cm apart in an east-west orientation; (c) the second layer was covered with closely spaced willow twigs (< 2.5 cm diameter) in a north-south alignment; (d) followed by a thick layer of adobe; (e) upon which shredded juniper bark was layered; and (f ) another thick layer of adobe packed on top completed the roof and served as a floor for the ­second-story room. Cottonwood (Populus) and oak (Quercus) probably provided some of the smaller latilla and tertiary layers. One carbonized juniper cone (berry) was found in a layer of artificial sand laid over the Chacoan floor by the Secondary occupants of the Tower Kiva (Lentz 1979, 2006). Juniper berries and seeds were found beneath clay floors in some of the rooms at Mug House (Rohn 1971:​257) on Mesa Verde. Similarly, the Hopi would spread meal mixed with piki crumbs over the subfloor of a kiva as part of groundbreaking ceremonies prior to laying the clay floor (Mindeleff 1891:​118); juniper berries were also used by the Hopi in the preparation of piki bread (Whiting 1939:63).

Fuel

Although charred wood was not a major focus of archaeobotanical analysis, some evidence of fuel use is available. Leftover maize cobs were used regularly as fuel or tinder in hearths, as indicated by 79

the presence of cob parts and cupules in thermal features of Room 93W (Adams 1980a, 2006b) and in trash deposits (Doebley 1976, 2006; Bohrer and Doebley 2006; Doebley and Bohrer 1983). Some of the juniper twigs and scale leaves recovered within hearths and trash suggest regular use of juniper as a fuel, although some of these parts could also derive from medicinal preparations of the twigs. A pilot study of hearth charcoal revealed that conifers constituted a common firewood choice during both occupations, with charred juniper wood most common (Rossi 1977). This finding reflects the location of the Salmon community within a predominantly juniper woodland (Lentz 1979). Limited evidence of charred pine (Pinus) wood and nonconifer wood was also documented within hearths. Bark left over from debarking ponderosa pine trees for construction timbers may have also been used as a fuel or tinder source (Adams 2006c).

Household Uses

Conifers in the region provided for the household needs of both occupations (Adams 2006c). Long, loose strips of pliable juniper bark that shredded easily and bent without breaking were widely used for daily household needs. Masses of shredded but otherwise unmodified juniper bark were recovered from half the excavated rooms representing both occupations. Chacoan occupants fashioned juniper bark cordage. Secondary occupants also used juniper bark as cordage, to make pot rests and burden rings, and as toilet paper (Room 62W); it was also stored as raw material (coiled bark, twisted bark) for other uses. They also gathered a juniper twig bundle. A juniper bark needle was found, consisting of a long strip of bark with the layers tightly adhering at one end and shredded the rest of the way. This would have served as a sewing tool with the needle built in, similar to the sharp, narrow, spine-tipped ends of broad-leaf yucca leaves. The presence of spruce bark scales in both Chacoan and Secondary trash suggest occasional use of spruce wood. The use of conifer materials in household needs is rivaled only by yucca (Yucca) leaves, discussed in greater detail by Webster (2006, this volume). Yucca leaf fibers were routinely made into cordage,

Karen R . Adams knots, plaited mats, coiled and plaited baskets, and plaited and twined sandals, all discussed fully and compared to other great houses by Webster (2006. this volume).

red-stained perishable artifacts described by Webster (2006) was identified as iron or hematite, and most are associated with the Chacoan occupation (Webster, this volume). Webster reported that redstained objects at Aztec Ruin and Pueblo Bonito were found in rooms containing quantities of ritual materials, suggesting that the stained Salmon objects may also represent ritual use. Another possible ritual item was a stick that had been shaped as the letter C, similar to sticks recovered from Pueblo Bonito ( Judd 1954:269). Many references to the use of hooked or bent sticks in ceremonial events have been assembled by Judd (1954:​ 269). Also, several examples of reedgrass (Phragmites) cigarettes, one still containing dottle in the burned portion of the cane, strongly suggests use in a ceremonial context, as documented in the historic literature (Adams 1990). One example of offering food to a departed spirit may include a small burned canteen jar in Secondary occupation trash (100W) that was filled with burned squash and winged pigweed (Cycloloma) seeds as well as maize pollen. Such jars are not usually used in cooking. Puebloans believe that upon death the spirit only needs the odor or smoke of prepared food, and Zunis say that ghosts eat food put into a fire and burned (Parsons 1939:302). Although this canteen jar was not associated with a burial, there were several burials nearby. Puebloans are not known to consume either ­yerba mansa or Arizona poppy. Yerba mansa leaves and roots have been used to treat various ailments (Swank 1932:​26–27; Kearney and Peebles 1960:​ 207). Arizona poppy is also a valued medicinal plant (Curtin 1965:68). The low frequency of recovery of both of these items in hundreds of flotation samples (< 2 percent) is reasonable for a plant sought occasionally as a medicine. Widespread presence of juniper twigs or scale leaves in 40 percent (87 of 210) of broadly distributed flotation samples suggests widespread use of juniper boughs in prehistory (Adams 2006c). Clearly, fuel use of juniper wood with twigs still attached would contribute to this record. However, documented post-contact Native American family medical practices have included use of juniper in tea form for earaches, as binding for bruises and

Medicinal and Ceremonial Plant Use

Determining medicinal and ceremonial uses of plants can be difficult when the same plants are also used in ordinary ways. However, some examples from Salmon fit into this category, most of them associated with maize. An association of maize parts with an unusually high occurrence of maize pollen (37 percent) on one of the Secondary floors in Room 93W suggests that prayers or rituals were conducted in this room (Adams 1980a, 2006b). Cattail pollen on this same floor may also represent ceremonial use, as this pollen is easily gathered like maize pollen for ritual use. On another Secondary floor in Room 93W, it appears that an offering of groundcherry (Physalis) fruit was made to the fire, a custom found among modern Puebloans (Adams 1980a, 2006b). The presence of maize pollen aggregates on seven Chacoan floors and eight Secondary floors implies the gathering of immature maize tassels, similar to the practices of modern Puebloans who gather maize pollen for only medicinal or ceremonial use (Bohrer 2006). Two Chacoan rooms (30W and 62W) considered to have had ceremonial usage due to the presence of maize aggregates also had maize pollen percentages above 80 percent. In another Chacoan room (31W), the association of maize pollen (83 percent) with chopping tools and bones suggests blessing rituals associated with the butchering of meat. This is similar to a modern Zia custom that involves a war chief visiting the homes of successful hunters to sprinkle maize pollen over game later butchered in a ceremonial house (Stevenson 1894:120). Maize pollen aggregates in 12 of 23 Secondary occupation trash units may represent the floor sweepings of ceremonial locations, including maize tassel fragments (Bohrer 2006). Other evidence suggests medicinal or ceremonial plant use (Bohrer 2006). A number of objects from Room 62W (broken pieces of peeled sticks, yucca leaves, young prickly pear pads) all carried traces of red or pink stain. The pigment on nine 80

Subsistence and Plant Use sprains, and for other ailments (see Adams 2006c). A neatly wrapped bundle of juniper twigs from the roof of a Secondary occupation room (129W) implies they were in storage for later use. Other conifers were sought on occasion. Limited recoveries of spruce bark scales within trash deposits of both the Chacoan and Secondary occupations may relate to uses in nearby rooms serving ceremonial functions (Adams 2006c). Although spruce trees were not routinely used as construction beams, a spruce bark scale associated with the Great Kiva (130W) suggests that the planking of the foot drum was fashioned of spruce wood. Spruce twigs and branches have served ritual needs among modern Puebloans (White 1942:287; Vestal 1952:​ 12). Two lots of ponderosa pine (Pinus ponderosa) needles of mixed cultural affiliation may once have been used for ritual (Robbins et al. 1916:41; Whiting 1939:​63) or as a medicine (Vestal 1952:13–14). Although technically not a conifer, limited numbers of stems of the gymnosperm Mormon tea, often used as a medicinal treatment in the postcontact era, were preserved in Salmon contexts of mixed cultural affiliation (Adams 2006c). Some items associated with burials may have been offerings or part of the burial wrappings. For example, juniper seeds were recovered in eight strata associated with burials, all enveloped by matting (Lentz 1979, 2006). The single Chacoan burial had many other grave goods as well (brush of monocotyledon leaves, prayer sticks, a bow, Phragmites arrow shaft). The association of the seeds with the burials is reliable for some burials, but not for others. In the post-contact era, graves included food placed with the deceased for the long journey after death. Limited ponderosa pine bark in association with a burial pit (Room 102B) of Secondary occupation may have been used to line the burial pit (Adams 2006c).

Storage Rooms/Rooftops in the Secondary Occupation Seven burned Secondary occupation roofs, sandwiched between unburned strata, were considered prime locations to study burned associated plant remains (Bohrer and Adams 2006b). These were characterized by the presence of maize cobs and 81

kernels, beans, squash, a diversity of wild foods, raw plant materials (wood, bark, stems), finished items (leaf knots, plaited mat fragments, cordage, textile fragments), lithics, and ceramic sherds. Altogether these assemblages suggested these rooms functioned as storage rooms. The most consistently recovered plant items were cultivated maize and beans. Wild foods in storage included chenoam seeds, along with the fruits of lemonade berry, juniper fruit/seeds, piñon nuts, prickly pear pads, and cholla flower buds. Unidentified tubers indicate use of underground plant parts, which rarely preserve in the archaeobotanical record. Raw materials in storage included coils of smooth bark, grass stems, and yucca leaves. Basketry fragments in association with fused maize kernels suggest a storage container. Pieces of textile, matting, netting, twine, and braided fiber were either parts of storage containers or items in storage. The presence of rodent feces suggests rodent problems with stored ­materials. A storeroom on the ground floor (30W) contained maize cobs stacked along the south wall of a floor, next to the wall. Beans were also in storage there when a roof burned and fell onto the stored crops, sometime around ad 1210 (­archaeomagnetic date). Two other ground-floor rooftops of the Secondary occupation (100W and 127W) served as outdoor processing areas as well as storage. The plant inventory of these two outdoor locations is quite similar to that of the interior storage rooms discussed above (e.g., maize, beans, squash, cholla flower buds, juniper berries). Perhaps cholla flower buds were in storage in a vessel, as they are the only food product in this assemblage available in the springtime, and all other foods identified there ripen in the fall.

Plant Remains Associated with Kivas Interest is high in the types of activities that occurred in kivas or other recognized ceremonial rooms. Plant evidence can suggest ceremonial usage of plants or everyday uses, such as for food preparation. The evidence for ceremonial use of maize pollen in blessings is considered first (Bohrer 2006). Two Chacoan non-kiva ceremonial rooms (30W and 62W) have both maize aggregates and maize

Karen R . Adams

Diversity of Habitats Visited

pollen percentages above 80 percent. Two Secondary kiva floors (6A, 124A) do not have broadly distributed maize pollen aggregates, although aggregates in a slab-lined bin in Kiva 6A may be residue of grinding maize tassels on a metate within the bin. Some ceremonial activities with maize were also associated with non-kiva locations, including Chacoan rooms with features (93W, 101W) and without features (30W, 56W, 62W, 82W, 89W), and Secondary rooms with features (93W) and without features (51W, 57W, 58W). Some of these rooms (51W, 57W, 58W, 62W) are close to the Tower Kiva (64W), but any pollen there would have been destroyed by the extremely intense fire. A variety of seeds of economic importance also were preserved on three Secondary occupation kiva floors (Rooms 6A, 64W, and 124W), suggesting that secular aspects of life occurred in these locations (Bohrer and Adams 2006b). Whether this food was consumed during special ceremonies cannot be determined. Within Kiva 6A, seeds were concentrated in and near a milling bin whose grinding stone had been removed. There, the concentration of cheno-am and tansy mustard seeds is 100 times greater than elsewhere on the kiva floor, and heavy concentrations of purslane, stickleaf, and spurge seeds are unique to the bin area. These remains accompany a high concentration (47 percent) and presence of maize pollen aggregates within the bin, perhaps where maize tassels were also ground to remove the pollen. Most of the seeds types from this milling bin area were also recovered from beneath the metate on a habitation floor of Room 93W (Adams 1980a, 2006b), lending support to the notion that the former grinding stone was also used to grind foods. The two other Secondary kivas (64W and 124W) preserved a broad spectrum of small seeds that are regarded as foods. When the fire precipitated the swift depopulation of the Tower Kiva (64W), many food plants were burned on the floor. Among these were maize ears, beans in pods, squash rind, tansy mustard seeds, stickleaf seeds, cholla flower buds, young prickly pear pads, beeweed seeds, purslane seeds, piñon nuts, and juniper seeds. These clearly reflect more than one season of gathering.

The plant record allows construction of a list of ­local habitats that provided resources in prehistory. Riparian locations were visited for resources recovered in trash (Adams 2006a; Doebley 1976, 2006) and on floors of Room 93W (Adams 1980a, 2006b). These include sedges (Carex, Cyperaceae, Eleocharis montanus), rushes ( Juncus spp.), reedgrass (Phragmites), chokecherries (Prunus), bulrushes (Scirpus, S. acutus), and cattails (Typha). Nine centuries ago, stream banks along the nearby San Juan River provided ideal habitats for these plants, just as they do today. Because of a regular water supply, many riparian plants are able to offer a very dependable yearly food and material culture source. Disturbed locations such as fallow agricultural fields, trash heaps, and pathway edges provided places for weedy resources to thrive (Amaranthus, Chenopodium, Cleome, Cycloloma atriplicifolium, Descurainia, Helianthus, Mentzelia albicaulis, Polanisia, Portulaca, Salvia reflexa, Trianthema portulacastrum). Many of the rest of the wild plants, but not all, are known to grow in juniper woodlands.

Distant Travel or Trade for Resources It seems likely that travel groups/expeditions explain much of the conifer evidence at Salmon Ruin (Adams 2006c; Windes and Bacha, this volume). Conifer species exhibiting a relatively large number of parts and widespread distribution within the pueblo likely grew closer than those conifers exhibiting a smaller number of parts and a restricted distribution within the pueblo, assuming that the plant species in the region today are not significantly different than those of the eleventh through thirteenth centuries ad. With this in mind, juniper was a readily available conifer in prehistory, although some exploration of the woodland would have been needed to find all the appropriately sized juniper beams harvested by the Chacoans for construction. However, the extensive record of shredded juniper bark, berries (cones) and seeds, twigs/ scale leaves, and charred wood together suggest ease of acquisition of juniper. On the other hand, the restricted and limited nature of piñon parts, with the exception of nutshells, suggests that people trav82

Subsistence and Plant Use eled some distance to harvest the nuts. Presence of ponderosa pine construction beams, coupled with limited recovery of any other parts (bark and needles), strongly implies that the Chacoans traveled to higher elevations for these construction elements, especially from ad 1088 to 1090. Likewise, the only evidence of the use of Douglas fir, white fir, and spruce comes in the form of very limited construction timbers or bark scales, implying longdistance travel or possibly the occasional use of driftwood logs brought by the San Juan River. At the present time, travel of some distance is required to find large stands of conifers other than piñon and juniper (Windes and Bacha, this volume). A number of plant resources were likely obtained through trade. These include limited cotton textiles, as there is no direct evidence (seeds, bolls) that cotton was grown locally. It also includes a gourd container fragment that likewise has no direct evidence (seeds, rind fragments) of being grown nearby. Screwbean trees grow below 4,000 feet and have not been located anywhere in the Salmon region, so the presence of screwbean seed coats in numerous flotation samples suggests they were brought to Salmon via trade. Possibly people traded for a variety of cholla cactus that has larger pollen grains than those of the local varieties of cholla. The question of what resources the Salmon occupants could potentially offer in trade has been recently addressed by researchers trying to locate where maize was grown to support large ­labor forces within Chaco Canyon in New Mexico (Benson et al. 2003). A study of strontium (Sr) isotope ratios and elemental analyses of Chaco Canyon maize suggests that some of it was likely grown along the San Juan or Animas River floodplains over 90 km to the north. So, it is reasonable that the Chacoan occupants of the Salmon great house were trading some of their maize to Chaco Canyon for other items such as cotton textiles and gourd containers. Possibly screwbeans and an unusual variety of cholla cactus came via a Chaco Canyon trade network or via other trade relationships. The Chacoan occupants of Salmon could also have easily gathered quantities of piñon nuts and juniper berries to trade south, as both piñon and juniper

trees were decreasing during Prehispanic occupation of Chaco Canyon (Betancourt and Van Devender 1981; Samuels and Betancourt 1982). It is unlikely the Salmon great house provided construction timbers to Chaco Canyon because: (a) Salmon folks had to travel long distances to acquire their own timbers; and (b) recent chemical analyses of Prehispanic roof timbers have suggested the Chuska and San Mateo mountains northwest and southeast of Chaco Canyon, respectively, provided Chaco Canyon with some building supplies (English et al. 2001).

A Possible Reason for Final Pueblo Depopulation A study of conifer evidence at Salmon Ruin was utilized to speculate on its final depopulation (Adams 2006c). The distribution of conifers in prehistory appears similar to that of today, suggesting similarities in general climate. However, this obscures the relevance of sudden and relatively shortterm climatic fluctuations that would affect both subsistence and social systems (Van West and Dean 2000). Around ad 1275, a prolonged drought ­began that was accompanied by falling water tables, floodplain erosion, and a period of unpredictable and irregular seasonal rainfall. For the Secondary occupants, these environmental changes apparently resulted in lowered agricultural productivity of both maize and squash, and an increased reliance on a range of wild plants. However, when conditions are difficult for agricultural crops, they are often difficult for wild plants as well. During arid growing seasons, annual plants may not make an appearance, and perennial plants may put more effort into staying alive than in reproducing. Clearly some Secondary occupants were also forced to resort to famine foods such as juniper bark, yucca leaves, and corn cobs. Yet even increasing reliance on wild plants and famine foods may have been less than adequate to sustain people until the drought period abated around ad 1300. The result was a complete depopulation of the Salmon great house sometime during this extremely difficult environmental period. Other factors related to social and political circumstances certainly contributed to the depopulation, but plant evidence must be used 83

Karen R . Adams in conjunction with many other lines of evidence from Salmon and from the regional archaeological record to best evaluate such influences.

was agricultural production of maize, beans, and squash. Maize of the two occupations differed in notable ways. Gourd containers may have been traded in, along with limited amounts of cotton ­fibers and cloth. The Chacoans may have managed a type of cholla plant or traded for its fruit. Wild plant gathering was a regular activity, particularly of weedy plants that grew in their fields and other disturbed locations. Grass grains, cactus fruit, ­piñon nuts, and juniper berries were also among the gathered resources, along with many others. The long list of foods spans nearly all the seasons of the year, except winter, when plant products such as juniper berries, cactus pads, and tree bark could still be found. A comparison of foods between the Chacoan and Secondary occupations reveals some differences. Chacoans utilized maize more often and sought fewer species of wild plants than the Secondary folks. Study of coprolites reveals that the Secondary residents occasionally ate corn cobs, juniper bark, and yucca leaves; none of these were particularly nutritious, but they were filling. A shortage of maize and an increasing reliance on wild plants and less-preferred foods may have been one result of a late-thirteenth century environment that became less predictable and dependable for maize farmers. In addition to subsistence, plants served many other needs: juniper seeds were used for adornment; a range of conifer and some non-conifer trees were sought as roofing elements; piñon and juniper fuel was burned regularly for cooking, heating, and light; juniper bark served many household needs; and yucca leaves were fashioned into a wide range of baskets, mats, sandals, and other items. Maize pollen was used in ritual contexts, as was cattail pollen. Foods placed in jars may have been offerings to departed ancestors or spirits. Items (peeled sticks, yucca leaves, prickly pear pads, and perishable artifacts) with remnants of pink or red staining may also relate to ritual needs. Juniper twigs in many contexts were perhaps routinely used for medicine. The storage habits of Secondary occupants are known from a number of burned rooms. Maize and beans were present most often, along with weeds and many seeds and fruits of perennial plants. Coils of bark, grass stems, and yucca leaves in storage suggest the materials were intended for use in making

Plants Whose Distribution or Abundance has Changed Since Prehispanic Times Some of the plants considered foods for the Salmon occupants are not common in the area at present (Bohrer and Adams 2006b). For example, winged pigweed seeds (Cycloloma atriplicifolium), stickleaf seeds (Mentzelia albicaulis), and horse purslane seeds (Trianthema portulacastrum) all appear to have been gathered at least occasionally as foods, yet these annual plants were by no means easy to find during six years of regular plant collecting. Locating perennial chokecherry (Prunus) trees, yerba mansa (Anemopsis californica) plants, and wild onions (Allium macropetalum) in the region was also quite difficult. Other perennials whose availability in Prehispanic times may have been higher include hackberry (Celtis), hawthorne (Crataegus), and buffalo berry (Shepherdia argentea). Some of these plants may have been depleted due to introduction of domestic grazing animals, and others may have had their niches usurped by foreign taxa. There is no insight into why juicy-fruited prickly pear plants are scarce in the region at present, unless they have also been preferred by livestock. Of all the places where plants prefer to live, the ones most impacted in the post-contact era may be those riparian resources (Prunus, Crataegus, Shepherdia, Celtis, Typha, sedges, and reedgrass) that have given way to thick stands of introduced Russian olive.

Summary and Conclusions Drawn from many studies and reports, the plant record of the Chacoan and Secondary occupants of Salmon Ruin is rich in detail and insight. Despite the fact that the relative recovery of Secondary materials was significantly higher than that of the Chacoan period, multiple seasons of systematic archaeological sampling of macrofossils, flotation, and pollen samples have provided a comparative basis for viewing subsistence and other uses of plants through time. The basis of subsistence for both occupations 84

Subsistence and Plant Use textiles, matting, twine, and braided fibers. In one case, maize was stacked along the south wall of a Secondary room floor. Kivas contained evidence of both ceremonial activities and everyday life. Maize pollen aggregates can only be explained as the intentional collection of maize pollen. Charred seeds considered to represent economic resources on three Secondary occupation kiva floors point to secular use of the structures. When the Tower Kiva (64W) burned at the end of the Secondary occupation, plants in storage reflected more than one season of gathering and represented many of the foods found in nonceremonial contexts and in middens. Most of the plant materials used during both occupations of Salmon Pueblo was available locally or within a day’s journey. Some, such as the large conifer roof elements, clearly required a longer trip of greater effort. The presence of multiple parts of juniper trees strongly suggests that a juniper woodland surrounded the pueblo. Piñon trees likely were not as abundant or were simply not used as often. Plants obtained through trade included cotton, gourds, screwbeans, and a rare form of cholla cactus not currently found in the local area. Salmon farmers may have traded maize for these non-local resources. The question of final Puebloan depopulation is complex. Inter- or intra-Puebloan stresses were likely responsible for the fire in the Tower Kiva (and many other rooms) that caused numerous deaths. The plant record suggests that final Pueblo depopulation was related in part to difficulties in acquiring maize, either because the environment had become unpredictable or because the productivity of fields had declined. One result seems to have been increased use of wild plants, some of them of low nutritional value; however, the possibility that these subsistence differences related to social or political influences, or simply to differences in food preferences between the Chacoan

and Secondary occupants, cannot be ruled out. Further exploration of such explanations would require use of multiple archaeological data sets in addition to plant remains and a regional comparative framework of subsistence and plant use information from contemporary great house and post– great house communities. The archaeobotanical record has contributed a number of insights to the original model of Chacoan society developed by Irwin-Williams. These insights rely on comparisons of the Chacoan plant record to that of the Secondary group that came later in time. The Chacoan occupants appear to have depended more on agricultural products, in part because environmental conditions for agriculture were good until the climate began deteriorating around ad 1130. Chacoan maize was more uniform than that of the Secondary folks, and maize in Chacoan ceremonial contexts was unlike the maize of residential units, although the same could be said of some Secondary ceremonial contexts. These findings generally supports Irwin-Williams’s hypothesis of a religious elite that also maintained economic control. The Chacoans had trade relations with other groups within the region, or traveled farther for certain plant goods and resources, including their construction timbers, revealing the larger regional system within which they operated. Speculation that Salmon provided Chaco Canyon with maize (Shelley 2006) is reasonable based on chemical studies that suggest some maize in Chaco Canyon was likely grown along the San Juan or Animas River floodplains, perhaps in return for cotton textiles, gourd containers, and other plants that entered the Chacoan society trade network. In return, the Chacoan occupants of Salmon may have supplied juniper berries and possibly piñon nuts to the south. The ancient plant record plays a key role in understanding the structure and control of Chacoan society and economy, and is central to greater knowledge of the Chacoan world.

85

5

Parasite Pathoecology of Salmon Pueblo and Other Chacoan Great Houses The Healthiest and Wormiest Ancestral Puebloans Karl J. Reinhard Archaeoparasitology and Pathoecology in the Southwest

affected parasitism at Antelope House and Salmon Ruin is another application of pathoecology. Ancestral Pueblo communities have long been the focus of archaeoparasitology. Samuels (1965) developed the methods for helminth (parasitic worm) egg recovery with coprolites from Mesa Verde. Subsequently, Stiger (1977) provided the first intersite comparison analysis for sites on Mesa Verde. Fry and his colleagues conducted the first regional comparisons of parasitism, focusing on Canyon de Chelly and Glen Canyon (Fry 1977; Fry and Hall 1975, 1986). Fry (1977) also presented the first cross-cultural analysis of Archaic, ancestral Pueblo, and Fremont sites, and pioneered the comparison of parasitism among populations practicing different subsistence strategies. Building on this previous work, I have analyzed the diversity of helminths that parasitized ancestral Pueblo peoples (Reinhard 1985a, 1985b, 1985c, 1990; Reinhard et  al. 1987). By 1985, archaeoparsitologists had identified eight species of helminth that infected ancestral Puebloans (Figure 5.1). Aidan Cockburn’s insight into the origins of ­disease influenced the development of pathoecology in the archaeoparasitology of ancestral Pueblo sites. Cockburn (1967, 1971) argued that the evolu­ tion of infectious diseases followed human evolution and the development of human cultures. Inspired by Cockburn, Reinhard (1985a) compared the parasitic state of Colorado Plateau Archaic peoples to ancestral Puebloan sites. He verified Cockburn’s hypothesis that occasional infections

Two fields of paleopathological investigation originated in the Southwest. Archaeoparasitology is the study of ancient parasite infection (Reinhard 1990, 1992b). It includes comparisons between time ­periods of single societies as well as comparisons of parasitism between different, contemporaneous cultures. For example, Fry (1980) compared Fremont and Anasazi parasitism, and also Archaic hunter-gatherer and ancestral Pueblo parasitism. All of these studies fall into the definition of archaeoparasitology. By contrast, pathoecology is the reconstruction of relationships among behavior, environment, and disease organisms in the development of illness (Martinson et al. 2003; Reinhard and Buikstra 2003; Reinhard et al. 2003; Santoro et al. 2003). This field developed from the need for fine-grained analysis of prehistoric ecological and behavioral conditions to assess factors that affected disease. Pathoecological interpretation depends on archaeological information regarding parasitism, community size, trade patterns, water sources, subsistence practices, environment, medicinal use, and many other topics. Although the term is new, pathoecology developed over several decades. I view El-Naijar­et al.’s (1976) study of ancestral Pueblo anemia as the first pathoecology study. Perhaps the most advanced example of pathoecology is Stodder and Martin’s (1992) multifactorial perspective on ancestral Pueblo disease. My study (Rein­hard 1996) of the factors that 86

Parasite Pathoecology

Figure 5.1.  Diagram showing the wide spectrum of parasites that infected ancestral Puebloans.

ver­micularis) prevalence in coprolites (Reinhard 1988). The pinworm was chosen as an indicator of general infectious disease because it is transferred by person-to-person and by environmental contamination (Figure 5.3). Over millions of years of mutual evolution with hominids and modern humans, pinworms have evolved multiple routes of infection, including anal-oral, hand-to-hand, and airborne routes. Pinworms are exceptionally remarkable among ­human parasites because the female worm wriggles out of the anus of her host at night to scatter her eggs. Once outside of the intestine, she disperses eggs by two different mechanisms. Two types of eggs are produced in two parts of the pinworm uterus: light and heavy. Heavy eggs are laid on the perianal folds with an irritant excretion. The resulting itching (pruritis) and nocturnal host scratching transfers the infective eggs to the host fingers. Other eggs are distributed by aerosol when the female’s desiccated body bursts, which releases thousands of light eggs into the air. Ultimately, these light eggs contaminate the environment, settling on food, in water, and throughout the

in hunter-gatherers became major health hazards in agricultural populations. Reinhard (1988) presented a number of pathoecological findings regarding the development of parasitic disease in ancestral Puebloans relative to earlier hunter-gatherers­. Parasitism was limited in hunter-gatherers due to small band size, band mobility, diffuse regional populations, and the presence of natural anti-helminthics (worm poisons) in hunter-gatherer diets. Huntergatherer parasitism was promoted by the consumption of uncooked meat and insects. Parasitism was promoted in ancestral Puebloan communities by contaminated water sources, concentrated populations, a more sedentary lifestyle, crowded (­apartment-style) living conditions, establishment of large latrines, activities centered on water (agriculture), and activities that expanded wetlands (including irrigation of all types). By the 1990s, Reinhard (1992a) had identified wide variation in parasitism among ancestral Pueblo villages (Figure 5.2). At some settlements, parasitism was controlled, but others were overwhelmed by their pathogens. This topic was explored with a comparison of pinworm (Enterobius 87

Karl J. Reinhard

Figure 5.2.  Map showing variation in percentages of pinworm parasitism among ancestral Pueblo villages.

Figure 5.3.  Diagram showing modes of pinworm transmission to human hosts.

­ abitation. How long these eggs ­remain infective h depends on warmth and humidity. In general, even in arid environments, ­human habitations have an elevated humidity. Thus, several ­infection routes result from the pinworms’ nocturnal excursions. Retroinfection occurs when the eggs hatch on the

perinanal region, and the larvae wriggle back into the host. Hand-to-hand transfer of the eggs occurs when ­humans interact upon waking. Autoinfection occurs when humans eat food contaminated with the eggs from their own hands. Airborne infection occurs when humans inhale the eggs, or 88

Parasite Pathoecology

Figure 5.4.  Graph comparing pinworm parasitism with porotic hyperostosis prevalence for several southwestern locales. The chart shows that the prevalence of pinworm parasitism covaried with porotic hyperostosis prevalence at ancestral Pueblo sites where both coprolites and skeletons were studied (Reinhard 1992a).

when the air ­dissemination of eggs ­results in the contamination of food and water. Of course, other pathogens follow the same hand-to-hand, handto-mouth, and aerosol routes as pinworm infection. Therefore, high rates of pinworm prevalence suggest high rates of infection by other pathogens that are passed through the same modes of infection (see Figure 5.1). Some ancestral Pueblo communities were extremely parasitized. In fact, some sites have the highest levels of pinworm infection recorded for ancient or modern peoples. In a modern clinical setting, only 5 percent of feces from pinworminfected­people are positive for pinworm eggs. The percentages of coprolites positive for pinworm from several sites far exceed this level. For example, 29 percent of the coprolites from Antelope House, 19 percent of those from Inscription House, and 21 percent of those from Chaco Canyon sites were positive for pinworm eggs. This indicates that pinworm parasitism was unavoidable, and that in all probability people had heavy infections. In such populations, pinworm infection is not just a nui-

sance, but reflects a serious health risk, particularly when one considers that other pathogens are spread by the same means. Reinhard (1992a) showed that the prevalence of pinworm parasitism covaried with porotic hyperostosis prevalence at ancestral Pueblo sites where both coprolite and skeletons were studied (Figure 5.4). Porotic hyperostosis is an indicator of general skeletal pathology that has been used to assess maternal-infant health. The fact that these indicators of disease had a positive, statistically significant correlation underscores the use of pinworms as a general gauge of ancestral Pueblo disease state (Reinhard 1992a). Pinworms are not very pathogenic but are a good proxy for the infectious disease environment (Reinhard 1996). Rates of pinworm infection at ancestral Pueblo and Fremont culture sites were explored by Hugot and his colleagues (1999). They found that sites in rockshelters without walled villages (some Glen Canyon sites) had the lowest levels of parasitism. Such sites had pinworm prevalence comparable to hunter-gatherers. Next, village 89

Karl J. Reinhard sites outside of rockshelters had intermediate levels of parasitism. Finally, walled villages built within rockshelters had the highest prevalence of pinworm (see Figure 5.4). If we look at data from the Chacoan realm, we find that Chacoan great houses are anomalous: they include both the wormiest and healthiest sites. Salmon Pueblo has among the lowest prevalence values (7 percent). In contrast, Pueblo Bonito and Pueblo Alto in Chaco Canyon are among the highest (21 percent). Clearly, the pathoecology of great houses was defined by factors other than size. Puebloans living in great houses adapted their use of the structures in ways that either promoted or limited parasitism. The remainder of this chapter explores the factors that could have limited parasitism at Salmon relative to other great house communities.

persal capability and the ability to find hosts when needed. Also, they must possess attributes enabling them to survive in the external environment. Features of the host have less effect on survival and reproduction of these parasites. Nidicolous parasites live in the host’s immediate environment: in beds, walls, granaries, caves, rockshelters, and under floors. Fleas, mites, bedbugs, triatomiid bugs, and the diseases transmitted by these bugs are examples of nidicolous parasites. They depend upon the host not only for food but for creation of their habitat. Permanent parasites live on or in the host except when dispersing between hosts. These include most protozoa, roundworms, flukes, and tapeworms. They are completely dependent upon their host for food and all other environmental requirements.

Chacoan Great Houses as Nidi for Infection

Factors Outside Great House Environments Water Source, Giardiasis, and Amoebic Dysentery

Realizing that for parasitic disease to occur, all factors related to the survival and reproduction of the parasite must be present, Pavlovsky (1966) combined ecological factors into a predictive tool for infection. These can include vectors, reservoir hosts, humans, and favorable external environments. He defined a nidus as that portion of a natural geographic landscape that contains a community consisting of a pathogen, vectors, reservoir hosts, and recipient hosts, and that possesses an environment in which the pathogen can circulate. He further found that pathogens possessed nidality: the tendency of an infectious agent to occur in distinct nidi, such as being associated with particular geographic, climatic, or ecological conditions. Thus, a nidus is a focus of infection. For humans, a nidus can be as confined as a single room accessed by rodents carrying plague-infected fleas, or as large as an entire community and its agricultural use-area, as is the case for the transmission of hookworms. Various types of parasites circulate in nidi. Temporary parasites, which live in the external environment and come to the host only to feed, include mosquitoes, chiggers, ticks, and leeches. In these species, every individual must have good dis-

Water sources in desert environments are foci for human activity and can therefore become nidi. As long as they are plentiful and flowing, and popu­ lations are not too concentrated around them, ­water sources are not necessarily a pathoecological factor in the spread of parasitism. However, when ­water sources are few in number and stagnant, and when populations aggregate around them, these sources often becomes contaminated, providing a nidus and becoming significant pathoecological problems. Giardia lamblia has been found in ancestral Pueblo coprolites (Gonçalves et al. 2002). This parasite is not highly pathogenic in most adults. In fact, most infected people show no symptoms. However, when G. lamblia becomes established in stagnant water sources, it becomes a problem. It is most perilous to pregnant women and their babies. Disease in mothers and children is due to poor maternal nutrition caused by malabsorption, resulting in intrauterine growth retardation. G. lamblia causes malabsorption when the intestinal villi become blunted and the function of intestinal mucous di-

90

Parasite Pathoecology minishes (Carden and MacLeod 1988). Clinical symptoms include cramps, watery diarrhea, nausea, vomiting, and sometimes fever. Among pregnant women who exhibit symptoms, G. lamblia causes malabsorption and dehydration at a period when there is a need for accentuated nutritional requirements. Such women fall into a negative nutritional balance, as demonstrated by Carden and MacLeod (1988), who summarized the effects of G. lamblia on the fetus and newborn. With protracted maternal infections, normal fetal development is impeded. With asymptomatic maternal infections, low birth weight and infant anemia are common (De Morais and Suzuki 1997). Generally, infants become infected after three months of age. Islam et al. (1988) found that some immunity is conveyed from mother to infant, but this immunity is not effective in infected infants. Immunity to G. lamblia increases with age (Shetty et al. 1992). Thus, the pathology caused by G. lamblia is most significant in infants and toddlers (Hjelt et al. 1992). The epidemiology of giardiasis is well known (Taus et al. 1998; Hjelt et al. 1992; Harter et al. 1982). Sullivan et al. (1991) showed that giardiasis is highly prevalent in children with chronic diarrhea and malnutrition, and that giardiasis does not respond to standard therapeutic measures. Children who have low iron or vitamin B12 levels have more severe giardiasis symptoms (Awasthi and Pande 1997; Olivares et al. 2002). Subadults in the age range of 9 months to 11 years are most susceptible to infection, though infections can occur at 3 months of age. In developing nations, 91 percent of infants of infected mothers become infected by 6 months of age. Of infected infants, 86 percent have diarrhea. Infected people tend to live in dwellings with dirt floors, simple latrines, ground­water drinking sources, and close contact with dogs. These aspects of life were common at ancestral Pueblo villages (Reinhard 1996). In addition, ­person-​to-person transmission of G. lamblia is common (Birkhead and Vogt 1989; Black et al. 1977; Keystone et al. 1978). The parasite Entamoeba histolytica also afflicted ancient Pueblo groups (Gonçalves et al. 2002). Relative to G. lamblia, E. histolytica causes more dramatic pathology, creating ulcerations in

the large bowel or ileum. Amoebas can cause nodular granuloma formation, colitis, and diarrhea. The disease can become systemic and eventually an ulcerative disease of the large intestine, liver, lung, brain, or other ­organs. Amoebiasis can be symptomatic or even fatal during pregnancy (Abioye 1973; Lee 1929; Lewis and Antia 1969; Rivera 1972). Deaths that occur are due to a rapid onset of profuse diarrhea with dehydration and severe anemia. Premature delivery results from colitis, diarrhea, dehydration, ketosis, or shock (MacLeod and Carden 1988). Weigel et al. (1996) found that high E. histolytica load in asymptomatic infections was associated with decreased maternal serum hemoglobin and hematocrit ­levels, and iron-deficiency anemia. Among women who had severe problems (spontaneous abortion, stillbirth, low-birthweight babies), there was a fourfold increase in the prevalence of amoebiasis relative to normal births (Czeizel et al. 1966). In infected but asymptomatic mothers, Weigel et al. (1996) found increased indicators of diminished intrauterine growth. Despite immunity conveyed by antibodies passed through the placenta and in milk, infants can become infected. When this happens, infants exhibit fever with severe ­watery, sometimes bloody, diarrhea. Colitis, appendicitis, intestinal rupture, and peritonitis result in a high mortality among infected infants (MacLeod and Carden 1988). The Pueblo III occupation of Antelope House, at Canyon de Chelly in Arizona, is the bestdocumented­ case of an ancestral Pueblo village that suffered declining health due to water source nidi. Morris (1986) describes the pathoecological conditions that led to water contamination. Towards the end of the occupation, drought affected the region. As more distant water sources dried up, the population of Antelope House and Canyon de Chelly burgeoned. The increased population and decreased water resulted in contamination. Gonçalves et al. (2003) found both E. histolytica and G. lamblia in Antelope House coprolites. El-Naijar (1986; El-Naijar et al. 1976) found increased skeletal evidence of systemic disease during the Pueblo III occupation of Canyon de Chelly relative to other time periods. Thus, there is a relationship

91

Karl J. Reinhard b­ etween environmental stress, increased parasitism, and skeletal indicators of morbidity in mother and infants. For Chacoan great houses, coprolites from Salmon Ruin were tested for G. lamblia with negative results (Wilson et al. 2006). The absence of giardiasis at Salmon is logical given the presence of a flowing water source (the San Juan River) within 200 m of the community. Given these conditions, there was little chance for contamination. No coprolites from Chacoan great houses have been tested for E. histolytica.

cate human hosts and burrow through the skin. Also, A. duodenale can achieve transmammary and transplacental infection. Thus, fetuses and infants can be infected without ever coming in contact with contaminated soil. Hookworm causes specific problems in pregnancy. One of the most common causes of death in labor in the developing world is cardiac failure from severe anemia attributed to hookworm infection (Cintron Villaronga 1967). As many as 90 percent of pregnant women are infected in endemic areas (Ananthakrishnan et al. 1997; Navitsky et al. 1998). Crompton and Whitehead (1993) present calculations comparing effects of hookworms on a nonpregnant woman versus a pregnant woman. The model predicts that hookworms more rapidly deplete stored iron, with a rapid effect on red cell density per milliliter of blood in pregnant women. MacLeod (1988) verified this model from the clinical perspective. Each worm consumes 0.27 ml of blood per day, and only 20 weeks after initial infection, hypochromic, macrocytic anemia can develop. The minor symptoms of infection are indistinguishable from complaints of pregnancy (epigastric pain, heartburn, and so on). However, with moderate infections there is low-grade fever, fatigue dyspnea, heart palpitations, flow murmurs, and anemia. In heavy infections, constipation or diarrhea, jaundice, emaciation, cardiac failure, or pre-eclampsia occur. If a woman survives labor, she cannot recover as easily from post-partum hemorrhage, which can contribute to maternal death. Hookworms also have a negative impact on fetuses and infants (MacLeod 1988). Abortion, stillbirth, and premature labor are associated with severe hookworm infection. Women infected with hookworm give birth to low-birth-weight infants (a 2 percent hematocrit drop in the mother correlates to a 100 gram decline in birth weight). Because of transplacental migration, infants are infected at birth. Severe and sometimes fatal hemorrhage occurs in infants less than four months of age. Chaudhary and Jayaswal (1984) first described an anemic infant resulting from transplacental migration. In a survey of hundreds of transplacental-infected infants in China, Yu et al. (1995) defined the symptoms of transplacental infection, which include

Irrigation, Hygiene, and Hookworm

Hookworm has been found in coprolites from Antelope House and Pueblo Bonito, but at no other ancestral Pueblo. Hookworm is the greatest parasitic threat to the mother, fetus, and infant. Irondeficiency­ anemia resulting from intestinal blood loss is the major consequence of hookworm infection (Variyam and Banwell 1982; Ali et al. 1990). Treatment for this type of anemia is administration of iron supplements. According to Gilman (1982), development of hookworm-induced irondeficiency anemia is dependent on the intensity of infection, the species of hookworm, and the ability of the host to resist infection and to maintain adequate stores of iron. Loss of blood is caused by direct ingestion of red blood cells and by tissue trauma produced by worm attachment and feeding. The species that causes the more ­serious pathology and that has been identified in ancient New World remains is Ancylostoma duodenale (Allison et al. 1974). This is a fascinating, humanspecific­ parasite that has evolved several infection modes and adaptations. Perhaps the most remarkable aspect of A. duodenale is its hypobiotic ability. Hypobiosis occurs when a parasite suspends it development in host tissues in a way that prevents a strong immunologic response. A. duodenale can go into hypobiosis in winter and come out of hypobiosis in summer. This is a significant adaptation because the females can lay their eggs in the season that is optimal for larval survival. The larvae hatch within a few days, exit the feces, and develop through three larval stages as free-living soil nematodes. Subsequently, as third-stage larvae, they lo92

Parasite Pathoecology village location that aggravated pinworm infection. Although pinworm prevalence was highest in walled villages built within rockshelters, it is important to note that no thoroughly studied ancestral Pueblo site has been found to be pinworm free. Related cultures also were infected. The earliest Basketmaker II coprolites from Bighorn Cave (Grand Gulch, Utah) have a prevalence of 25 percent. Later, the diffuse populations of ancestral Pueblo and Fremont in the Glen Canyon area were infected. Even the Sinagua inhabitants of Elden Pueblo were infected (Hevly et al. 1979). The infections resulted from air humidified by human activity and contaminated with floating eggs within confined spaces. Although pinworm infection tends to be asymptomatic, a high prevalence of heavy infections can result in severe pathology, including secondary bacterial infections in juveniles. However, to my mind the real relevance of pinworm relates to other diseases that are also airborne transferred. For the ancestral Pueblos, tuberculosis was the other airborne disease. Among the most poignant epidemiological descriptions of the tuberculosis threat to Pueblos, is applicable to pinworms as well, is provided by Fink (1985), who examined details of Anasazi life such as communal living, lack of knowledge of germ theory, and cramped living conditions that promoted infectious diseases. The San Juan period at Salmon Pueblo is enigmatic in the context of pinworm prevalence at other sites. Only 7 percent of 112 coprolites studied contained pinworm eggs—approximately onefifth the prevalence recorded for other Pueblo III sites, including Pueblo Alto and Pueblo Bonito. A probable explanation for this relates to Paul Reed’s (2006c) finding that the San Juan residents of Salmon used primarily the second-floor rooms for human activities. Air conditions in the lowest rooms, and those closest to the windowless rear wall, would have been more likely to increase pinworm infection. Such rooms would have had the most stagnant and humid air, promoting airborne infection with pathogens. Use of hearths in secondfloor rooms would have produced a much less humid environment, and any rooms opening towards the large Salmon plaza would have been healthier

bloody stools, melena, anorexia, listlessness, and edema. A. duodenale was the species implicated in these types of infection. Transplacental migration is not rare. Nwosu (1981) documented that 10 percent of 316 Nigerian newborns (four to five weeks old) were infected with A. duodenale. Transmammary infections from mother to infant also occur, with similar health results (MacLeod 1988). Studies of many groups from around the world link hookworm disease, especially from A. duodenale, to severe iron deficiency and anemia in children (Albonico 1998; Stoltzfus et al. 1998). Hookworm infection is dependent on moisture, shade, and warmth. The Colorado Plateau is normally too dry to promote infection, and in historic times hookworm was unknown. Thus, the discovery in Anasazi sites of hookworm eggs and another parasite with a similar infection mode, Strongyloides stercoralis, was surprising (Reinhard 1985c; Reinhard et al. 1987). Clearly, ancestral Puebloans created microenvironment nidi where parasite larvae could hatch and mature in moist, warm, and shaded soil. Puebloans also spent time in these nidi, where they spread eggs and became infected by larvae. It is very likely that irrigated fields were hookworm and S. stercoralis nidi. Studies of hand and foot washing in Bengal show that the larvae can be washed off easily within a few minutes of coming into contact with the skin. The infection occurred in defecation grounds, and washing was prescribed by religious rules (Nawalinski et al. 1978). We do not know if ancestral Puebloans had similar rules, but it is very likely that hookworms could penetrate the skin of Puebloan farmers as they worked in irrigated fields. If the ­division of labor resulted in men working more in irrigated fields, it may be that they were more often infected than women.

Internal Great House Factors Apartments, Plazas, Kivas, and Second-Floor Living

Pat Horne (1985) attributed the remarkable pinworm prevalence among ancestral Pueblos to crowded, apartment-style living conditions. As noted above, Hugot et al. (1999) elaborated this theme by detailing the aspects of architecture and 93

Karl J. Reinhard with ventilation from the dry, relatively breezy air outside the pueblo. Subterranean rooms such as kivas would have provided the primary vector for transmitting airborne disease. In kivas, the air would have been ­humid, and the air flow around the ventilator would have been sufficient to transmit particles around the room, but not to remove infectious particles from the structure. This would explain, in part, why Basketmaker groups, who lived in pithouses with little or no circulation, developed a high prevalence of infection. After Basketmaker times, however, kivas would have been the most likely subterranean nidi of pinworm dissemination. One way to test this ­hypothesis is through analysis and comparison of sediment samples from kiva, living room, and milling room floors. By far the healthiest place to work and live at Salmon was outside, in the plaza and on rooftops. Sunlight would have desiccated and radiated pathogens, thereby reducing the number of infectious airborne contaminants, and the clean air moving across the plaza would have provided people with alternate, healthy air. Humidity also was undoubtedly much lower in these open spaces compared to confined rooms and kivas.

the scolex. As the they progress downward along the length of the tapeworm, their ovaries and testes mature, fertilization occurs, and eggs mature. When a proglottid is filled with mature eggs, it is said to be gravid. In some tapeworms, such as those that infect humans who eat undercooked fish, the eggs are laid through gravid proglottid genital pores. In other species, such as those that infect humans who eat undercooked beef, entire gravid proglottids break off of the tapeworm. These proglottids are partly motile and squirm their way out of the host body. Hymenolepis lays eggs through genital pores, and these eggs are infective when they pass into the environment. Although they have been found in Canyon de Chelly coprolites, they have not been found in Chacoan great houses.

Conclusion The Chacoan great houses provided many potential nidi for temporary, permanent, and nidiculous parasites. Great house inhabitants created or eliminated nidi through different activities and practices. Although there is was no way to completely eliminate the transmission of permanent parasites, some aspects of life at Salmon reduced the prevalence of pinworm relative to other great houses. A lower population density, among other factors, would have accomplished this. The absence of fecal-borne parasitism indicates that nidi of fecal exposure were eliminated at Salmon Ruin through the use of specific rooms as latrines, an effective way of stopping the spread of parasites such as Giardia lablia. Nidi external to Salmon Ruin where hookworm and S. stercoralis transmission could have taken place did not exist. This was probably due to a different type of irrigation (perhaps using the free-flowing San Juan River) and gardening relative to that of Pueblo Alto and Pueblo Bonito, where hookworms did infect humans. In the future, more extensive analysis of ancestral Pueblo coprolites should be conducted using a variety of research methods. Some sites, such as Antelope House, are currently the focus of molecular, immunological, and microscopic analysis. Other sites, such as those in Glen Canyon, were studied

Sanitation and Hymenolepidid Tapeworm Infection

With regard to tapeworms, there are two main types of hosts. The definitive host is the animal in which tapeworms accomplish sexual reproduction, whereas intermediate hosts are infected with nonsexual stages. Usually, tapeworm infections in humans occur through ingestion of infected intermediate hosts. The most common tapeworm (Hymenolepis nana) found in ancestral Puebloan coprolites took a different infectious pathway. This tiny tapeworm evolved the ability to use an intestinal villus as its intermediate host. The larvae emerge and become adults in the intestinal lumen, thus using humans as both their definitive and intermediate hosts. Tapeworms have two methods of laying eggs. The tapeworm’s anterior end, the scolex, attaches to the intestinal wall. Proglottids are the sexually reproducing tapeworm organs that develop from 94

Parasite Pathoecology only through microscopy and could yield beneficial data with newer approaches. Once a range of methods has been applied to a larger sample of sites, we will be able construct a more complete picture of ancestral Pueblo parasite pathoecology. In addition, parasitological methods must be developed for analysis of remains other than coprolites. Many nidiculous pathogens such as bed bugs and kissing bugs live in walls and roofing. There-

fore, archaeological excavations should include soil samples from architectural remains to search for the presence of insect exoskeletons. Also, analysis of trash sediments for all types of parasites must be developed in order to obtain parasitological data from sites that lack coprolites. Once these approaches are developed, then a true archaeology of parasitic disease will emerge.

95

6

Animal Bone from Salmon Ruins and Other Great Houses Faunal Exploitation in the Chaco World Kathy Roler Durand and Stephen R. Durand Over 150 years after Lt. Simpson (1850) first traveled through Chaco Canyon, Chacoan great houses still capture our imagination. Why were they built? What was their purpose? Despite decades of research, much still remains unknown about great houses, particularly those in outlier communities. One thing that is clear about the outlier communities is how variable they were (e.g., Kantner 1996, 2003; Marshall et al. 1979; McKenna and Toll 2001; Powers et al. 1983; Van Dyke 1999, 2003a; and the papers in Kantner and Mahoney 2000); therefore, each outlier can potentially broaden our understanding of the Chaco world. We have come a long way in our understanding of Chaco since Cynthia Irwin-Williams’s work at Salmon in the early 1970s, and it is worth revisiting our interpretation of this important outlier. This chapter summarizes our analysis of the Salmon faunal remains in an attempt to realize a portion of Irwin-Williams’s original vision (this volume, 2006b) and to place these data and conclusions within the modern context of our understanding of the Chacoan and post-Chacoan world. We go beyond bones as indicators of subsistence to explore the frequencies and diversity of non-economic fauna (ritual fauna) and conclude that there is clear and robust patterning in both aspects of the faunal assemblage from Salmon Ruin. We use the terms Chacoan and post-Chacoan generically to refer to the Primary and Secondary occupations at Salmon, as well as the pre- and postad 1130 occupations in Chaco Canyon and the middle Rio Puerco Valley. We prefer these terms

because they work well in different parts of the San Juan Basin. For the Chacoan period we have included samples from Chaco Canyon (from the Una Vida great house and two levels from the great house of Pueblo Alto), the Guadalupe Ruin community (from the Guadalupe Ruin great house and six small houses at the base of Guadalupe Mesa), and Salmon Ruin (from five rooms with Chacoan deposits). For the post-Chacoan period our samples include a small sample from Aztec’s West Ruin (from Room 202, excavated to solve drainage problems in 1983), two assemblages in Chaco Canyon (the post-Chacoan remains at Pueblo Alto and 29SJ 633), two assemblages from the Guadalupe Ruin community (from the great house and ENM 883, one of the associated small houses), three rooms at Salmon Ruin with post-Chacoan remains, and eight sites in Sand Canyon (including Sand Canyon Pueblo). These sites are listed in Table 6.1 (which includes the references for the faunal data not collected by K. Durand as part of this study), and Figure 6.1 shows their locations. This study of the Salmon fauna was initiated in 2002 in conjunction with the Salmon Ruins– Center­for Desert Archaeology partnership, as part of the Salmon Research Initiative (Reed, Chapter 1, this volume). The faunal samples from Salmon Ruin came from trash deposits in seven rooms (33, 36, 43, 62, 63, 81, and 130 [the great kiva]) (Figure 6.2). The remains are described in more detail in a technical report on the fauna (Durand and Durand 2006). Several other rooms were considered 96

Animal Bone from Salmon Ruins and Other Great Houses

Figure 6.1.  Locations of sites and localities in the San Juan Basin and Four Corners region discussed in the text.

Changes in Dietary Faunal Exploitation at Chacoan Sites

for inclusion in this study (including Rooms 31, 61, 64, and 127), but they contained no or very few bones. Samples from Rooms 100 and 129, plus a larger sample from Room 62, are planned for future phases of analysis. The seven rooms included in the present study were chosen because they represent some of the best contexts at the site. Most of the analyses that follow are based on the number of identified specimens (NISP). NISP allows us to calculate comparative indices and will allow others to use these data in their own research.

In the Southwest, there are three primary categories of economically useful fauna: artiodactyls, lagomorphs, and turkeys. Other species, including rodents and fish, were eaten on occasion, but these three provided the vast majority of the meat in prehistoric Southwestern diets. For Salmon Ruin there are patterns in each of these categories that, in a broad sense, mirror patterns found at contemporaneous sites across the northern ­Southwest. Faunal 97

Kathy Roler Durand and Stephen R . Durand Table 6.1.  References for Sites Included in This Study

Chaco Canyon Great Houses   Pueblo Alto   Una Vida

Chacoan

Post-Chacoan

Data Reference

X X

X

Akins 1985, 1987 Akins 1985

X

Akins 1985; Gillespie 1991

X

Chaco Canyon Small House   29SJ633 Guadalupe Ruin Community   ENM838 — Guadalupe Ruin   ENM845   ENM846   ENM850   ENM852   ENM881   ENM883 — Eleanor Ruin

X X X X X X X

X

Pippin 1987; Roler 1999 Roler 1999 Roler 1999 Roler 1999 Roler 1999 Roler 1999 Roler 1999

Salmon Ruin

X

X

Durand and Durand 2006

X X X X X X X X

Muir 1999 Muir 1999 Muir 1999 Muir 1999 Muir 1999 Muir 1999 Muir 1999 Muir 1999

Sand Canyon Community   Sand Canyon Pueblo   5MT262   5MT1825   5MT3901   5MT3936   5MT3967   5MT5152   5MT10508

changes at Salmon through time point to an increasing reliance on domestic fauna for protein, possibly reflecting changes in the local environment. The NISP present in several categories of dietary fauna from Salmon and other sites in the Chaco region are listed in Table 6.2. Table 6.3 provides the indices that are commonly calculated for these fauna to facilitate comparison of assemblages of varying sizes.1 While changes occur through time in the frequencies of all the economic fauna, the greatest changes are found for the artiodactyls and turkeys. Artiodactyls (bighorn sheep, deer, elk, and pronghorn) show less variation among sites during the Chacoan period, as reflected in a narrower range in the artiodactyl index (AI) during this period (from .27 to .60) (see Table 6.3). This

index is computed by dividing the number of artiodactyls and large mammals (likely all artiodactyls) by this sum plus the total number of lagomorphs ([artiodactyls + large mammals]/[artiodactyls + large mammals + lagomorphs]). Thus, higher AI values indicate assemblages with more artiodactyls in comparison to the number of cottontails and jackrabbits present. The AI values from the Chacoan period assemblages are fairly high, suggesting that artiodactyls were an important component of the diet at that time. For the post-Chacoan ­period, the range in AI values expands slightly (.07 to .45), and most assemblages have lower indices than they did in the previous period. Sites with low indices in the post-Chacoan period include both great houses (.12 at Aztec and .21 at Salmon Ruin) and

98

Figure 6.2.  Generalized ground plan of Salmon Ruin showing rooms sampled for this study (shaded). Analysis is ongoing for the rooms shown hatch marked. Late intramural kivas and late room-dividing walls not shown (from Durand and Durand 2006, adapted from R. Adams 1980:211).

Table 6.2.  Dietary Fauna from Chacoan Sites Artiodactyls

Lagomorphs

Turkeys

NISP

Percent

NISP

Percent

NISP

Percent

Total NISP

Chacoan Assemblages   Guadalupe Ruin   Guadalupe Ruin Community   Pueblo Alto   Salmon Ruin   Una Vida

178 131 548 283 217

58.7 21.3 6.9 46.2 15.4

123 474 7,356 319 1,172

40.6 77.2 92.2 52.1 83.4

2 9 71 10 17

0.7 1.5 0.9 1.6 1.2

303 614 7,975 612 1,406

Post-Chacoan Assemblages   Aztec West   Guadalupe Ruin   Guadalupe Ruin Community   Pueblo Alto   29SJ 633   Salmon Ruin

9 1,176 54 284 8 29

2.9 39.5 5.0 7.0 0.4 6.6

290 1460 652 2,906 1,452 363

93.9 49.0 60.3 71.4 65.2 82.9

10 343 375 878 766 46

3.2 11.5 34.7 21.6 34.4 10.5

309 2,979 1,081 4,068 2,226 438

Note: Data for Pueblo Alto taken from Akins 1987:624, Table 8.144; Data for 29SJ 633 taken from Akins 1985:420, Appendix 2; Data for Una Vida taken from Akins 1985:413, Appendix 2.

Kathy Roler Durand and Stephen R . Durand Table 6.3.  Artiodactyl, Lagomorph, and Turkey Indices  

Artiodactyl Index

Lagomorph Index

Turkey Index

Chacoan Assemblages   Guadalupe Ruin   Guadalupe Ruin Community   Pueblo Alto - Red Mesa   Pueblo Alto - Gallup   Salmon Ruin   Una Vida

0.59 0.47 0.27 0.39 0.60 0.27

0.83 0.84 0.58 0.55 0.83 0.47

0.07 0.12 0.002 0.01 0.14 0.06

Post-Chacoan Assemblages   Aztec West   Guadalupe Ruin   Guadalupe Ruin Community   Pueblo Alto   29SJ 633   Salmon Ruin

0.12 0.45 0.15 0.41 0.21 0.07

0.87 0.88 0.81 0.45 0.67 0.71

0.10 0.19 0.53 0.23 0.46 0.35

Note: Artiodactyl and Lagomorph Indices for assemblages from Pueblo Alto, Una Vida, and 29SJ 633 were provided by Nancy Akins (pers. comm. 2004). Turkey Indices for Pueblo Alto do not include large bird values. This is also true for 29SJ 633, but is unlikely to affect the TI value for the site as only three unidentified bird bones were reported (Gilllespie 1991:247). The TI for Una Vida includes the unidentified aves, which may slightly inflate the TI value if some of these aves were not turkeys. Without including them, the TI value for Una Vida is .01.

small houses (29SJ 633 in Chaco Canyon with .07, and Eleanor Ruin in the Guadalupe Ruin community with .15). The drop in AI values may be tied to the increased reliance on turkeys in the post-Chacoan period. In general, turkeys make up only a small portion of the Chacoan period assemblages (turkey indices, or TIs, range from .002 to .14), whereas they make up a much larger part of the assemblages in the post-Chacoan period (with TI values ranging from .10 to .53). Furthermore, every site that has remains from both occupation periods shows an increase between periods: TI values increase from .07 to .19 at Guadalupe Ruin, from .12 to .53 in the Guadalupe Ruin community, from .002/.01 to .23 at Pueblo Alto, and from .14 to .46 at Salmon Ruin. This pattern matches that documented by Muir (1999) and Munro (1994) for the Mesa Verde region (indices for sites in Sand Canyon are given in Table 6.4) and supports the interpretation that turkeys were being raised across the northern Southwest after the Chacoan period. Akins (1987), Gillespie (1991), and Windes (1987c) have all argued that turkeys were being raised by the early ad 1100s in

Chaco Canyon. As Spielmann and Angstadt-Leto (1996:​93) note for the Chaco region, “where temporally separable faunal data are available..., the frequency of turkey increases over time.” This is clearly the case at the sites in our study. Note that we prefer the term raising turkeys to domestication of turkeys. Munro (1994), in an excellent study on Southwestern turkeys, was unable to distinguish wild from domesticated turkey based on osteological evidence. We concur with her conclusions and could not say it better:

100

The individuals involved in the pioneering research concerning prehistoric southwestern turkeys (McKusick, Hargrave, Schorger) were well seasoned in the field of avian zooarchaeology and performed hands-on research with a range of turkeys from all corners of the Southwest. Certainly this has provided them with an eye for noting variation within turkey populations. However, if only researchers with a specialization in turkey identifications are capable of recognizing the distinctions, they are undoubtedly subtle and subjective and should be

0.04 0.05 0.06 0.06 0.00 0.06 0.01 0.19 0.82 0.87 0.79 0.91 0.88 0.85 0.96 0.80 209 727 1357 136 283 293 753 6,344 94 334 761 60 189 108 530 3,408 47 229 230 23 102 57 348 1,961 47 105 531 37 87 51 182 1,447 120 413 636 81 94 196 226 3,604 5 20 40 5 0 11 3 668 115 393 596 76 94 185 223 2,936 Source: Muir 1999.

Cottontails

94 340 473 69 83 157 213 2,337 5MT262 5MT1825 5MT3901 5MT3936 5MT3967 5MT5152 5MT10508 Sand Canyon

A Artio­dactyls

L+A

Turkeys

Large Bird

T+LB

T+LB+L

L

Indices

101

All lagomorphs

Table 6.4.  NISP Data and Indices for Selected Sites in the Sand Canyon Region

This, of course, does not mean that turkeys were not domesticated, but rather that the osteological evidence for separating out a domesticated sub­ species is tenuous at best (Munro 1994:141). There is nothing tenuous, however, about the evidence that turkeys were being raised. We have already mentioned the dramatic increase in the proportion of turkeys in post-Chacoan assemblages, but many other sources of evidence also indicate they were being raised at these sites. This evidence includes the presence of immature birds (Akins 1987:​482), burning, eggshell (Windes 1987c), and healed wing and leg bone fractures (Durand and Durand 2006; Gillespie 1991). Five healed limb fractures were observed on turkey elements at Salmon and the Guadalupe Ruin community, whereas only two healed fractures were found for all of the other species (the pubis bone of a mouse and a small mammal rib). Similar patterns have been cited as evidence of turkey husbandry at other late prehistoric sites including Alfred Herrera, Arroyo Hondo, and North Bank (Lang and Harris 1984:​98–​102). In fact, Lang and Harris (1984:​102) found a nearly identical set of healed fractures at Arroyo Hondo, with six healed turkey fractures but only one other healed fracture for any other species (again, a mouse element). As they suggest (Lang and Harris 1984:​102), “Under natural conditions, there is little chance that birds thus injured would survive long enough to experience full healing.” Although we have not analyzed the eggshell remains from Salmon, we can present some initial summaries of the occurrence of eggshell at the site. The following summary is derived from the site’s artifact catalog, and we simply counted the strata from which eggshell was collected. The amount and kind of eggshell from each stratum has not

T

interpreted as representing different sub­species with extreme caution. Subtle differences may simply represent natural variation within a population composed of one species. It is indisputable that osteological variation exists within the species Meleagris gallopavo, but splitting the population into distinct subspecies or breeds based on this variation may be unwarranted. (Munro 1994:138)

0.45 0.46 0.56 0.44 0.67 0.37 0.70 0.54

Animal Bone from Salmon Ruins and Other Great Houses

Kathy Roler Durand and Stephen R . Durand Table 6.5.  Salmon Ruin Strata with Eggshell ­Present

Primary Strata Secondary Strata

With Eggshell

Without Eggshell

101 212

840 1,152

been determined. According to Paul Reed (pers. comm., 2005) 2,851 unique stratigraphic units were recorded during the excavations at the site. Of these strata, 436 contained eggshell, and the temporal affiliations of the strata with eggshell are presented in Table 6.5. A little more than twice as many postChacoan strata contain eggshell than Chacoan strata; however, there also are more post-Chacoan than Chacoan strata at the site. The percentage of Chacoan strata with eggshell is 10.7 percent, while 15.5 percent of the post-Chacoan strata had eggshell present. Although this difference is statistically significant (χ2 = 10.97, p = 0.001, df = 1), it is smaller than we expected based on the temporal difference in turkey bone frequency. This finding is difficult to evaluate due to the potential for varying collection strategies during the excavation of Salmon and the variability in the quantity of eggshell per bag. In the examples we have seen, one bag may contain a single eggshell or dozens of fragments. Thus, we are unable to get a true picture of the quantity of eggshell from the artifact catalog alone. Nevertheless, the presence of so many strata at Salmon with eggshell strongly supports the argument that turkeys were being raised there. More than half of the strata that contain eggshell (236 of 436) are from five rooms at the site (Rooms 36, 62, 91, 100, and 129), all of which contained a great quantity of faunal remains as well. Finally, the presence of turkey pens also has been used to document turkey rearing. For example, 18 turkey pens were discovered at Arroyo Hondo (Lang and Harris 1984:102–105). Although such structures were not mentioned for Guadalupe Ruin (Pippin 1987) or Pueblo Alto (Akins 1987), Rooms 62B, 127E, and 128 at Salmon Ruin may have been used as turkey pens (P. Reed 2006c). It is clear that turkeys were being raised during the post-Chacoan period, but questions still re-

main about their potential contribution to the diet. Akins (1987:​482–483) notes that turkeys raised in Chaco Canyon would likely have been fed corn because the canyon environment would have provided poor foraging. Consequently, she (Akins 1987:​ 483–​484) proposes that turkeys were used primarily for their feathers, not for their meat, until the last period of occupation at Pueblo Alto. Lang and Harris (1984:​133) suggest that in the Arroyo Hondo region during times of adequate rainfall, forage and supplemental corn would have been easy to obtain for raising turkeys. It seems likely that the environment around Salmon and Guadalupe Ruins would have been similarly fruitful for turkey husbandry, at least during non-drought years. Thus, any corn that the turkeys consumed is not likely to have had a negative impact on the communities’ diet. Turkeys also can contribute to maize production by helping to keep insects out of the fields once the crops are established (Phillip Shelley, pers. comm., 2004). Thus, turkeys were an important component of the post-Chacoan diet at Salmon, although their feathers no doubt remained important in the construction of ritual paraphernalia. In summary, there is an inverse relationship between artiodactyls and turkey at many sites in this study. While artiodactyl frequencies decrease through time at most sites (Pueblo Alto being the notable exception), turkey frequencies increase, suggesting that animal husbandry began to replace or supplement hunting as a source of meat in the prehistoric diet. This fits Speth and Scott’s (1989) proposal that, at an “unknown threshold of resource depletion,” communities may have needed to obtain protein through means other than hunting. In addition to this dietary change through time, a regional difference appears among the great houses during the post-Chacoan period that is not found in the Chacoan period. Both of the northern great houses considered here (Aztec and Salmon) have low AI values in the post-Chacoan period, with that for Salmon Ruin dropping from .60 to .21 over time. Pueblo Alto and Guadalupe Ruin continue to have comparatively higher values (.41 and .45, respectively), suggesting that artiodactyls were still being heavily utilized at these sites. In fact, the AI value for Pueblo Alto steadily increases through

102

Animal Bone from Salmon Ruins and Other Great Houses time, from .27 (for remains dating to ad 920–1020) to .39 (ad 1020–​1120) to .41 (post-ad 1120). In addition to this pattern, at the Guadalupe Ruin community there is a much larger difference between the great house and associated small house community in the post-Chacoan period than was found for the Chacoan occupation. Pueblo Alto and 29SJ 633 show the same pattern in the post-Chacoan period (although it must be noted that 29SJ 633 is much farther from Pueblo Alto than the small houses at the Guadalupe Ruin community are from the Guadalupe Ruin great house). This suggests that at Guadalupe Ruin, and likely at Chaco Canyon, utilization of artiodactyls was increasingly restricted to the great house. Whether this was due to restrictions on hunting rights, tribute, or the use of artiodactyls in feasting activities is unclear. While artiodactyls vary in importance from site to site, and turkeys change in frequency over time, lagomorphs (cottontails and jackrabbits) were important components of the diet at nearly all prehistoric communities. The lagomorph index (LI) is calculated by dividing the number of cottontails by the total number of lagomorphs: thus a high value indicates that the lagomorphs were mostly cottontails, while lower values indicate that jackrabbits were proportionately more common. The LI values we have calculated for sites across the northern Southwest tend to be high (see Tables 6.3 and 6.4), often .8 or greater, indicating a high frequency of cottontails. The great houses in Chaco Canyon, however, have quite low LI values (below .6), indicating that cottontails and jackrabbits are almost equally represented in these assemblages. Because jackrabbits adapt more readily to habitats that have been disturbed by humans (Szuter and Bayham 1989), it is possible that these LI values are due to differences between the habitats in and around Chaco Canyon and those in other parts of the northern Southwest. However, the small house site of 29SJ 633, also in Chaco Canyon, has a higher LI value, suggesting that this may not be the entire explanation for the low LI values at Pueblo Alto and Una Vida. Another contributing factor could be feasting at the great houses. Jackrabbits can be hunted in large numbers by groups with nets (Beagle­hole 1936; Gnabasik 1981; Szuter and

Bayham 1989) and are argued to have been used for feasts at other sites in the Southwest (Potter 2000a). If this was a contributing factor, it would fit with the ceramic evidence suggesting feasting at Pueblo Alto (Toll 1984). Pueblo Alto also had the second highest AI value among the post-Chacoan assemblages considered here, which would support an interpretation of feasting. Akins (1987:​639) argues that the artiodactyl remains do not support an interpretation of feasting, based in part on the lack of higher utility elements in the Pueblo Alto ­assemblage. However, there is a much larger than expected frequency of thoracic vertebrae and ribs at the site (51.0 percent observed versus 24.1 percent expected) during the Red Mesa period (ad 920–​1020) (Akins 1987:488, Table 8.11). These ­elements would be ideal for venison stew, which was used for feasting at Taos and Cochiti pueblos (Gnabasik 1981:​44, 49). Interestingly, the higher than expected frequency of thoracic vertebrae and ribs occurs in the earliest occupation period and again in the last period of occupation, when the avi­ fauna, which are likely ritually related, also flourished (see below).

Food For Thought It is a commonly held idea that Southwestern farmers were heavily dependent upon maize, especially during the time frame that Salmon Ruin was occupied. Spielmann and Angstadt-Leto (1996:80) persuasively argue that “a maize-dominated diet is deficient not only in high-quality protein but also essential minerals, such as iron, and vitamins, such as B12.” In order to maintain nutritional health, Salmon residents would have had to consume meat and/or other plants on a regular basis (Spielmann and Angstadt-Leto 1996:80). Regardless of whether the post-Chacoan inhabitants of Salmon were descendants of the original Chacoan inhabitants, the change in diet between the two occupations resulted in a nutritionally adequate diet within each occupation. Meat from large mammals in the Chacoan occupation was replaced by meat from turkey in the post-Chacoan period. Bohrer and Adams (2006b) report that beans were recovered in nearly half the rooms excavated at Salmon (21 of 48) and occur almost exclusively

103

Kathy Roler Durand and Stephen R . Durand in post-Chacoan contexts. Nutritionally, beans are highly complementary to corn in terms of the amino acids present, and they are an excellent source of protein (Spielmann and Angstadt-Leto 1996:​95). Furthermore, many consumable wild plants also are excellent sources of iron and vitamin C (Spielmann and Angstadt-Leto 1996:96–99). Adams (this volume, 2006c) summarizes the studies of wild plant use at Salmon, which indicate an increased reliance on wild plant foods between the Chacoan and post-Chacoan periods. Changes also are seen at Salmon in the variety of corn grown during each period, shifting from a flinty, Chapalote variety during the Chacoan period, to a flour corn (Maize de Ocho influenced) in the post-Chacoan period (Bohrer and Doebley 2006). Paralleling the change in maize at Salmon, Shelley (2006) documents a change in grinding stone technology from trough to slab metates that probably relates to differences in maize preparation techniques (Bohrer and Doebley 2006). Taken as a whole, subsistence practices were very different in the two occupations at Salmon. Nutritionally, the post-Chacoan occupation may have been a healthier adaptation. As stated earlier, we see an increased reliance on food production in the post-Chacoan period at Salmon. By this we refer to the increased dependence on raising turkeys and growing beans to provide the main sources of protein. Given the substantial population residing at Salmon, the available game would have decreased over time ( Janet­ski 1997; Speth and Scott 1989). The increasing cost of obtaining artiodactyl meat was made up in part through raising turkeys, which is reflected in their dramatic increase during the post-Chacoan­ occupation at Salmon. The cumulative effects of these changes on the local habitat are seen in the increased frequency of jackrabbits at the site over the same period. At least some of these patterns are similar at sites across the central and southern San Juan Basin.

Chacoan Ritual Activity The most popular explanation for Chaco great houses today seems to be their use as ritual centers. As Mills (2002:79) recently noted, “Proponents of the ritual center model [of great houses] domi-

nate the recent literature on Chaco.” The difficulty with these models is finding relevant evidence to test them. Items used in ceremonies often are made of perishable materials that do not preserve well; they are seldom left in situ when a site is abandoned; and it is not always clear which items had ritual significance. Although it is likely that many aspects of modern and historic Pueblo ritual practices are fairly recent developments, the elements common to all or most modern Pueblos probably go back much further in time. Ware and Blinman (2000:​403) suggest that some of these elements may include the kiva and sipapu, meal altars, and “an emphasis on cardinal directions and color associations, to name just a few.” We argue here (and in more detail in a recent article by K. Durand [2003]) that the use of feathers, claws, and pelts from birds and carnivores is another element of Pueblo ritual practices with great antiquity. If this hypothesis is correct, bones from these creatures would have been deposited as a by-product of obtaining feathers, claws, and pelts, thus creating a signature for identifying this activity. Ritual fauna are those that were collected and used primarily for non-dietary purposes. In the Southwest, these include many species of colorful birds and carnivores. Ritual fauna are, by definition, rare in an assemblage, and the current sample size from Salmon Ruin must be kept in mind in the following analysis. The results reported here may change when larger samples from three additional rooms (Rooms 62, 100, and 129) are included in the analysis. Nonetheless, as scholars increasingly converge on a model of great houses as ritual centers of their communities (Durand 2003; Judge 1989; Kantner 2003; Lekson 1999; Lekson et al. 1988; Mal­ville and Malville 2001; Mills 2002; Renfrew 2001; Roney 2001; Stein and Lekson 2001; Toll 1984; Van Dyke 2000), any evidence of ritual activity in these structures is important for our understanding of them. In a recent article (Durand 2003), it is argued that some of the patterns of bird and mammal use for ritual activities at modern pueblos (Gnabasik 1981; Schroeder 1968; Tyler 1979) can be traced back to prehistoric pueblos. The assemblages from Pueblo Alto, Pueblo Bonito, and the post-Chacoan assemblage at Guadalupe

104

Animal Bone from Salmon Ruins and Other Great Houses Table 6.6.  Summary of Macaw and Turkey Burials at Salmon Ruin Room 33 Room 42 Room 86 (Chacoan) (Unknown) (Chacoan) Articulated macaw skeleton Loose macaw elements Other macaw remains Articulated turkey skeleton Human burials also present

1 2 No 0 Yes

2 Possible Unknown Unknown Unknown Unknown

Ruin have an amazing diversity of birds. The importance of some species is also demonstrated from the context in which they were recovered (such as in kiva niches [ Judd 1954:323], articulated burials, and a pit beneath the plaza at Pueblo Alto [Akins 1987:​596–​607]), as well as from other types of artifacts at these sites (such as the wooden bird objects from Chetro Ketl [Vivian et al. 1978]). A similar use of fauna has been found in the prehistoric Mississippian region, where Jackson and Scott (2003) have identified a number of species at Moundville that are associated with ritual contexts, including peregrine falcons, swans, and several species of small, colorful birds. Thus, it is with some measure of confidence that we search for similar evidence at Salmon Ruin.

Avian Burials at Salmon

The remains of nine macaws were recovered at Salmon. The vibrant plumage of these birds, their rarity at Chacoan sites, and the long-distance trade necessary to procure them provide some of the best evidence for ritual fauna at Salmon. In addition, one of two macaws recovered in 2005 from an area of heavy rodent disturbance in Room 42 (immediately north of Room 33) was covered in red ochre prior to burial. The context of the macaws provides another clue to their importance: five of them were recovered as intact burials, and there were also two turkey burials at the site (Table 6.6). The turkey burials and all but two of the macaw burials date to the Chacoan, or Primary, occupation period. The other two macaw burials may date to this period as well, but the mixed nature of their strata make their cultural association unclear at the present.2 Four of the bird burials came from a single room; Room

0 0 Yes 0 Yes

Room 91 (Mixed) 1 0 No 0 No

Room 100 Room 129 (Chacoan) (Mixed) 2 0 No 2 Yes

1 0 No Possible Yes

100 contained both of the turkey burials and two macaw burials, one in the southeast corner of the room adjacent to the two turkey burials, and the other in the doorsill between Rooms 97 and 100. With the exception of Room 91, all of the rooms with macaw burials also contained human burials. This includes Room 86, which did not have an articulated macaw burial but contained several wing and skull elements within the floor stratum, suggesting a macaw skin or bag had been deposited in the floor. Macaw burials also have been found near or with human burials at Cameron Creek, Freeman Ranch, and Galaz in the Mimbres region (Creel and McKusick 1994:519) and at Grasshopper Pueblo (Olsen 1990:59). Although the macaws are restricted to one half of the structure at several Southwest sites, that does not appear to be the case at Salmon Ruin. For example, all of the macaws at Pueblo Bonito came from its eastern rooms (Creel and McKusick 1994; Durand 2003; Pepper 1996 [1920]). Patterning in the distribution of macaws also has been reported for Galaz, Cameron Creek, and Old Town in the Mimbres area, where the macaws were located in the northern halves of these sites (Creel and McKusick 1994:​519). The differential distribution of various avian species has been suggested to represent social divisions at a number of sites (Creel and McKusick 1994:​519–521; Olsen 1990:79–81; Potter 2000b:​307–​308). At Salmon Ruin, Webster (this volume) has found that during the Chacoan ­period, many perishable weaving elements were dyed red before use, thus giving a strong red color to the finished artifacts. These red objects were all located in the eastern half of Salmon Ruin or adjacent to the Tower Kiva. Irwin-Williams (this volume) found a

105

Kathy Roler Durand and Stephen R . Durand

Figure 6.3.  Turkey burial from Room 100. Note the macaw skull near the tail of the turkey skeleton.

dichotomous pattern among the ceramics as well, with most imported wares in rooms in the western half of the pueblo. The macaws, however, appear to be equally distributed across the site rather than in one half. The turkey burials at Salmon (Figures 6.3 and 6.4) are not unique for the Southwest. Akins (1985:​ 380) and Windes (1987c:685) report that turkey burials have been found at numerous sites in Chaco Canyon, including the great houses of Pueblo del Arroyo, Una Vida, and the Pueblo Alto trash mound, as well as at several small house sites. In addition, turkey burials have been found at a few sites in the Mesa Verde region (Munro 1994:​147), at TA32 in the Taos area (Lang and Harris 1984:​ 97–​98), at Galaz in a room with macaw and parrot burials (Creel and McKusick 1994:519), and five are reported from Grasshopper Pueblo (Olsen 1990:​ 82). Turkeys are an especially interesting species in terms of their ceremonial role in prehistory. At sites across the northern Southwest, the number of turkeys is initially small, and whether they were being raised or hunted, it appears they were used mainly for their feathers (Akins 1987:​483–484). By the Pueblo III period, turkeys increased dramatically

at many sites, and this, coupled with the disposal of disarticulated elements in trash deposits, suggests that by this time they were an important component of the diet (Munro 1994). Indeed, this is the pattern found at Salmon Ruin, as discussed above. However, it seems equally clear that turkeys continued to be important for their feathers, as reflected at Salmon in the proportion of turkey feather blankets among the perishables (Webster, this volume). In fact, as ritual activity involving avifauna appears to have increased through time at many Chacoan sites (see below), the feathers may have continued to be as important as the meat in these communities. The presence of turkey burials in the Chacoan deposits at Salmon, the absence of such burials in the post-Chacoan period, and the tremendous increase in turkeys in the post-Chacoan period all reflect the broader trends seen in turkey remains across the northern Southwest.

Faunal Diversity: A Quantitative Look at Ritual Fauna

Comparing assemblage diversity gives us a more objective way of exploring faunal variation among sites. Given the model of great houses as ritual cen-

106

Animal Bone from Salmon Ruins and Other Great Houses

Figure 6.4.  The second turkey burial from Room 100.

ters, one would expect their faunal assemblages to be more diverse than those of small house assemblages, especially their ritual fauna. Faunal diversity is frequently used as one marker of elites and/or ritual activity (e.g., Hamblin 1984; Maxham 2000; Pauketat et al. 2002; Potter 1994; Szuter 1990). Indeed, Jackson and Scott (2003) have documented such differences in both dietary and ritual fauna between elite and non-elite deposits at Moundville. They found that elite contexts were associated with “greater-than-expected large mammal meatbearing­elements,” as well as nearly exclusive access to avian species such as passenger pigeons, peregrine falcons, swans, and a variety of small colorful birds ( Jackson and Scott 2003:555). In looking at differences in assemblage diversity between great houses and small houses, we must be careful when using samples of varying size, such as those in our study. Much ink has been spilled in the last 20 years on the relationship between

s­ ample size and diversity (e.g., Grayson 1984; Jones et al. 1983; Kintigh 1989; Leonard and Jones 1989; Rhode 1988; Ringrose 1993). It is generally held in quantitative zooarchaeology that direct comparisons of assemblage diversity are problematic. Below a given (and variable) threshold, larger samples are more diverse than smaller samples, regardless of the underlying cultural processes generating those assemblages. Two techniques are commonly used for dealing with the sample size effect (although see Baxter [2001] and Kaufman [1998] for other approaches): using regression for testing whether sample size and diversity are correlated in a given set of assemblages (Grayson 1984), and using a Monte Carlo–based simulation technique to create levels of diversity expected at different sample sizes (Kintigh 1989). We have chosen the latter, because the Monte Carlo–based technique helps to highlight those samples that significantly diverge from expected levels of diversity. This Monte Carlo–based

107

Kathy Roler Durand and Stephen R . Durand

Figure 6.5.  The relationship between sample size and diversity using all fauna taxa identified to the species level (90 percent confidence interval depicted). The analysis was accomplished with Kintigh’s DIVERS program in his Tools for Quantitative Archaeology software package (Kintigh 1998). As discussed in the text, the sample size/diversity relationship was simulated using the Pueblo Alto’s Gallup period faunal assemblage (Akins 1987) as the model population. PA = Pueblo Alto, GR = Guadalupe Ruin, ENM = Guadalupe Ruin community sites, 5MT = Sand Canyon Locality sites (Muir 1999), -C suffix = Chacoan age, -PC suffix = post-Chacoan age.

simulation technique takes the diversity from a designated model or a pooled group of assemblages and simulates “draws” from this universe to create expected diversity values (specifically, how many different taxa are likely to be present) for assemblages of various sizes. In this way it allows for the comparison of assemblages that have very different sample sizes. Rather than using the pooled data, we opted to use the Chacoan faunal assemblage from Pueblo Alto’s Gallup period deposits (ad 1020–1120) as the model for this analysis (these data came from Akins 1987:​624). We chose this assemblage because it is very large, it was collected using modern ­methods of excavation and analysis, and it comes from a Chaco Canyon great house, our hypothetically most diverse source. In the first phase of our analysis, all recorded mammalian and avian taxa

were used, although closely related species in the same genus were counted together as one variable (such as various species of Neotoma, or woodrats). In the second phase of our analysis we compared only the potentially ritual component of the assemblages, removing all dietary fauna, rodents, and bats, but leaving the carnivores and avifauna (except turkeys). The data sets for both phases of analysis are available at the San Juan County Archaeological Research Center, Salmon Ruin Museum. The results of the first phase of our analysis are presented in Figure 6.5. In this graph, the range of expected diversity for various sample sizes is shown along with the actual diversity of each of the samples considered here. All of the Chacoan period samples fall within or very close to the level of diversity expected if the samples were taken from a collection identical to that of the Gallup (late Cha-

108

Animal Bone from Salmon Ruins and Other Great Houses

Figure 6.6.  The relationship between sample size and diversity using all ritual fauna taxa identified to the species level (90 percent confidence interval depicted). The analysis was accomplished with Kintigh’s DIVERS program in his Tools for Quantitative Archaeology software package (Kintigh 1998). As discussed in the text, the sample size/diversity relationship was simulated using all Chacoan age faunal assemblages included in this chapter as the model population. PA = Pueblo Alto, GR = Guadalupe Ruin, ENM = Guadalupe Ruin community sites, 5MT = Sand Canyon Locality sites (Muir 1999), -C suffix = Chacoan age, -PC suffix = post-Chacoan age, -RM suffix = Chacoan/Red Mesa age, -G suffix = Chacoan/Gallup age.

coan) component of the Pueblo Alto assemblage. This includes the Chacoan components from the great houses of Guadalupe Ruin, Pueblo Alto, Salmon Ruin, and Una Vida. All of the small house assemblages also fell within the expected range of diversity based on their sample sizes. The situation is different for the post-Chacoan samples: three of the samples from this period (from the great houses of Guadalupe Ruin and Pueblo Alto, plus Sand Canyon Pueblo) are well above the expected range. The other post-Chacoan samples fall within the expected range, including the great houses of Aztec West and Salmon Ruin, plus the small house assemblages from Chaco Canyon (29SJ 633), the Guadalupe Ruin community (ENM 883), and Sand Canyon (5MT262, 5MT1825, 5MT3901, 5MT3936, 5MT3967, 5MT5152, 5MT10508).

In the second phase of our analysis, we ran a Monte Carlo-based simulation using only the fauna that were likely ritually related. We attempted to use the ritual fauna from the Gallup component at Pueblo Alto as the model for this simulation, but the sample size was no longer large enough to model the expected range for the largest samples. Therefore, as we were interested in comparing Chacoan and post-Chacoan use of ritual fauna, we used the pooled counts of the ritual fauna from all of the Chacoan assemblages and then plotted the postChacoan assemblages against this model. The resulting plot is shown in Figure 6.6, which mirrors Figure 6.5 in many ways. Two sites (post-Chacoan Salmon Ruin and one of the Guadalupe Ruin community small houses) contained no ritual fauna, so these assemblages were not plotted. Again, the

109

Kathy Roler Durand and Stephen R . Durand post-Chacoan components of Guadalupe Ruin and Pueblo Alto continue to be much higher than expected given their sample sizes. The Gallup component of the Pueblo Alto assemblage and the assemblages from Aztec West, the Chacoan period at Salmon Ruin, and Una Vida, as well as the assemblage from Sand Canyon Pueblo, all had the expected level of diversity among their ritual fauna. The earliest sample from Pueblo Alto had substantially less diversity than expected, probably because of the relatively high numbers of raptors in the ­sample from the sub-plaza pit, creating a large ­sample size but little diversity during this period. Thus, based on our analysis, the post-Chacoan Pueblo Alto and Guadalupe Ruin assemblages are very different from the others, both in terms of their overall faunal diversity and in the frequency of their ritual fauna. We found this a surprising result, given our expectations about the role of Chacoan great houses in their communities. We had expected a greater difference between the great house and small house assemblages in the Chacoan period, based on the common interpretation of these structures as ritual centers. Furthermore, we expected that there would be a breakdown in this pattern in the post-Chacoan period, during which great and small houses might have become more, not less, similar. Instead, our results suggest the opposite: ritual activities occurring in the Chacoan period do not appear to be restricted to the great houses, but during the post-Chacoan period the amount of ritual activity at some of the great houses increases tremendously. This was not true at all great houses, however, as Salmon Ruin and Aztec West seem to have been left out of this trend. In fact, at Salmon Ruin, overall faunal diversity decreased over time, and the ritual fauna essentially disappeared in the post-Chacoan period. These results suggest that, unlike some great houses, Salmon Ruin was primarily a residential structure in both occupation periods (a conclusion also reached by Reed, Chapter 3, this volume). Some ritual activity did take place at the site, with slightly more evidence of this in the Chacoan period; however, the majority of the remains from both periods are from economic fauna, with macaws being the notable exception.

Reflections on Irwin-Williams’s Model As Vivian (this volume) points out, Cynthia ­Irwin-​ Williams’s research at Salmon was structured around an explicit, testable model of Chacoan culture. Lest we forget her contribution to our understanding of Chaco prehistory, we think it is important to recast our results in terms of her ideas. The linchpin of her argument is that the Chacoan adaptation resulted in a more efficient subsistence base, with an economy that more effectively dealt with the vagaries of an agriculturally marginal environment. Our data from Salmon do show differences between the Chacoan and the post-Chacoan­ period, but they also suggest the latter period was more intensively focused on food production. The post-Chacoan period was not a return to the old ways; rather, it was a development out of the Chacoan adaptation with an increased reliance on different varieties of maize as well as a significant contribution from beans and turkey. The ideological realm is a different matter, and here the faunal evidence tends to better support her predictions. Both occupations at Salmon are rather impoverished in terms of exotic or ritual fauna, with the exception of macaws in the Chacoan ­period. To the extent that macaws are a proxy for an underlying trade network oriented to the south, the Chacoan period occupants of Salmon appeared to have had a larger, or at least a different, social environment than their descendants in the post-Chacoan period. In general, Irwin-Williams saw Chacoan developments as the beginnings of complex society in the northern Southwest, with a subsequent return to less complexity. We argue that the faunal data do not support this model, but instead show a change in the organization of the community between the two periods. Although the trade network that allowed residents to obtain macaws (and possibly other goods from the south; see Webster, this volume) did not continue in the post-Chacoan­ period, we do see a previously unseen intensification of food production at the pueblo. Thus, by one measure, trade, complexity at Salmon appears to decrease through time, while by another measure, food production, complexity appears to in-

110

Animal Bone from Salmon Ruins and Other Great Houses crease through time. Perhaps the better question to ask regarding Chacoan outlier communities is not how complex they were, but in what ways they were complex (Nelson 1995).

from those at other Chacoan sites (Durand 2003). Unfortunately, without better data this hypothesis cannot be tested. The faunal diversity in the post-Chacoan ­period rose dramatically at two of the great houses (Guadalupe Ruin and Pueblo Alto), but not at the others. We suggest this pattern represents a change in the level and centralization of ritual activity during this period that was not seen previously. These changes did not occur at Salmon Ruin, nor are they seen in the small sample we have analyzed from Aztec’s West Ruin. This pattern indicates that ritual practices at great houses in the central and southern San Juan Basin were changing in ways not seen in the northern San Juan Basin. From our analysis one thing seems clear: the Chaco outlier communities were not the same, ­either during the period of Chaco Canyon’s florescence or during the subsequent post-Chacoan ­period. This is not the first study to suggest that there was variability among Chacoan communities; many others have drawn similar conclusions (e.g., Durand 2003; Kantner 1996, 2003; McKenna and Toll 2001; Van Dyke 1999, 2003a). In her study of the architecture at outlier communities, Van Dyke (1999:​471) notes that “A variety of relationships probably existed between outlier communities and Chaco Canyon.” Thus, we will not gain a thorough understanding of Chacoan outliers or their relationship to Chaco Canyon through the analysis of one outlier community. Instead, a sample of many communities will be necessary. The centralization of ritual practices at some post-Chacoan great houses may represent a trend toward increased control over ritual knowledge. Many have argued that secret ritual knowledge may have been closely guarded by some residents of the ancient pueblos (Brandt 1994; Hegmon 2005; Potter and Perry 2000; Roler 1999). Ware and Blinman (2000) suggest that the secretive medicine societies are some of the oldest components of Pueblo ritual practices. It is possible that ritual knowledge was more accessible initially, but gradually became more tightly controlled. Hegmon (2005:​218) notes that in the Hohokam region, access to rituals became increasingly restricted, as seen in the eventual “construction of steep-sided

Conclusions Analysis of the faunal assemblage at Salmon Ruin has much to tell us about the economy and ritual practices of the inhabitants of this great house and the Chaco world. At Salmon Ruin and most of the other sites in our study, there was a greater reliance on artiodactyls in the Chacoan period than in the post-Chacoan period, whereas the reliance on turkeys increased dramatically in the latter ­period. This suggests a regional shift in emphasis from hunting to food production, which is also reflected in the increased contribution of beans to the postChacoan­diet at Salmon. Again, while we find evi­ dence of Irwin-Williams’s (this volume) hypothesized “adaptive response of Chacoan society,” we find the “new and improved technology for food production” not in the Chacoan period, but in the post-Chacoan period. There also were some important differences between the dietary fauna at Salmon and those at other great houses. Whereas artiodactyls decreased dramatically through time at Salmon, they decreased much less at Guadalupe Ruin, where they continued to be a large component of the diet. At Pueblo Alto the frequency of artiodactyls actually increases through time, again making up a large portion of the diet in the post-Chacoan period. With respect to diversity among the ritual fauna, there was a surprising similarity in the diversity between great houses and small houses during the Chacoan period. Samples from structures of all sizes fell within the expected level of diversity based on sample size; however, we believe that if NISP or count data were available for the fauna from Pueblo Bonito, this assemblage might be different from those included in our study. An earlier analysis based on presence/absence data (the only type of data available for Pueblo Bonito and many of the great houses in Chaco Canyon) grouped the great houses of Pueblo Bonito, Chetro Ketl, and Kin Kletso, as well as several nearby small houses, suggesting their faunal assemblages are distinct 111

Kathy Roler Durand and Stephen R . Durand

Acknowledgments

mounds and surrounding walls” as venues for ritual activity. This is in stark contrast to the earlier ballcourts, which “seem to have been designed so as to maximize open access” (Hegmon 2005:220). The increase in faunal diversity at Guadalupe Ruin and Pueblo Alto over time may reflect such a shift in access to ritual knowledge at post-Chacoan sites. The unusually diverse assemblage of avifauna found at post-Chacoan Guadalupe Ruin may indicate the presence of an individual (or individuals) at the site who had acquired knowledge of Chacoan rituals, which then became a focus of activity at this great house. The post-Chacoan sample from the associated community structure (ENM 883) shows that at that time ritual activity was restricted to the great house, whereas in the previous period there is some evidence of ritual fauna in the small house community structures. We were surprised by the lack of diversity in post-Chacoan remains at Salmon. Although we agree with Hegmon (2005:221) that there is no evidence in the Pueblo world for restricted access to the rituals themselves, our data show that in the post-Chacoan period, ritual activities came to be centralized at great houses in some communities (e.g., Guadalupe Ruin and Pueblo Alto) but not at others (such as the largely residential Salmon Ruin and possibly Aztec West). Finally, we hope that this study illustrates that faunal remains can be as useful as other types of material culture for exploring questions in prehistory. Because of the important role of feathers and pelts in the ceremonies of many Native American cultures, faunal remains can give us clues about prehistoric ritual behavior that cannot be obtained from any other artifact class. Although we have come a long way from the days in which faunal reports were included as brief lists of species in the appendices of site reports, these remains continue to be a neglected resource on many projects. As our study has shown, the fauna from each site and region may well tell us something unique about the past.

We would like to thank Paul Reed and Larry Baker for inviting us to study Salmon’s faunal remains. Paul Reed and Nancy Espinosa have been extremely helpful in tracking down the relevant field notes and photos needed for this research. Phillip Shelley has been an invaluable colleague, sharing his wealth of knowledge about all things Salmon. We are extremely grateful to Nancy Akins for sharing with us her unpublished, revised artiodactyl and lagomorph indices for various sites in Chaco Canyon, and to Virginia Butler for analyzing the fish remains for us on short notice. Last but not least, we would like to thank the numerous students at Eastern New Mexico University who helped clean and label the bones, including Beau DeBoer, Robin Gillespie, Noel Lanci, and Thomas Lloyd.

Notes 1. The frequencies of artiodactyls, lagomorphs, and turkeys at different sites are often compared using three standard indices. Each of these uses the frequency of lagomorphs in the denominator, as these taxa are ubiquitous at sites in the Southwest. At Salmon, the lagomorph NISP, as a percent of total NISP, is fairly stable between the Primary (26 percent) and Secondary (34 percent) periods. Following Szuter and Bayham (1989), but with the addition of large mammals, the artiodactyl index is calculated as: AI = (artiodactyls + large mammals) / (artiodactyls + large mammals + lagomorphs). The lagomorph index is calculated as: LI = cottontails / lagomorphs. The turkey index is: TI = (turkey + large birds) / (turkeys + large birds + lagomorphs). These indices facilitate the comparison of assemblages of differing sizes by giving the proportion of each group in comparison to the number of lagomorphs. By including the numerator values in the denominator of each equation, each of the indices ranges from 0 to 1.00, giving standardized values that are ­easier to compare between sites. Nancy Akins also is using the above formulae to calculate these indices (pers. comm., 1994). 2. The partial remains of two macaws recovered in 2005 from rodent disturbance in Room 42 have not yet been dated.

112

7

Sighting Along the Grain Differential Structural Wood Use at Salmon Ruin Thomas C. Windes and Eileen Bacha

In Chaco Canyon in the late ad 1000s, unparalleled labor efforts were devoted to building new great houses and remodeling older ones, while at the same time a number of new great houses were being built in the region north of the San Juan River. Traditionally regarded as the San Juan–Mesa Verdean cultural area, settlement in this northern region was distinct from the Chacoan society to the south. Only a few of the largest northern great houses have seen the research spade, and two of them, Salmon and Aztec ruins, have become symbols of the enigmas that drive archaeological inquiry about the development, spread, and integration of Chacoan culture beyond the interior San Juan Basin. Aztec Ruins’ three clustered great houses (Aztec East, West, and North) were built between the midto-late ad 1000s and 1200s (Brown et al. 2002, this volume; McKenna and Toll 2001:137–138; Windes and McKenna 2001). Although Aztec North remains untouched, the East and West ­ruins were partly excavated in the early days of southwestern archaeology (Morris 1928; Richert 1964). However, detailed notes that would answer many of the questions regarding the Chacoan and subsequent occupations are lacking. Aztec’s structural wood dates harvest and construction episodes through time and reveals other behavioral traits useful to understand the society that created these spectacular structures. Importantly for this chapter, our understanding of structural wood use at Aztec pro-

vides a useful model for comparison with wood data from the Salmon Ruin.

The Salmon Project Between 1972 and 1978, a large-scale excavation project at Salmon Ruin investigated the first large Chacoan great house in the Northern San Juan region since Earl Morris’s work in the 1910s and 1920s at nearby Aztec Ruins. This project, directed by Cynthia Irwin-Williams of Eastern New Mexico University, explored the nature of Chaco great house outliers and the difference between the initial Chacoan society and the subsequent later occupations that produced much San Juan–Mesa Verdean cultural material and remodeling within the same bounded architectural space (IrwinWilliams­2006a, this volume). Recent research has blurred this dichotomy between separate occupations, focusing instead on the changing nature of societies and the integration of diverse groups as regions were settled and abandoned (see P. Reed, this volume; Brown et al., this volume). Salmon saw extensive use, with massive construction in the late ad 1000s and continued use, perhaps intermittent at times, through the late ad 1200s. The poorly known intermediary occupation provides cultural continuity between the initial Chacoan occupation and the later “Mesa Verdean” one (see Cameron 2005 for a discussion of Chacoan and Mesa Verdean “cultures”); we are unsure

113

Thomas C. Windes and Eileen Bacha of the relationship of the intermediary occupation to the others (see Brown et al., this volume; see also 2002 for a similar situation at Aztec). Many architectural changes through remodeling, renovation, and ­urban renewal resulted in massive deposits of refuse and architectural remains that marked the site during its final use, creating a cultural treasure trove for research. Research problems with this long-used site and its complex fill stratigraphies were related to basic excavation logistics, staff turnover, a large, inexperienced work force, and lack of staff continuity within many room and kiva excavations—​problems common to any large site. No final, detailed excavation report was produced from the original Salmon Project although a substantive report to funding agencies provided much data (Irwin-Williams and Shelley 1980). To redress this situation, the Center for Desert Archaeology partnered with Salmon Ruins Museum in 2001 to complete a final, comprehensive report on the site, which has now been completed (Reed 2006a). This chapter is part of the larger Salmon Research Initiative developed over the last six years. This volume explores reasons for the Chacoan presence north of the San Juan River in the Middle San Juan or Totah region (McKenna and Toll 2001; Reed, this volume), an area incorporating most of the San Juan River valley in northern New Mexico and its lower northern tributaries. The core area borders the San Juan River between the modern towns of Shiprock and Bloomfield, New Mexico. Broad stretches of terrain bordering the river provide low terraces and a number of side drainages that would have been optimal places to practice horticulture. The south side of the river is scattered with small houses and occasionally larger ones that probably served as community centers (Wheelbarger, this volume). The north side, too, probably once provided acreage for housing and farming, but the spread of modern towns has made an accurate assessment of this habitation difficult (Toll, this volume). Puebloan movement into the Middle San Juan probably occurred in the ad 900s or ad 1000s, coincidental with much of the spread of Chacoan-like communities throughout the San Juan Basin. It is during the late ad 1000s that the

first planning and construction occurred at the Salmon Ruins.

Research Focus This chapter focuses on the harvesting and use of structural wood as one avenue to address the basic tenets of this volume: that Salmon was founded ­either by a migration of Chacoans from Chaco Canyon (e.g., Irwin-Williams 2006a; Judge et al. 1981:​88; Vivian 1990:​483–484), as an outpost of direct Chacoan control of a local population (Bradley 2004; Wilcox 1993, 2004), or by a local group imitating the Chacoan hierarchy as a way to concentrate power and participate in a broad regional network (Powers et al. 1983:​345; Roney 2004; Sebastian 1992:​135–136). Resolution of this problem does not automatically mean that all northern great houses derived from the same origins or were even used in the same way, but it does provide clues as to how some of the northern great houses evolved. The employment of specific technical knowledge used during the harvest and preparation of structural great house wood may help reflect group identity and symbolic systems useful for resolving questions about Salmon’s origins (see Carr and Neitzel 1995; Hegmon 1992; Lemonier 1993; Stark et al. 1998).

Woodland Expectations in the Salmon Region Understanding Puebloan behavior in regards to normal wood harvesting efforts for house construction provides a base for understanding the construction norm at Chacoan great houses. Several avenues provide insights into prehistoric tree harvesting behavior. Knowledge of conditions that affect different tree species and the nature of the local environment produces reasonable expectations for local tree stands in the Salmon area. Tree resources in the vicinities of both Salmon and Aztec ruins are similar today and may have provided the prehistoric architects and designers of the two great houses with similar materials and procurement choices. The perceived similarity of tree resources is important in light of the findings of the actual species selected for construction at the two great houses. Wood use among small-house popu-

114

Differential Wood Use at Salmon Ruin lations is critical to “reading” local tree availability and past environmental conditions. Finally, historic and modern tree stands provide information about local conditions and change, if any, over time. Across much of the Colorado Plateau, small domestic groups used wood in the near vicinity, from marginally forested lands of cottonwoods, ­piñons, and junipers. The high altitude conifers (firs and spruces) did not grow in the same areas except in isolated groves within canyons where some peoples utilized favorable farming lands. But long, straight tree limbs were not necessary for most prehistoric construction except in great houses and great kivas.

several thousand years (Adams and Bowyer 2002; Betancourt 1984; Gillespie 1985:​36). Wood ­samples recovered from prehistoric and historic sites along the San Juan, La Plata, Animas, ­Florida, and Los ­Piños rivers, in the Gobernador and Largo canyons, and in the mesa high country show an overwhelming preference for piñon and juniper (Tables 7.1, 7.2). Occasionally, ponderosa pine and Douglas fir are recovered in site locations where modern groves of these trees are present. Unless there has been a dramatic shift in forest cover, only local, native species of juniper, piñon, willow, and cottonwood would have been suitable for most prehistoric building construction. It is not surprising, then, that juniper, willow, and cottonwood are represented in the wood inventories from both Salmon and Aztec ruins. Piñon is seldom found in pueblo great house construction because of its crooked growth, lack of long straight stems, and its value for piñon nuts, pine gum, and fuel (e.g., Hovezak 1993:10; Miller and Albert 1993). ­Piñon is rare in the area today, becoming somewhat common only at about 8 km (5 miles) distance and more from Salmon (K. Adams 1980b:495, 2006a). Piñon (Pinus edulis, Colorado piñon) requires abundant summer rainfall for establishment, while Utah juniper ( J. osteosperma) flourishes in a winterwet/summer-dry climate (Neilson 1987:95, 97). If local availability had been much higher prehistorically, then we would expect piñon to have been common at Salmon. Although piñon nuts were common to both occupations, few piñon-tree parts were found at the site, suggesting that in prehistoric times it was also rare (K. Adams 1980b:494–499, this volume). Mammals and birds associated with piñon forests that might have been used for their feathers and food are also absent in the Salmon collections (see Durand and Durand, this volume). On the other hand, juniper parts were commonly used throughout the occupation for construction timbers, fuel, and many other needs (K. Adams 1980b:​493–494), indicating that it was locally plentiful. A substantial amount of firewood would have been needed each year for a large habitation site, with between 400 and 500 cords of wood consumed annually for a 200-room pueblo (Fish 1996:

Reconstructing the Tree Resources in the Salmon Region Knowledge of past and present tree resources in the general vicinity of the Salmon Ruins is germane for understanding the potential wood available to the builders of its multistoried great house. In recent years, new scientific advances have provided accurate determinations of trees sources (e.g., Durand et al. 1999; English et al. 2001; Reynolds et al. 2005), but studies utilizing these methods have not been done for Salmon. Three areas of research were used at Salmon to determine areas of timber procurement that were available to Salmon builders, based on the premise that the most desirable trees would have been those closest to the site. Information derived from excavated small-house structural timbers, historical accounts, and the locations of trees today has been used to model the prehistoric woodland setting. Large conifers were typically scarce in the Salmon and Aztec areas even before Anglo settle­ ment took place. Accounts by explorers in the ad 1700s and mid-1800s (Miller 1976; Newberry 1876:​79–81, 103, 110) and more recent observations (Morris 1919a:​7–9) supported by historic and prehistoric site investigations suggest that local tree resources were little different during the construction of Salmon and Aztec ruins; environmental conditions, which limit the extent of different tree species, have changed little in the region over the past

115

87

100 67 58 38 15 78

all 154 107

42

15 1,538

87 97 91

67 121 830

58

96

56

common 9

124

100

%

11

No.

88 224

4

— 18 7

7

8 4 83

5

— —

—

No.

Piñon

85 11

4

4

10

10 3 9

4

%

— 187

65

— 49 68

1

— — —

—

— 4

—

No.

10

58

21 37

2

25

%

Ponderosa pine

—

—

— — —

—

— — —

—

— —

—

No.

%

Douglas fir/ spruce-fir

— 8

1

— — 4

—

— see note —

—

common 3

—

No.

0.5

1

2

19

%

Populus sp.

— 11

—

— 8 —

1

2 — —

—

—

—

No.

0.5

4

2

3

%

Artemisia/ non-conif.

103 1,968

112

— 229 186

67

77 125 913

129

common 16

11

No.

3

4 6 46

7

1

1

%

5 100

6

— 12 9

Totals

In the same area, LA 89865 yielded juniper (dates at ad 650 and 664), piñon, and hackberry (Chuck Wheeler, pers. comm., 2005. Site is located north of Gobernador, east of Navajo Dam, and west of the forest.

Pump Mesa is located between Pump and Simon canyons, north of the San Juan River. Excavations yielded all juniper wood (Hefner 1985:154–155).

I ncludes wood recovered from cave sites in the mesa country north of Ditch Canyon: Cueva Grande (n = 13), Grotto Cave (14), Mesa Mt. Cave (5), Simon Canyon Cave (43), and Site H (38). Excavations by Temple Cornelius and Mr. Hoofnagle, Durango Public Museum Project (Daniels 1940; Flora and Daniels 1940–1941). Counts from the Tree-Ring Laboratory. Dates in the 900s–1200s, with eight in the 1400s and 1500s from Grotto Cave.

b

c

d

 reat kiva roofing dated in the ad 800s excavated by Hibben and Dick (1944). Counts from the Tree-Ring Laboratory. Herb Dick’s 1940 UNM field school notes at the Maxwell Museum (cat. 2005.22.53) indicate the presence G of several “cottonwood” (Populus sp.) roof timbers that were discarded.

a

Sources: San Juan River/Bolack Ranch: Linda Wheelbarger, San Juan College, pers. comm., 2005, based on Laboratory of Tree-Ring Research analyses of samples from the Tommy site and possibly the Mine Canyon site; San Juan River/Box B site: after Hogan and Sebastian (1991); Sterling site: University of Arizona Laboratory of Tree-Ring Research records; La Plata Valley, NM 170: Toll, pers. comm., 2004, and in press, and collections from Morris 41; Mouth of Largo: after Bussey et al. 1973: chronology, p. 1, LA 8662, LA 8665; Gobernador: from Hall 1944; all non-local areas except Pump Mesa and Mesa Mountains: from Dittert et al. 1961 and Eddy 1966.

Non-Local Areas San Juan and lower La Jara   and Pine rivers b Pump Mesa c Upper San Juan River Upper Bancos River,   NM/CO line Upper Piedra and   San Juan rivers, CO Mesa Mts.d Totals

Far-Local Areas La Plata River valley below   CO line, NM 170 Mouth of Largo Largo and Salt Creek a Gobernador

San Juan River/Bolack   Ranch San Juan River, Box B site Sterling site

Local Areas

Juniper

Table 7.1.  Construction Wood Recovered from Prehistoric Small Sites in the Salmon Region, ad 600s–1500s

17

—

—

0

—

—

42

—

60 693 639

25 5

1

5

14

20

— — — 100 61

Table 1),with potentially serious impacts on local tree stands. Certainly, if old-growth piñon stands had been present near Salmon, the amount of dead wood available for fuel (Floyd et al. 2003) would have made it a highly desired product and thus a frequent find during excavations.

Note: Dinétah area includes Blanco, Largo, Gobernador, Frances, La Jara, and Los Piños canyons; data from Towner 2003. Data for other three areas from Dittert et al. 1961, and includes historic and historic/prehistoric mixed.

100 1,459 1

7

1 20 —

—

4 61 —

—

79 16 1,150 228 — 3 — 7 36 71

Dinétah Area San Juan River   and lower Pine   and La Jara Upper Bancos;   NM/CO line Upper San Juan and   Piedra in CO Totals

412 161

650 42

57 18

32 14

3 6

39 3

3 1

— —

— —

17 —

1 —

— 1

— T

%

Totals

No. No. %

Artemisia Oak

No. %

Populus sp.

No. % No. % No. % No. % No. % No.

Spruce or fir Douglas fir Ponderosa pine Piñon Juniper

Table 7.2.  Construction Wood Recovered from Historic Sites in the Salmon Region

%

Differential Wood Use at Salmon Ruin

A Survey of Ponderosa Pines and Douglas Firs in the Vicinity of Salmon Ruin The primary side drainages along the San Juan River are potential source areas for the non-local timbers used to construct Salmon. Today, the nearest groves of Douglas fir are in Gobernador Canyon, 34 km (21 miles) from Salmon, and isolated ponderosa pines grow in Largo Canyon, about 59 km (35 miles) away. Rare groves of aspen are first encountered in Largo Canyon, about 72 km (45 miles) from Salmon (Figure 7.1). Stands with hundreds of ponderosa pines and Douglas firs are substantially farther away. Recent fieldwork suggests that the closest major sources are far down Largo Canyon, 77 to 90 km (48 to 57 miles) away, and northeast in the upper drainages and high country of Ditch Canyon, which drains into the Animas River at Cedar Hill, New Mexico, about 20 km north of Aztec Ruins. The high country (in the Mesa Mountains) has substantial forest pockets that may have been utilized during construction at Aztec, 37 km (22 miles) away, but it is another 17.6 km (10.6 miles) from Salmon. Overall, the largest suitable stands potentially used for construction at Aztec and Salmon are located along the Colorado border between the Animas and Los Piños rivers. Today these stands are small and dominated by Douglas fir, but the prehistoric evidence (Table 7.1) indicates much former ponderosa pine along latitudes about 8 km (5 miles) south of Durango, Colorado, a trek north of about 45 km (28 miles) from Aztec and 61 km (38 miles) from Salmon.

The Salmon Sample Field documentation of the structural wood collection at Salmon Ruin is incomplete. Since primary excavation emphasis was on the deposition histories across the site, an important goal, the features and architectural units received little attention. 117

Thomas C. Windes and Eileen Bacha

Figure 7.1.  The region around Salmon Ruin showing major drainages and canyons where the closest stands of aspen, Douglas fir, and ponderosa pine are found.

The long use of the architectural spaces over two centuries and the extensive remodeling that took place hampered the separation of the initial construction wood from later wood materials. Most rooms from which structural wood was recovered, including reused roofing, provided a variety of species that may have derived from several construction episodes. Most important, the identification of specific pieces of wood to their various functions was only cursorily noted, so that with few exceptions, the sample cannot be separated into its many component architectural uses. For the most part, structural elements in situ were not sampled during excavation (Larry Baker, pers. comm., 2005). Roofing fragments recovered from the fill account for the vast majority of ­samples recovered from the site. These fragmented structural elements created from extensive fires and natural deterioration at Salmon resulted in a mixture of samples from multiple features and time periods that lacked recording control. Intact or partly intact Chacoan roofs were rare, and only re-

covered from Rooms 37W, 58W, 118W, 119W, and 129W; a Mesa Verdean secondary roof was found in Room 60A.

The Prehistoric Sample

Nearly 1,800 prehistoric wooden elements have been documented from Salmon Ruin, with 1,667 submitted for dendrochronological analysis, including 237 recently gathered from the ruin and the Salmon Museum collections by the authors and volunteers. Another 61 samples, thought to be Populus sp., were sent to the University of Minnesota Plant Pathology Laboratory for species identification (e.g., Tennessen et al. 2002); these were not submitted for dating but are listed here. Approximately 71 percent of the tree-ring samples were burned or were charcoal fragments created by the massive fires that swept the site; another 11 percent were rotted. Of the total sample, 562 yielded treering dates. In comparison, the sample from the Aztec West Ruin (n = 4,112 as of April 2006) yielded almost 118

Differential Wood Use at Salmon Ruin

Figure 7.2.  Tree species used in the Chacoan construction of Salmon Pueblo as identified in the tree-ring sample. PP = ponderosa pine, SF = spruce-fir, DF = Douglas fir, PSF = pine/spruce/fir, POP = Populus sp., COT = cottonwood, Jun = juniper, Pnn = piñon.

no burned samples, although Earl Morris admittedly discarded a large number of burned samples during his early excavations. The degree of preservation among the different great houses is surprisingly different and may relate to different episodes of violence or cultural practices that affected each site’s final state once the inhabitants decided to leave and not return (see Schlanger and Wilshusen 1993). Salmon was burned extensively in the early ad 1100s, perhaps many times throughout the remaining ad 1100s, and then finally when the last occupants left in the late ad 1200s. Aztec West also suffered from extensive fires, but these were less widespread than at Salmon. An archaeomagnetic sample from burned Kiva H at Aztec West indicates that some of the fiery destruction was caused by protohistoric groups.1 The number of dendrochronological samples retrieved from the Salmon Ruin excavations is impressive, but caution must be exercised in light of the various conditions hampering the sample’s retrieval. Unlike the wood resources at Aztec Ruins, where thousands of wood elements are still in original contexts, the devastating fires at Salmon left much of the roofing in fragments mixed with other materials in the room fills. Roof deposits were ­neither mapped nor piece-plotted, making it dif-

ficult to trim the sample’s inevitable redundancy. The number of years to complete excavation units and the changing field personnel may also have contributed to the collection of duplicate roofing elements, which may skew the findings. In few cases did the tree-ring laboratory provide matches of pieces from the same tree stem (3 per­ cent from the 1972 sample of 420 specimens, the only year that matches were noted), but the true number of elements from the site must be much lower than the raw sample count. It was the practice of the tree-ring laboratory in the 1970s to discard samples that had very short ring series with little future scientific value that would only take up precious storage space. These were not assigned laboratory numbers, so many Salmon samples may have gone unanalyzed and uncounted.

Species

Aside from the unknown quantities not analyzed, the large tree-ring sample provides an accounting of the tree species selected for Salmon construction (Figure 7.2). Of these, 1,563 were identified to species by the University of Arizona’s Tree-Ring Laboratory, the University of Minnesota, and the senior author. The Salmon sample is dominated by two species: non-local ponderosa pine (37 percent) 119

Thomas C. Windes and Eileen Bacha samples, and another 35 unanalyzed sacks). It probably came from the mass of closing materials that covered the roof latillas. From our recent sampling, we know that both Populus sp. and ponderosa pine were favored for lintels. At Aztec West, aperture lintels were a mix of high-altitude conifers and ­aspen in the core unit, but juniper dominated the elements in subsequent construction periods. Juniper vigas were used during the later Aztec building periods, whereas only ponderosa pine and an occasional spruce-fir or Douglas fir were used during the core construction. Construction at Salmon followed similar patterns but was far less reliant on the harvesting of non-local trees.

Table 7.3.  Unanalyzed Dendrochronological Samples from the 1972 Salmon Excavations Individual specimens No. Ponderosa pine Spruce-fir Populus sp. Juniperus sp. Non-coniferous Mixed Totals

354 61 354 227 363 — 1,359

Sacks

%

No.

%

26 4 26 17 27 — 100

16 6 16 33 35 59 165

10 4 10 20 21 36 101

Note: The unanalyzed samples were returned to Salmon Ruins by the University of Arizona Laboratory of Tree-Ring Research, as noted on laboratory work sheets.

Beam Preparation and End Treatment

“Sacks” presumably represent the large paper sacks (10 lb.) delivered from the field full of fragmented and burned roofing samples that were not worth analyzing.

and local juniper (38 percent). There also is a large amount of spruce and fir (14 percent) along with lesser amounts of Douglas fir (2 percent) and Populus sp. (cottonwood and aspen: 7 percent). Almost no piñon was used for structural wood (see Figure 7.2); a meager nine pieces were recovered (0.6 percent). The relative use of these different species of trees in construction is probably accurate except for the use of Populus sp. (cottonwood and aspen). The analyzed sample does not include small, fragmented specimens or species not useful for tree-ring dating.2 If the wood returned from the tree-ring laboratory in 1972 is any indication, early Populus sp. was more prevalent than suggested from the tree-ring analyses (Table 7.3). Populus sp. represented 26 percent of the total returned specimens, and 10 percent of the sacks of returned specimens. Much of the in situ Populus sp. recently documented at the site was used as lintels, so the true impact of Populus sp. at the site undoubtedly has been underestimated. The number of returned Populus sp. in 1972 also suggests that much of the roofing was probably Populus sp. latillas. Unfortunately, we do not know how much of this may have been nonlocal aspen. Non-coniferous (probably willow [Salix sp.]) material was common among the returned tree-ring materials (27 percent of the individual

Chacoan great house construction typically reveals extraordinary efforts to flatten the beam ends (Windes and McKenna 2001: Table 2). Field and laboratory notes indicate that flattened beam ends were present at Salmon. Our recent field sample revealed a mixture of end treatments (Figure 7.3) dominated by expedient cuts rather than the flat ends common in other great houses, produced by the labor-intensive process of nibbling the ragged ax-cuts. Most of the recent sample of end treatments came from Room 62, where an abundance of small tie-pole ends were evident (Figure 7.4). These were cut in ad 1116 and 1118, long after the initial room construction, and may not be representative of earlier workmanship. A few other beam-end cuts were studied in the collections at the University of Arizona Laboratory of Tree-Ring Research. Although the end-treatment sample is spotty, it does indicate that the Chacoan practice of removing all vestiges of the stone ax-cuts on the beam ends was used at Salmon. The presence of bark, however, suggests preparation differences between Salmon and Aztec. Whereas bark is almost never found on Chacoan elements at the Aztec West Ruin nor in Chaco Canyon great houses, the Chacoan sample at Salmon yielded bark on almost every dated juniper ­sample (98 percent; 49 of 50). This suggests less careful beam preparation (lack of bark removal) at Salmon than at other great houses.

120

Differential Wood Use at Salmon Ruin

Figure 7.3.  Beam end treatments recorded during recent fieldwork at Salmon Ruin. Majority are from Room 62.

Total Structural Elements Used at Salmon

how much new tree-cutting took place. Ideally, if all new wood was required, about 2,900 to 5,600 more trees, predominantly juniper, would have been cut after about ad 1120. The short stems of juniper trees would have provided a maximum of a single viga or latilla each, but the tall, straight conifers growing in the higher, more moist environments probably furnished two or three elements per stem (see also Wilcox and Shenk 1977:85). One tree could have provided an estimated four or more lintels. Aztec West builders used an average of about 8–10 lintels per door and 10–14 lintels per ventilator, with the highest numbers set in the thick outer walls of the structure. Based on estimates from Chaco Canyon’s Pueblo del Arroyo (Windes et al. 1994), approximately 400 to 1,000 juniper trees would have been needed for roof closing material if only juniper splints were used. We know that in at least some instances at Salmon, an additional layer of tightly packed small poles covered the latillas, which were then covered by a layer of branches and twigs before the final application of a thick layer of adobe (Figure 7.5). If the 38 percent presence of juniper, as indicated by the tree-ring sample, is relatively accurate, and if much of the closing material were juniper and willow, then the impact to the ecology of removing thousands of local trees would have been severe. The

Estimating the number of elements and the potential number of trees required for construction at Salmon is fraught with difficulties but provides general baseline figures that help to determine the magnitude of the tree-harvesting effort. Although the number of rooms for much of Salmon was determined through excavation and wall clearing (see P. Reed 2002, 2006c, this volume), the west wing was a difficult area that saw much remodeling and reconstruction during the McElmo and Mesa Verde phases.3 Thus, Chacoan room counts must be extrapolated from the patterns known for the rest of the site. Beam sockets and general Chacoan building practices provide reasonable estimates for the different types of structural elements required for the building (Table 7.4; see Windes and McKenna 2001). We estimate that the initial Chacoan building required between 7,500 and 9,500 trees to provide the 15,000 to 17,000 elements needed primarily for roofing and wall apertures. The San Juan (McElmo and Mesa Verde phases) occupation resulted in many small kivas and subdivided rooms that required far less timber; the samples from Salmon suggest that much of this was re­cycled Chacoan wood. Without dates for most of the secondary elements, however, it is impossible to know

121

Figure 7.4.  Profiles of the west and east walls of Room 62 showing the roof vigas, their supporting posts, and the two retaining walls that buttress the east wall. Note the numerous tie poles, which indicate additions made in ad 1118. Original maps by J. Schubert, G. Jordan, and E. Bacha (2002).

Table 7.4.  Estimated Construction Elements and Total Trees Needed for the Chacoan Greathouse and San Juan (Secondary) Constructions at Salmon Numbers per Unit Vigas

Latillas

Door lintelsa

Vent lintelsa

Splintsb ( Juniper Posts Misc. trees)

Numbers Required per Room 6–8

26–50

8–10

10–14

Chacoan/Primary Occupation Rooms (type) No.c Large, sq. (1) 19 152 950 152–190 380–532 Long, narrow (2) 162 972 4,860 1,296–1,620 3,240–4,536 Large rect. (3) 6 36 180 48–60 0 6 48 360 48–60± 0 Gallery (4) e 104 520 216–270 0 Misc., small (5, 14) f 52 Kivas 2 500± 40± 0 0 Large, cribbed b 1 30 643 40 0 Great kiva h 1,842 7,553 1,800–2,240 3,620–5,068 Total elements j 1,250 5,212 — — Adjusted)2 j 1,062 4,431 — — Adjusted)3 j — — 450–560 905–1,267 Adjusted)4 j Total Trees 1,062–1,842 4,431–7,553 450–560 905–1,267 San Juan/Secondary Occupation Rooms k Subdivided rooms 35 35–70 525–2,100 Kivas k Small, non-cribbed 8 16–48 120–240 Medium, cribbed 12 1,200–1,800 180–360 1 30 643 Great kiva h Total elements 1,281–1,948 1,468–3,343 Total Trees 1,281–1,948 1,468–3,343 a

0–4

Totals (365–938)

48 d ? ? ? ?

? ? ? ? 100g

— — 48

40 i (9–25) 580 — (14–35) 713 140 — 15,003–16,891

48

140

388–998

1,682–1,872 10,368–11,988 264–276 456–468 940–994

?

0

0

?

(57–146)

0 0 40 40 10

112 168 0 280± 70±

0 6 0? 6 6

? ? ? ? ?

(5–13) (16–41) (14–35) — 92–235

7,424–9,475

560–2,170 248–400 1,554–2,334 713 3,075–5,617 2,927–5,612

Calculated for a single door and two ventilators per room.

b

Total room roof area × 5 cm ) .236 – .606 m3 = number of juniper trees (after Windes et al. 1994). Based on work at Pueblo del Arroyo (Windes et al. 1994), a single one–seed juniper tree yields between .236 and .606 m3 of splints. Other species of juniper might yield more. Salmon roof areas based on 4,426 m2 for Chacoan rooms, 58 m2 each for elevated kiva and unnumbered kiva in NW corner, 165 m2 for great kiva, 691 m2 for San Juan rooms, and 259 m2 for San Juan kivas (included 2 unnumbered in west wing). c

Numbers of room types include estimated second and third stories.

d

Posts are difficult to estimate without detailed notes about them. I assume that the large square living rooms had roof support posts based on the notes from Room 56 (4 per room, not counting post steps).

e

Gallery–like rooms (93, 102, 128, 142, 147, and 158). Estimates based on paired sets spaced equally along room with 15 latillas resting on each pair.

f

Small miscellaneous rooms and irregular-shaped-interstitial spaces estimated for an average of 2 vigas each and 10 latillas per room.

g

This number approximates only those small poles photographed for Room 181 (see Figure 7.5). There may be far more rooms similar to this one.

h

Estimate from Morris’s (1921) field notes and his rebuilt great kiva. Includes antechamber but not extramural rooms.

i

Later support beams set in recess around elevated kiva (Room 64) walls.

j

Adjusted for two and three non-juniper vigas and latillas cut from a single tree. Four lintels are calculated for each tree cut. Assume one viga or latilla per juniper tree.

k

Based on number of crib logs recorded at Spruce Tree House, six-pilaster Kivas C and D (102 and 126 logs, respectively) and at 42SA6651, Kiva A (4 pilasters: 100 crib logs, 32 flat roof–top logs) in Natural Bridges. Street (2001:156) estimates 200 logs for flat-roof Kiva I (non-cribbed), with four large recesses, at Long House. For this exercise, assume 15–30 roof-top logs (“latillas”) for small San Juan rooms and tops of kivas, 100 crib logs for a small kiva (4 pilasters) and 150 crib logs for a medium-size six-pilaster kiva. Crib logs are listed under the viga column. Small, non-cribbed kivas yield 2–6 vigas.

Thomas C. Windes and Eileen Bacha

Figure 7.5.  Collapsed roofing in Room 118. This unusually complex roof had small poles packed together over the vigas and covered by a ­layer of twigs and branches, and then adobe. Photograph by Peter George.

continued occupation of the site and the need for firewood and wood for other uses would have impacted the local forest cover to an alarming degree (see Kohler et al. 1984; Kohler and Matthews 1988; Samuels and Betancourt 1982). During the secondary occupation, builders scavenged logs from the Chacoan structures when rooms were subdivided and when smaller rooms were built inside the large Chacoan rooms. Because so little juniper dated from the secondary occupation, it is difficult to assess how much new wood was procured as an alternative to reusable timbers

without correlating the excavation data with the wood samples; however, the high number of juniper elements associated with the secondary constructions suggests that many late roofs were made from freshly cut local trees, and that long-distance procurement was rare.

The Dated Sample from Salmon The success rate for dating wooden elements from Salmon Ruin is 34 percent (n = 559), generally on par with other southwestern sites, even though Populus sp., willow, spruce-fir, and one-seed juni-

124

Differential Wood Use at Salmon Ruin

Figure 7.6.  Tree-ring dates from Salmon Ruin.

per provide few, if any, dates. Paul Reed (2006b) covers in detail the relationship of the tree-ringdated sample to construction events at the site. Based on architectural information and associated tree-ring dates, he has defined six separate building episodes that created the overall final design of the great house. Several of these episodes yielded few, if any, tree-ring specimens. In addition, much remodeling took place during the secondary occupations. A brief overview of these tree-ring results is germane for this chapter (Figure 7.6).4 Only a few cutting dates prior to the massive harvesting in ad 1088 were obtained from the sample, and these were at first dismissed as reused pieces gathered from earlier small-house structures in the area (Adams 1980:​215). Six early cutting or near-cutting dates for ponderosa pine and Douglas fir clustered in the east wing of the site ranged in age between ad 1068 and 1072; this suggests a singular episode of construction using non-local wood. Baker (this volume) argues that these samples and their associated rooms represent an early roomblock at the site before the overall main complex was built. Small, initial roomblocks incorporated within the larger, later great house format are not unusual and have been noted in several of the early great houses in and around Chaco Canyon (Lekson 1984; Windes

2004; Windes and Ford 1996). The placement of this potential early block of rooms in and around Room 122 at Salmon, however, is unusual because it is located within the east wing rather than the central core of the great house, where it exists in all other known cases. It is also unusual that the east wing location would have required the initial roomblock to be oriented east, given that nearly all Puebloan domiciles are set to face somewhere along an arc of east-to-south to take advantage of the sun. The eastern orientation, of course, is predicated on the initial use of the early building for habitation. Later, when the east wing was expanded and incorporated within the larger structure, it was oriented to face west into the interior plaza. More likely, this small roomblock was razed or much modified, and the timbers salvaged once Salmon was started; a similar small, early, razed roomblock was found ­under Pueblo Alto (Windes 1987a, 1987b).

Central Roomblock: The West and East Sections

West and east of the elevated kiva (Room 64) were many deep, two-story rooms and several with a third story. Many of these were excavated. The complex depositional histories in these rooms made it difficult to separate the probable different-story roof elements recovered from them. For the most part, 125

Figure 7.7.  Map of Salmon Pueblo showing probable third-story room additions (shaded) in ad 1094 and 1095, based on tree-ring results.

Differential Wood Use at Salmon Ruin timbers for the first and second stories were cut between ad 1088 and 1089 to roof the basic rooms of the Chacoan great house. Small clusters of dates between ad 1092 and 1094 mark a second round of harvesting that may mark new additions. These ­samples are mixed with ad 1088–1089 samples in some rooms. Plotting the newer dates by room revealed a curious clustering along a single row of rooms on both sides of the elevated kiva, one tier in from the back rooms. The late samples were found in a string of rooms (30-31-33-36) west of the elevated kiva and in an equivalent row (55-79-90-84) to the east with one exception, Room 79, which failed to produce any dated samples. In addition, Room 43, in a similar position to the west as Room 79 but along the outermost tier, produced numerous ad 1094 dates (Figure 7.7). These late dates probably reflect the construction of two blocks of third-story rooms on both sides of the elevated kiva (see P. Reed 2006b), but the context of the late dates may be misleading. If a single tier of thirdstory rooms had been built along the very back of the pueblo, it is possible that after abandonment these would have collapsed into the rooms directly in front. The dates may also be the result of sampling biases, perhaps marking a larger block of three-story rooms along the back tiers of the central roomblock. ad 1092 to 1094 dates, recovered from beams in Room 121 in the southeast roomblock and in front of the elevated kiva in Room 60A, mark other architectural modifications, although these samples were probably elements reused for repairs.

Rooms Not all of the rooms can be discussed in detail, but the sample examined here spotlights the trends observed in wood used for building throughout the roomblocks. The primary great house construction centered on the elevated kiva and then extended east and west. The core unit is of interest as the central place and because of its importance to the overall plan of the structure. How wood was selected and treated for use in construction in this block of rooms may reveal conformance (or not) to a generalized Chacoan model of great house construction. Certainly some units within the core, such

as the elevated kiva (see below), reflect maximum efforts in tree selection and harvesting. As a unit, however, the effort appears variable, with the use of both local and far-distant trees. Overall, the three blocks comprising the core unit—the elevated kiva and its attendant rooms (Construction Episode 1) and the rooms around and below the elevated kiva (Episodes 2 and 3)—are dominated by ponderosa pine (non-local), followed by juniper (local) and fir (non-local) (Table 7.5). In contrast to the core unit at the Salmon Ruin, almost no non-local wood was used during the initial constructions at Aztec West. When Aztec wood was first harvested at about ad 1098 or shortly thereafter, Salmon was barely in its tenth year of existence. The large elevated kiva (Kiva L) at Aztec and a few attendant rooms were built first, and over the next decade the overall footprint of the great house was prepared and foundations were poured (Brown et al., this volume). The large sample from the Aztec West core unit (n = 1,066 samples) reveals that non-local conifers (ponderosa pine, spruce-fir, and Douglas fir) made up 72 percent of the wood materials (roofing and lintels); another 25 percent came from Populus sp. (see Windes and Bacha 2006: Table 52.11). Given the widespread presence of cottonwood along the nearby Animas River, it could be erroneously assumed that the Aztec Populus sp. was mostly cottonwood. Chaco builders chose instead to carry to the site thousands of high-altitude, non-local aspen, a species now present in large numbers at least 58 km (36 miles) from Aztec; 83 percent of the core unit Populus sp. was aspen (n = 198). Nearly 800 Populus sp. specimens from the site examined for subspecies (see Tennessen et al. 2002) confirm the overwhelming preference for aspen over cottonwood in the construction (mostly as roof latillas). Salmon yielded few samples that could be classified as one of the two species (most were too rotted), with cottonwood slightly favored over aspen. Many more samples were curated but not analyzed, although those returned from the tree-ring laboratory could not be relocated in the Salmon Ruins Museum collection. Much of the juniper sample from the core unit at Salmon derived from the large set of tie poles

127

207 101 168 476 117

Aztec West Kiva and suite East suites West suites Totals Totals w/o lintels c

21 39 35 95 41

25 33 31 89

SF

9 29 27 65 15

— 2 1 3

DF

12 75 66 153 43

— 2 — 2

PSF a

249 244 296 789 216

72 81 120 273

All

66 72 75 71 65

88 42 69 61

%

9 27 33 69 18

— 15 2 17

Pop. b

15 10 9 34 2

— 1 — 1

Cot.

101 34 29 164 97

— — — —

Aspen

125 71 71 267 117

— 16 2 18

All

33 21 18 24 35

8 1 4

%

Local Woodlands

1 22 25 48 1

10 94 51 155

Jun.

Field calls; did not distinguish between ponderosa pine, Douglas fir, and spruce-fir.

Not identified as to aspen or cottonwood.

Aztec lintels removed from totals because majority of Salmon samples presumably do not include lintels.

a

b

c

2 1 3 6 —

— 2 — 2

Pnn

3 23 28 54 1

10 96 51 157

All

1 7 7 5 T

12 50 29 35

%

Low-Altitude Conifers

Note: Cot. = cottonwood; DF = Douglas fir; Jun. = juniper; Pnn = piñon; PP = ponderosa pine; Pop. = Populus sp.; PSF = ponderosa pine/spruce-fir; SF = spruce or fir.

47 44 88 179

Salmon Kiva and suite East suites West suites Totals

PP

High- and Low-Altitude Populus sp.

High-Altitude Conifers

Mixed

NonMixed Local local

Non-local Woodlands

Table 7.5.  Use of Tree Species in the Architectural Core Units at Salmon and Aztec West

377 338 395 1,110 334

82 193 173 448

Total

34 30 36 100 100

18 43 39 100

%

Differential Wood Use at Salmon Ruin added to a new support wall in Room 62 in ad 1118. With this sample removed from the totals, the emphasis upon non-local roofing elements is more pronounced. Room 62 must have suffered from structural failure caused by the weight of the elevated kiva after its initial construction in ad 1089. Repairs in Room 62 were initiated in the fall or winter of ad 1116 but not completed until ad 1118 or later, at which time the room may have been used for refuse deposits to help alleviate the structural forces imposed by the elevated kiva’s architectural mass. The deliberate in-filling of rooms to help distribute the weight of adjacent architecture was also noted at Aztec West (Brown et al., this volume). The reuse of small poles and probable reused vigas for upright posts to support a sagging roof suggest that part of the great house roofing elsewhere in the site was dismantled by ad 1118. A single cutting date of ad 1257v from a viga-sized Douglas fir element marks some late use in the room. Room 62 at Salmon yields the latest dendrochronological evidence, dating to the mid-1200s, but many other lines of archaeological evidence indicate that remodeling and new construction was ongoing throughout the ad 1100s, similar to efforts at the Aztec Ruins (Brown et al., this volume). However, these later events did not match the intensity and short timing of the initial great house construction. Although dendrochronological evidence from Salmon is lacking for most of the ad 1100s and early 1200s, there was widespread scavenging and reuse of the initial Chacoan timbers as well as local tree harvesting for continued building episodes. The latter wood failed to produce any dates because of the nature of the erratic tree growth, the local species (juniper and Populus sp.) selected for building, and the probable use of small poles with few growth rings. The suite of door-connected rooms adjacent to the elevated kiva on the east side (Rooms 51, 57, 81, 82) yielded 84 analyzed tree-ring samples dominated by ponderosa pine (51 percent), spruce-fir (31 percent), and juniper (13 percent), a small handful of Douglas fir (1) and Populus sp. (2), and a rare oak specimen. The few dated samples (25 percent) were harvested between ad 1088 and 1090, indicating that the room construction was completed by

ad 1090 or slightly later. Otherwise, there is little to add to our knowledge of wood use in this series of rooms, which is similar to others along the central main roomblocks.

West and East Wings

The lack of excavation in the west wing prevented a thorough look at the initial room constructions amidst much secondary rebuilding, but Room 129 in the east wing produced the largest wood sample from Salmon Ruin. Field notes indicate that 138 dendrochronological samples were collected during the two years of excavation (1974 and 1975), but only 64 samples were analyzed; it’s possible the remainder were discarded at the tree-ring laboratory because they had too few rings and were considered of little research value. Room 129, which burned early in the site occupation and was then used for trash deposition, is unique at the site for the number of small Chacoan and Mesa Verdean refuse piles in it. Tree-ring dates indicate tree-cutting for a burned roof took place between ad 1088 and 1090 or 1091, and it is the only roof at the site composed strictly of ponderosa pine. A single date at ad 1106 indicates some later activity that may not have been roof-related. The room’s location in the middle of the east wing and its association with the earliest site construction suggest its special place within the overall site, reflected in part by the selection of all pine for its roofing. Although many textiles were recovered from the roof, they have been assigned to the primary and secondary occupations (Webster, this volume). Perhaps the refuse piled in the room was placed specifically in this room rather than in more secular contexts. An equinox alignment passing through the room and the elevated kiva (Baker and Mantonya 2002) may have had some significance for the use of the room and for the later deposition of cultural materials.

Living Rooms

The largest rectangular rooms at the pueblo were those used for habitation by Chacoans (see Reed, Chapter 3, this volume). A series of contiguous living rooms was built to the east and west of the ­elevated kiva, but only those along the eastern row were excavated. These rooms yielded few roofing

129

Thomas C. Windes and Eileen Bacha specimens even though large quantities of wood would have been necessary to roof them. Large ­conifers (about 6 m long) would have been needed to span the 5.6 × 5.6 m rooms; it is unlikely that juniper trees would have been long enough. Such large timbers could only have been obtained at some distance from the pueblo. Notes from Room 56 indicate that roof support posts and post steps that aided egress through the doorways were used, a common characteristic of the large, early rooms in Pueblo Bonito and Pueblo Alto ( Judd 1964; Windes 1987a). The foundations for a possible elevated kiva, similar to Room 64, were also found under several rooms (41, 56, and 57). Other large living rooms were excavated in the east and west wings. Many of the living rooms were later modified into kivas, a pattern also observed at Pueblo Alto (Windes 1987a).

Kivas Room 64 (Central Elevated Kiva)

The architectural core of Salmon is dominated by the elevated kiva, Room 64, and its attendant rooms. Although this area may not have been the initial construction at the site (see Baker, this volume), it certainly was planned in the final layout as seen today. Known as the Tower Kiva, it is in the center of the great house and is the focal point of the site. A possible second large, elevated kiva offset to the east (see description of Room 56 above) was either never completed (only foundations were found) or was razed. This may have been the initial elevated central kiva, or may have been intended to be the paired twin to Room 64. The few timbers that dated from Room 64 indicate that primary harvesting for the roof took place between ad 1088 and 1090, which suggests that preparations and construction took several years. Room 64’s samples are the most unusual because of the predominance of spruce-fir. Given its size, elevated position, and massive construction effort, Room 64 is probably the most important enclosed space at the site, followed by the Great Kiva. Aside from its massive masonry foundation and buttressing that allowed the large, circular structure to be elevated to a second-story level, the roof was built entirely of rare wood ­materials

seldom seen in great house construction. Nowhere else do we have evidence of a single architectural unit employing timbers mostly of high-altitude spruce-fir. Even Chimney Rock Pueblo and its attendant small houses, built at 2,439 m (8,000 ft) elevation in the heart of spruce-fir country to the north in the Colorado mountains, failed to yield complete architectural units of this wood species; fir comprised only 10 percent of the total from the great house, and none from the small houses (Eddy 1977: Table 29). It is not just the architecture that distinguishes Room 64 from the surrounding roomblocks. The contents of the adjacent rooms were full of textiles and bird bones (Durand and Durand, this volume; Webster, this volume), signifying the cultural significance of the center core. Even without the unusual artifacts and faunal remains, the construction elements alone make this area distinct and indicate that it was specially planned from the beginning.

Rooms 130 and 151: The Great Kiva and Its Antechamber

The Great Kiva offers a number of insights into construction behavior at Salmon. Although it was built as part of the overall Chacoan construction effort in the late ad 1000s, substantial roofing remains were preserved by fire. They revealed that the structure had been rebuilt or reroofed at about ad 1263. The species used in the kiva bore the unmistakable imprint of the late occupants, who relied heavily on juniper trees. The scattering of Douglas fir, spruce, and fir in the roofing remains probably reflects fragments of the large primaries on which the weight of the roofing rested. The recovered juniper elements must have derived from the many roofing secondaries and leaner poles needed to complete the roof. The small number of spruce-firs (n = 13) and Douglas firs (n = 11) were cut in the same early ad 1200s period associated with the juniper roofing ­elements, suggesting use of a common harvest area in the higher elevations where north-facing slopes contained a mixture of juniper, spruce-fir, and Douglas fir. The area most suitable for this combination of species is in the Navajo Reservoir district, about 42 km (26 miles) to the east of Salmon, fol-

130

Differential Wood Use at Salmon Ruin lowing the San Juan or Los Piños rivers upstream to Colorado. Other single, late-construction events cannot be distinguished in the excavation record without, perhaps, time-consuming analysis of the extensive field notes. But given the paucity of post-Chacoan juniper dates from the site, it is presumed that the juniper for late room and small kiva construction, unlike that for the Great Kiva, was obtained from a different, most likely local, environment. For instance, juniper (n = 79) was the majority species in six of the nine late, small Mesa Verde style kivas, yet only a single sample (1 percent) dated. In contrast, juniper samples procured from kivas at Mesa Verde have an extremely high dating success rate because growing conditions were more favorable. Besides the Great Kiva, only a handful of ad 1200s samples from three rooms dated, but these reveal that juniper and Douglas fir were harvested together, again suggesting at least some long-range procurement in this time period. The initial Chacoan great kiva may have been burned (Adams 1980:241). The frequency of this common practice among southwestern great kivas suggests that they were burned as part of a standardized ritual terminating their final use. The timing of its first decommissioned status would be informative in determining the possible end of Chacoan political involvement in the area. Despite the numerous Chacoan timbers at the site, only newly cut timbers were used to rebuild the Great Kiva in the ad 1200s. Perhaps house wood was owned by the last residents living at the site and was not available for recycling. It is also possible that the use of new, sanctified non-local timbers was required for a structure of this importance, which was not to be contaminated by older recycled wood.

assemblages dominated by juniper are suspected to be late-cut wood, although few of these elements dated. Overall, most kivas yielded a mixture of species and tree-ring dates, indicating the reuse of Chacoan timbers.

Wood Harvests Most timbers obtained for constructing Salmon were cut during the growing season, but the mix of species cut during both the growing and dormant seasons helps to pinpoint the harvest strategies employed by construction crews. Based on the many buildings that have produced tree-ring dates, smallhouse construction normally required a single harvesting year to procure timbers for construction. Large construction efforts, however, required considerable planning, organization, and logistics to acquire the needed materials. As a group, the Salmon cutting dates do not indicate the same intensity of harvesting seen at the Aztec West Ruin, built two decades later. Instead, the Salmon dates suggest a minimal harvest effort that reached its maximum in a single year, ad 1089, and then declined. Of course, the sample contains many undated samples that might present a different interpretation if all could be dated. Assessing outer-ring growth for the dated ponderosa pine samples provides two interpretations. Those with incomplete outer rings were cut during the growing season in the year dated, but it is those that were cut during the dormant season that provide evidence of a switch from a fall to a spring harvest. Of course, these may have been cut during the winter, separate from other harvests, but the combination of both incomplete and complete outer rings in the same year suggests fall or spring harvesting when both types of outer ring growth occur together. It could be argued that harvesting took place each fall, when both incomplete and complete outer rings occur in the same year. Alternately, harvesting could have taken place in the spring of each following year, when dates with complete outer rings of the previous year are matched with incomplete rings of the next year. Thus the largest groups would date between ad 1088 and 1090, with ­major harvesting in the spring of each year (Table 7.6). For instance, the samples dated to ad 1089 could

Other Kivas

The remaining kivas at Salmon were small and built late. Most were constructed in existing square or rectangular rooms, but others were built as part of new additions added to the south ends of the wings, and some were built in the plaza, probably as units associated with late habitation suites nearby. Given the preference for juniper in structures built at Salmon and Aztec in the ad 1200s, 131

8

21

5

6

1

1

n=9

n = 44

n = 10

n=8

n = 23

n=6

n = 56

1093

1094

1105

1106

1116

1262

1263

8

3

30

110 3

3

SF C I C I C I C I C I C I C I C I C I C I C I C I C

Outer Ring a

1

1

3 2

1

3

1

DF

8

1 3

1

Fir

16 4 1 24 28

1 1 1 3 8 12

1 1 22 22 3

Jun

3

Pop

2

Pnn

4 3 4 6 7

1

1

3 2 7 19 74 29 27 1

PP

2 1

SF spr/sum/fall 1087 fall 1087-spr 1088 spr/sum/fall 1088 fall 1088-spr 1089 spr/sum/fall 1089 fall 1089-spr 1090 spr/sum/fall 1090 fall 1090-spr 1091 spr/sum/fall 1091 fall 1091-spr 1092 spr/sum/fall 1092 fall 1092-spr 1093 spr/sum/fall 1093 fall 1093-spr 1094 spr/sum/fall 1094 fall 1094-spr 1095 spr/sum/fall 1105 fall 1105-spr 1106 spr/sum/fall 1106 fall 1106-spr 1107 spr/sum/fall 1116 fall 1116-spr 1117 spr/sum/fall 1262 fall 1262-spr 1263 spr/sum/fall 1263 fall 1263-spr 1264

Calendar Date

local

local

local

non-local?

non-local?

far local

local, far local

far local

?

non-local

local, non-local

non-local

non-local

Procurement Area

a

I = incomplete outer ring, tree cut during the growing season; C = complete outer ring, tree cut during the dormant season.

Note: DF = Douglas fir; Jun = juniper; Pop = Populus species; Pnn = piñon; PP = pondersoa pine; SF = spruce or fir; SAL-160, 219, 302, 303, 306, 314, 428, 477, 836, 849, 853, 888, 967, 1022, 1058, 1107, 1109, 1119, and 1183 were removed from analysis because of uncertainty regarding the end tree-ring date (vv).

55

21

5

2

2

10

1

n=5

1092

3

1

1

2

n=2

1091

8

2

3

3

n = 42

1090

49

1

n = 163

32

1089

2

7

PP

3

1

Pnn

n = 38

Pop

1088

Jun

2

Fir

n=9

DF

1087

YEAR AD

Table 7.6.  Tree Harvest Periods by Species and Selected Years

Differential Wood Use at Salmon Ruin all have been cut in the fall of 1089, or the 19 complete outer-ring samples from 1088 could have been harvested with the 79 incomplete outer-ring ­samples of 1089 during a spring 1089 cutting. There is no way to ascertain which scenario is more likely, but other great house samples suggest that spring was the likely harvest period (Windes and McKenna 2001). It is clear, however, that ­major tree harvesting for Salmon was completed in the spring or summer of ad 1090. The juniper harvests present a different picture. The growing season for local juniper begins somewhat earlier than ponderosa pine because of the warmer temperatures. The equal numbers of 1089 dated juniper with incomplete and complete outer rings suggest a single harvest episode when juniper had almost finished its fall growth. In the Salmon area, that might have been in September, depending on the year. If juniper was cut at the same time as the ponderosa pine in the spring, we would expect that the delayed growth of ponderosa pine would yield greater numbers of trees still in the dormant stage. Considering either alternative discussed above, the great majority of pine was cut during the growing season—the opposite expectation when compared to juniper. Thus, two separate harvest sessions are suggested for the ad 1089 collection of these two tree species: one local harvest in the fall (juniper), and one non-local harvest in the spring (pine). The larger junipers destined for primary roof supports may have needed time for curing that was not necessary for the smaller elements (Lange 1959:146; Syngg and Windes 1998; Windes and McKenna 2001:125).

beyond (non-local). The northern locations provide different logistical requirements for travel and effort in acquiring the species of trees selected for construction. Two species provided the bulk of the structural wood requirements for the Salmon builders: juniper and ponderosa pine, a local and non-local species, respectively. Three construction phases show selection of one species dominant over the ­others: the elevated central kiva and its adjoining interstitial rooms, the Great Kiva, and the southeast roomblock (Construction Episode 11). The special significance of the elevated kiva probably accounts for the use of non-local wood that could be obtained only from the greatest distance of any ­species in our sample. Extra efforts were also made to procure wood at great distances for construction of the Chaco Canyon great houses (Dean and Warren 1983; English et al. 2001; Windes and Ford 1996; Windes and McKenna 2001), but the scarcity of suitable trees for great house construction in Chaco required the use of non-local resources. Salmon and Aztec ­ruins are located where the bulk of the needed wood resources could have been procured locally. Both were situated among similar locales of tree species, juniper and cottonwood, that were used in the great house construction. Plans by the builders to select specific species for structural timbers provide insights into the background and experiences of the building supervisors. For the Aztec West Ruin, builders of the central core rooms took the most arduous route for procuring wood. Ponderosa pine, spruce-fir, and aspen were widely used (­Table 7.5), but as construction extended out from the core units, the builders switched to local resources and an increased use of juniper (Brown et al., this volume). Special preparation of the structural elements and the use of non-local species indicate that the builders knew from prior experience the construction techniques employed in Chaco Canyon. Unfortunately, wood treatment observations for the Salmon sample are mostly lacking, but species use helps to compare the trends between the two great houses. The central core units at Salmon also reveal special treatment marked by the extensive use of

Local vs. Non-Local Procurement Information gathered from prehistoric and historic sites in the area (Tables 7.1 and 7.2) and general knowledge of tree species’ locations today suggest that juniper and cottonwood would have been locally available to the builders of Salmon. ­Piñon, along with isolated stands of ponderosa pine and Douglas fir, were farther away but within a day’s journey. Large numbers of young ponderosa pines, firs, spruces, and aspens are found only in the higher, cooler elevations along the northern drainages and mesa tops near the Colorado border and 133

Thomas C. Windes and Eileen Bacha non-local woods in construction (see Table 7.5), although only about 61 to 65 percent was non-local compared to 96 percent at the Aztec West core unit. How various species of wood were used in great house construction and the special treatment of the individual elements would not have been common knowledge to outsiders but would have been limited to those allowed access to the great house interior. The end treatment of many elements (see Windes and McKenna 2001: Figure 4) would have been hidden once placed in the walls; these techniques required specialized knowledge likely limited to the builders and their supervisors. Both central core units at Salmon and Aztec West reflect a great deal of labor in the selection of structural wood, and we know that this extra effort was extended to the treatment of individual elements at Aztec West. In both cases, specialists with required construction knowledge appear to have directed the projects. These specialists must have received their training and experience in Chaco Canyon in slightly earlier structures because the techniques and use of wood replicates those of Chaco great house projects. The massive building efforts in Chaco (Lekson 1984) in the late ad 1000s would have required the most laborers and building supervisors of any period, but they would have been available for new construction projects just as planning for Salmon was underway. We expect that these specialists would have moved onto other projects, much like the master builder groups that moved from project to project during the massive, ongoing public building projects (mainly cathedrals) in medieval Europe ( James 1982:​49–​63, 133–​138; Rodwell 1981:127). Despite the similarities between Salmon and Aztec West, the selection of different tree species suggests that different leaders carried out the expected model of the two Chacoan great house constructions in slightly differ­ent ways.

Symbolic Use of Wood Identifying the tree species used in construction helps determine potential harvest sources for much of the wood procured for building Salmon and many other great houses, and also provide detailed cultural preferences as indicated by the dif-

ferent groups involved in the varied aspects of construction. Van Dyke (2004) has argued that during the construction of McElmo style Chacoan great houses, which were built slightly later than Salmon, the horizontal dimensions and verticality within the cosmology of the Puebloan six sacred directions and the center place were reflected in the architectural design of the buildings in Chaco Canyon (see also Rapoport 1969:76). The rebirth and reformatting of Chacoan cosmology was an effort, according to Van Dyke (2004:426), to reestablish power and legitimacy among leaders and to connect with the glories of the past. At Salmon, this may be reflected in the duality of the high, centrally located, large, elevated kiva (Room 64) along the north-south alignment, with the lower great kiva (Room 130) in the plaza. Aside from the positioning at the site, the use of specific wood species in the construction may also have enhanced the tie to the sacred and the profane. Among historic Puebloans certain mountains and trees represented the cardinal directions and had ritual significance (Douglass 1917:​345–346, 358, 362–365; Jones 1931:41; Robbins et al. 1916:42–43; Stevenson 1894:28, 124, and 1904:​516; Tyler 1979:253; White 1962:110–111, 309; Whiting 1939:63). The unique use of fir from cool, wet, and high locations for the elevated kiva roofing reflects the symbolism of the north, the mountains, and life-giving water. We do not know which wood species were utilized for the Chacoan era great kiva superstructure (although ponderosa pine is suspected), but its later reconstruction using Douglas fir and juniper timbers from the warm, dry, lower elevations may symbolize the mesa-andplain country to the south in balance and symmetry with the central elevated kiva. A similar architectural plan was devised at Aztec West, where a centrally located elevated kiva (Kiva L), with pilaster logs of ponderosa pine, and a great kiva in the lower plaza were built slightly off a north-south alignment. The use of ponderosa pine for pilaster logs was unusual, although not unique, because juniper was usually favored for the practical purposes of rot resistance and the strength density necessary to support the great weight of a cribbed roof in Kiva L. We know nothing of the roofing elements employed in the two Aztec

134

Differential Wood Use at Salmon Ruin structures because the burned materials were not saved. Nevertheless, the importance given to the kiva ­positions is evident because they do not align within the overall symmetry of the building plan.

reused timbers already at Salmon—a considerable savings in labor and an apt response by groups not as tightly integrated as the original builders and occupants. After construction of the core units at Salmon and Aztec West had been planned and supervised by skilled Chacoan crews, work appears to have been turned over to local workers for the remaining construction. The differences in labor efforts and skill may reflect the labor demands on the ­local populace and the size of the available workforce, although we cannot be sure that even small, experienced work crews would have been adequate to carry out the Chacoan construction demands (Lekson 1984; Windes and McKenna 2001). It is the nature of the initial great house construction, however, that merits closer scrutiny regarding how the labor force was employed and how technical were its capabilities. The effort to obtain quantities of non-local wood implies the importance of certain species for construction, and we can assume that the subsequent finishing of the wood was also highly technical and labor intensive. Although we know little of the actual wood finishing treatment at Salmon, this behavior is well documented in other great house constructions (Windes and McKenna 2001). Observations of the Salmon wood by Shelley (2006), the tree-ring laboratory, and the authors suggest that similar detailed and laborious work was conducted during the initial Salmon construction, although its magnitude is unknown. It is more difficult to understand what happened to the structural wood at Salmon after its initial placement in the great house between the ad 1080s and 1095. It is clear that considerable scavenging for timbers was conducted sometime by or after about ad 1116, but the absence of treering dates and the lack of detailed room histories makes this process unclear. Recent work in Room 62 suggests that by ad 1118 some wood elements were being reused for repairs, and that these may have come from the initial Chacoan roofs, perhaps the nearby living rooms. A precedent for the short life of living rooms can be found at Pueblo Alto, where they were converted to other uses within a decade or two after their initial construction (Windes 1987a, 1987b), a period that overlaps

Discussion Although all great house construction in Chaco Canyon and at Aztec and Salmon involved massive tree harvesting to satisfy building needs, the magnitude of the timber harvests provides relative scales for the respective labor requirements and organizational capabilities. The construction at Pueblo Bonito in Chaco consumed 25,000 to 50,000 trees (Windes and McKenna 2001:123) and was the most demanding of the great house projects. Pueblo Bonito thus required harvesting efforts three to six times greater than at Salmon, but its construction was spread over many decades, whereas Salmon’s main harvesting took place in a mere three years. Pueblo del Arroyo in Chaco required between 6,300 and 18,800 trees for its construction (Windes et al. 1994) and had an equivalent number of rooms (approximately 239 rooms and three large kivas; Lekson 1984: Figure 66) as Salmon, but it is a smaller structure in terms of the basic plan. Although the estimates are approximate, Pueblo del Arroyo and Salmon required about the same numbers of trees. The higher numbers of elements estimated for Pueblo del Arroyo derive from the numerous door and ventilator lintels, which are not now fully exposed at Salmon. In general plan, Salmon is similar to Aztec West (and Hungo Pavi in Chaco) and probably required about the same number of trees, although Salmon’s harvests were dwarfed by the requirements needed for the three main Aztec great houses. It is not difficult to address the different strategies between the Chacoan and later San Juan construction episodes at Salmon: one was highly organized, and the other was not. The secondary construction of smaller rooms, perhaps taking place over a century or more, provided approximately 16 percent of the total Chacoan floor area and required about one-third to half of the trees needed for construction. Rather than organize for harvesting new wood, with the exception of the rebuilt Great Kiva, many of the secondary occupants 135

Thomas C. Windes and Eileen Bacha the initial ­construction at Salmon. At Salmon, too, many living rooms enjoyed only short habitation use before being converted to refuse areas or kivas (P. Reed 2006b, this volume). The period when building deterioration and roof scavenging became a major factor at Salmon coincides with the change in ceramic traditions (when McElmo-period pottery became prominent) and design layouts noted by Franklin (2006a), IrwinWilliams (this volume), L. Reed (this volume), and Washburn (this volume). At Salmon, many rooms were subdivided and made into smaller rooms or small kivas, a process also seen at other great houses (e.g., Pueblo Bonito, Chetro Ketl, and Aztec West), and one that appears to mark the presence of small habitation groups staking out claims within a larger architectural complex. At Aztec West, Pueblo Alto, and Pueblo Bonito, distinct architectural units, essentially small houses, were built in groups within the plaza or next to the great house (see Windes 1987a: Figures 10.12–10.15, 2003). Similar units existed at Salmon in the ad 1100s and 1200s. The late ad 1080s building effort at Salmon was extraordinary, but subsequent remodeling and construction were small-unit building efforts that were less formalized in the procurement and numbers of materials, in the potential size of the labor force, and in the more leisurely progress. Although there are no tree-ring dates from this period—which spans the ad 1100s and early ad 1200s during the McElmo phase, as it is known elsewhere—the extremely poor environmental conditions and the lack of organized labor probably forced scaledback efforts and adaptive measures to deal with the harsh circumstances. Large construction efforts were not attempted again at the site until the Great Kiva was rebuilt at about ad 1263, suggesting some organizational shift in the local society. Did the Great Kiva serve only residents of Salmon in the ad 1200s, or was it built for a poorly known, large, dispersed community? At Aztec, construction efforts were less intense, but the drive to maintain Chacoan standards is seen in the construction of Aztec East through the ad 1100s and into the 1200s, an effort not duplicated at Salmon. The lack of dates in the ad 1100s at Salmon may have resulted from a shift to local

procurement for trees that grew poorly during the ad 1100s drought and failed to date. There is little question, however, that Salmon in the ad 1100s represented a large resource center, full of useful cultural materials, including roofing. These materials could have been recycled by those living there, but who failed to maintain the standards imposed during the initial construction and subsequent use of the structure. There was habitation continuity at Salmon, but the use of space and the organization of the residents were very different in the ad 1100s than they were in the late ad 1000s—changes reminiscent of those observed for Pueblo Bonito during the same period (Windes 2003). Salmon is unmistakably a Chacoan great house. In the subtleties and details of its construction, we can see how the architects and builders closely adhered to the traditions, standards, and craftsmanship of the great houses built earlier in Chaco Canyon. Certainly, great house construction projects in the region would have benefited from the experienced labor pool that would have been available after the last massive construction efforts in Chaco in the early ad 1080s, shortly before Salmon was built. Unfortunately, our knowledge of local small-house settlement, whose residents might have supplied hands for the work at Salmon, is sparse. In contrast to the large contemporary settlement on the bluffs behind Aztec, Salmon appears to have lacked a similar large potential labor pool, which may have affected decisions regarding wood procurement and other construction strategies. Still, thousands of tree stems were gathered for the Salmon work, and we can only guess the impact of roving gangs of foresters as they cleared out trees in areas used by groups not invested in the same efforts that created Salmon. The area around Salmon must have suffered considerable devastation from the loss of thousands of junipers, a species that takes a long time to regenerate along the margins of conifer habitat. If fire was utilized as a tool to increase beam production (see Gruell 1985; Touchan et al. 1996; Weaver 1951), then concerns from neighboring groups may have impacted the social dynamics of the region. Clearly the size and nature of the small-house occupation in the region are germane to understanding the complicated tap-

136

Differential Wood Use at Salmon Ruin estry of ­human interactions in the middle San Juan region just prior to and during the settlement of Salmon. Before the final third-story rooms were added to Salmon by ad 1095, the construction of Aztec North and the planning for Aztec West may have ­incited rivalry and conflict between groups com­ peting for limited timber and other regional resources. The Aztec builders would have sought timbers closest to the construction; large stands were located mostly to the north-northwest, north, and northeast about 45 km away, along latitudes about 10 km south of Durango, in similar if not the same areas used by the Salmon crews. Potential conflict may have been resolved by the Aztec architects by selecting thousands of aspen latillas instead of ponderosa pine for the Aztec West roofs. Aspen would have been harvested from areas ­farther north than the potential conflict areas. The shift to greater use of local tree resources at Salmon after ad 1095 may have been a response to timber conflicts with Aztec resources, although ­builders at both Aztec and Salmon switched to mostly juniper elements after ad 1130. The potential for conflict over depleted timber resources may have affected the immediate areas around both Salmon and Aztec, as well as ­areas farther north along the lower margins of major conifer stands in the New Mexico/​Colorado border area to the northeast of both great houses. We must consider the extent of the potential forest resource areas along the Colorado border and slightly north and ask whether they could have supplied all the timbers for great houses in the region without serious depletion before we derive models of conflict. In any case, ­local resources around Salmon and Aztec would have suffered extensive depletion of woodlands, impacting any local, small households. For the most part, long, large-diameter, straight beams were not necessary in the construction of Salmon except for the primary, load-bearing roof beams in the larger rooms and kivas. Local mountain juniper was probably sufficient for construction needs at both Salmon and Aztec, although it is tougher to cut and prepare than other conifers and provides only a single roofing element per stem. With local trees suitable for construction needs,

the desire to import trees from long distances would have involved factors beyond mere practicality. The use of imported trees at Chaco Canyon was by necessity, but it probably was also imbued with symbolism that became embedded over time as one of the standards for great house construction. The logistics and effort to procure great numbers of non-local trees demanded organizational capabilities and human labor to engender production success. The timing of the harvests also would have had to fit with other crucial demands imposed by a horticultural life. The different procurement strategies for Salmon and Aztec West reveal variations on a Chacoan theme carried out by groups skilled in the nuances of great house planning and construction. We lack the detailed profiles that would link Salmon and Aztec directly to specific hallmark construction episodes in Chaco Canyon, but the overall patterns of construction reflected in the choices and craftsmanship of the wood resources mark the construction of the core units at Salmon and Aztec West as Chacoan. It is the choice of wood that most distinguishes the Salmon and Aztec constructions. Whereas both architects and builders could plan the use of local resources with few exceptions, the choices made by the planners and workers indicate different philosophies in procuring materials. As noted above, the impact of a potential small labor force at Salmon compared to larger labor resources at Aztec may have forced different procurement strategies. Possibly, the somewhat greater distances involved to obtain suitable wood for Salmon than for Aztec may have affected procurement strategies, although this problem did not deter procurement from at least 80 km away for building materials used in downtown Chaco. We know that much non-local timber was gathered for Salmon, but if the work force was small, they might have sought to reduce longdistance­demands. A large local settlement does not appear to have been present prior to the construction at Salmon, leaving local tree resources relatively unimpaired by previous harvesting. We argue that the different emphasis at Salmon and Aztec West for the procurement of non-local trees was invested in the leadership for designing and carrying out the actual great house constructions. At both Salmon

137

Thomas C. Windes and Eileen Bacha and Aztec West, the core units reflect knowledge and experience gained from work in Chaco Canyon of the proper standards for the use of structural wood, but its application differed between the two sites. After the core constructions, the greater use of local materials at both great houses suggests that ­local labor forces were increasingly involved in actual construction of the great houses, and that less strict standards were applied. Labor efforts associated with local and nonlocal­preferences appear to have been clearly differentiated between the Chacoan and later San Juan/ Mesa Verdean construction episodes at Salmon and between Aztec East and West. The later constructions employed juniper almost exclusively, and local procurement would have dramatically decreased the labor efforts needed during the initial construction. The predominance of non-local species in Aztec West is confirmation of the logistical capabilities and efforts sustained during the overall building enterprise, even though some local materials were procured after the core units were completed. The dichotomy between local and nonlocal efforts is less clear at Salmon, but it parallels the West Ruin in its adherence to extensive labor investments for the core room units. Afterward, as the wing roomblocks were built, labor investments tapered off with an increased use of local wood. Much of the evidence from Salmon suggests a progression of site use similar to that observed at Aztec (Brown et al., this volume). The initial construction was a massive effort requiring a high level of organization and probably a substantial ­labor pool that was mobilized for a relatively short ­period of time. Tree-ring dates for roofing and wall apertures suggest that this mobilization took place between ad 1088 and 1090, with a secondary effort between ad 1094 and 1095 to place a third story along the back room tiers to either side of the elevated kiva. Of course, it may have taken longer to lay out the general plan of the great house, set foundations, and to build the lower masonry walls of the first-story rooms at least to the level of roomwide platforms or level with the tops of the doorways before any wooden structural elements had to be embedded within the masonry. Finally, although it is possible that small groups

of residents from Salmon eventually migrated to found Aztec West (Lekson 1999:137; Stein and McKenna 1988), or contributed to the architectural changes to it after ad 1115 (Reed 2002:7), or later built Aztec East (Reed 2006d), the timing and wood construction philosophies employed at the two great houses offer evidence against it. If any of the three great houses had been planned and built by the same group, we would expect a much higher degree of similarity in construction. Importantly, the design and footings for Aztec West were probably being prepared less than five years after the third-story additions made to Salmon in ad 1095 or slightly later—little time to expect that the same planners and builders moved onto Aztec to implement construction of another great house. It is not clear, however, if those who designed and built the great houses were the same group that resided in or used them. The high degree of planning and craft specialization employed during the constructions suggests that those responsible for creating the structures may not have been the same people who used them.

Acknowledgments No research regarding Salmon can be conducted without giving special thanks to the efforts of the former Salmon field staff. Several chapters in the ­final report to the funding agencies provided lengthy, detailed information about wood use at Salmon and other relevant data (see K. Adams 1980b; R. Adams 1980; Shelley 2006) without which this report would not have been possible. It is our great honor to have followed in the groundbreaking footsteps of these earlier researchers. The senior author also wishes to thank several volunteers who helped document and sample the remaining wood in Salmon between 2002 and 2003. Jamie Schubert (CA), Margaret Kaiser (MD), ­Eileen ­Bacha (OH), and Gretchen Jordan (CO) all helped with many a wood project and brought a wealth of experience and good humor to the tasks. Nancy Espinosa of the Salmon Ruins staff was also a great help during our fieldwork. Thanks also to Beth Bagwell, who lent much of her field equipment for the wood sampling, and to Paul Reed, who organized the Salmon conference, invited our

138

Differential Wood Use at Salmon Ruin participation, and answered many an inquiry regarding the field data. Larry Baker, director of the Salmon Ruins Museum, gave wholehearted support of our fieldwork and has been a great friend and supporter of the project; our deepest thanks. Thanks to the University of Arizona Laboratory of Tree-Ring Research staff for discussions and assistance regarding the Salmon wood, particularly Rex Adams, Richard Warren, and Ron Towner. Finally, our appreciation extends to Beth Bagwell and Chris Millington of the University of New Mexico for graphics and computer assistance.

Notes 1. The results of the archaeomagnetic sample were too poor to derive a date, and the archaeomagnetic curve for this period is poorly known. Thus, we only know that the fire was protohistoric in time. Eric Blinman (pers. comm.) suggested a date of about ad 1500. 2. The discrepancy in the unexpected success of aboveaverage­numbers of samples with pith and outside rings may be explained by notes taken by field hand Debra

Autry on June 26, 1974. Her page 1 notes on Room 90 provide the noteworthy comment that “wood, carbonized or not, is not retained unless it is over 1" in diameter—preferably about 3" in diameter—and has been excellently preserved, with outside and inside rings intact, for dendrochronology and identification analysis.” 3. The McElmo and Mesa Verde phases have been widely used to denote the temporal periods following the Chacoan occupations (see Brown et al., Table 12.1, this volume; Cameron 2005). These phase names do not imply that peoples from the Northern San Juan reoccupied Salmon, although the changing structure of ­society throughout the occupations in the great houses deserves closer scrutiny. By the same token, a Bonito phase great house does not automatically indicate that it was built and occupied by Chacoans, although that is the focus of this chapter. 4. A more extensive article on the wood from Salmon was published in Thirty-Five Years of Archaeological Research at Salmon Ruins, New Mexico (2006), edited by Paul F. Reed and published by the Center for Desert Archaeology, Tucson, and Salmon Ruins Museum, Bloomfield, New Mexico.

139

8

Human Remains Recovered from the Tower Kiva at Salmon Ruins Nancy J. Akins

In 1973, excavators working in the east half of the Tower Kiva (Room 64W) at Salmon Ruins encountered what was described as the charred remains of 15–20 individuals (both sexes and ranging in age from under 2 years to mature). These remains were found just on top of or in the upper portion of the stratum, and it is assumed that these individuals were trapped on the roof at the time of the fire. There was very little other cultural material associated with this feature or with the rest of the stratum. (Whitten 1973)

blage included sherds from a number of corrugated and black-on-white vessels, matting, basketry, burden rings, a portion of a digging stick, ground stone pieces, a maul, and fragments of a lizard effigy.... The presence on the floor of a portion of a fresh pumpkin, fresh corn and beans still in their pods suggest that the kiva may have burned in the Fall (Room 64W 1975). This interpretation has been repeated several times in the limited literature on Salmon Ruin. In his volume on warfare, Steven LeBlanc describes the remains in the Tower Kiva:

Lower fill was removed from the west half of the structure the following year, revealing “the charred remains of 20–25 additional individuals, bringing the total to 35–45 for Feature 1. Mesa Verde pottery was found associated with the bodies, but little other cultural material” (Peterson 1974). A 1975 Pecos Conference paper summarized the two previous seasons of work as follows: The kiva was destroyed by an extensive fire during the secondary occupation and the charred remains of approximately 40 to 50 individuals (almost all children 0–5 yrs.) in the upper levels of the roof stratum indicate that some of the inhabitants were trapped on the roof when it burned.... Owing to the fire, preservation of artifacts on the floor was good and the assem-

At Salmon Ruin, one of the more tragic events of this period apparently occurred: Thirtythree children perished in a tower kiva, sometime after ad 1263 and before the end of the century. The best reconstruction of this catastrophe is that the children were on the roof of a tower kiva, which then burned, along with much of the site. It is not clear whether the youngsters were alive when the inferno began or were killed just before. Was this tower kiva considered a place of refuge during an attack, an attack that succeeded, and resulted in the site’s burning and the children’s deaths? (Le­ Blanc 1999:233) In this chapter, I present my own analysis of the burned human remains from the Tower Kiva at

140

Human Remains from the Tower Kiva Salmon—one that does not support the oftenrepeated­claim that dozens of children were killed in the fire that destroyed the structure. Nor does my work support LeBlanc’s warfare hypothesis. In contrast, I have found evidence that at least several of the individuals found in the kiva were intentionally cremated in an event separate from and after the fire that consumed the structure and destroyed Salmon Pueblo.

Prior Salmon Skeletal Studies The first analysis of Salmon’s human skeletal remains was undertaken by Jeff Shipman at the ­Human Identification Laboratory at the University of Arizona’s Arizona State Museum in 1975–1976 and 1979. It was reported originally in 1980 and reprinted in Shipman 2006. The remains were collected as 44 burials and 18 batches of bone, and Shipman reported that 15 of these batches, which he referred to as “skeletons,” had the remains of more than one individual. Shipman’s analysis suggested that no more than two adults (a male and a female) and 33 subadults, mostly infants and young children, were represented in the assemblage. They were described as extremely fragmented and commingled rather than actual burials. All degrees of burning were found, and Shipman felt the burning occurred when the bones were still fresh, indicated by warping and irregular transverse splitting (Shipman 2006). Shipman did not speculate about what happened at the Tower Kiva other than indicating he was told it was “an accidental situation”; he did find it “curious that the majority of the remains are those of young children” (2006:329). Shipman’s analysis attracted the attention of Christy and Jacqueline Turner, who examined the collection on June 6 and 7, 1992, and, in my opinion, predictably found evidence of cannibalism and violence in the remains (Turner et al. 1993). In a later publication summarizing their work on collections throughout the Southwest, the Turners again claimed that the Tower Kiva remains provide evidence of cannibalism and violence (1999:326–331). The Turners listed each “skeleton” separately, occasionally disagreeing with Shipman’s assessment of the age and adding observations on the presence

of abrasions, cuts, and perimortem breakage. They concluded that the two adults had been butchered and their remains scattered, whereas the children were “more or less spatially concentrated as specific individuals”; the presence of adult bones mixed with those of several children was viewed as an indication that the adult remains had been “scattered on the roof before the children burned up in the kiva fire” (Turner and Turner 1999:330, 331). The Turners reported finding perimortem breaks, cut marks, burning, and anvil abrasions, and the near absence of adult vertebra. Because later handling and the degree of burning prevented identification of pot polish, they were able to conclude that virtually all of their cannibal criteria were met. They interpreted the presence of caramelization (shiny, carbonized slag of heat-liquified tissue) as an indication that the children were burned in the flesh. The lack of carnivore damage was seen as an indication that body parts were covered with earth soon afterwards, possibly contributing to some of the mixing. Another important study of the Tower Kiva remains was conducted by Lisa Bergschneider (1996), who analyzed the collection for her master’s thesis at Eastern New Mexico University. Bergschneider’s study sought to test three possible explanations for the findings, including (1) an “intentional burning of living individuals during a time of conflict”; (2) “an accidental burning of the kiva”; or (3) cremations carried out “as part of normal mortuary practices” (1996:1). As the analysis progressed, she became convinced that most of the individuals did not die in the fire. Instead, Bergschneider believed the remains represented a cremation that was part of a secondary mortuary ritual that took place around the time the site was abandoned (1996:32), and that the bodies were in different states of preservation when placed on the kiva roof, indicating that some were exhumed and cremated (1996:55). She found evidence for at least 18 individuals, and probably more, based on the presence of 18 left (and 15 right) petrous portions (auditory meatus) of craniums. The next highest count was of femur and tibia portions (n = 11 individuals; 1996:35–36).

141

Nancy J. Akins

The Present Study My study of the Tower Kiva remains was developed as part of the Center for Desert Archaeology and Salmon Ruins Museum’s Salmon Research Initiative (Reed, Chapter 1, this volume). The purpose of this study was to evaluate previous hypotheses of butchery and cannibalism as well as Bergschneider’s cremation hypothesis. To accomplish this, I reanalyzed and reassessed the available data on the ages and numbers of individuals; the type, amount, and characteristics of burning on the bones; and the Turners’ reported cuts, abrasions, and perimortem breakage. To more accurately determine ages for the specimens and individual bones, the skeletal remains were compared to a number of precisely aged individuals from the Office of Archaeological ­Studies at the Museum of New Mexico in Santa Fe. Most of the individuals used for comparison came from excavations along the La Plata River: a male, a female, and children aged seven and eleven. Two younger individuals (a two-year-old and a six-yearold) came from a site in the Galisteo Basin. Dental aging charts were consulted when teeth were present. These charts are more reliable than relative size since tooth formation is not influenced by environmental conditions such as nutrition, which can retard skeletal growth (Larsen 1997:23–24). If teeth were not present, elements were assessed by comparison with individuals of known age. Although these comparisons are not as precise due to differences in growth trajectories, they provided a uniform set of standards for aging this collection. The initial goal was to record each piece of bone as accurately as possible using a coding format developed and refined over several years. This system includes (1) basic information for linking a bone to a provenience system; (2) a count of the pieces represented by that line of coding; (3) an assessment of how certain the identification of the pieces as human and the element identification are; (4) whether the bone(s) or pieces articulate or are from the same individual or bone; (5) Shipman’s original skeleton number; (6) identification of the element (e.g., femur, phalanx) and the side, and a detailed description of the portion of the element recovered (e.g., long bone: lateral or anterior

fragment of a proximal end,); (7) completeness of the element (complete to less than 10 percent); (8) general age (infant, young child, middle child, older child, young adult, middle adult, older adult, adult) and specific age; (9) criteria for aging (size, dentition, epiphyseal union, cortex compaction); (10) type and degree of environmental alteration (root etching, pitting); (11) type and location of animal alteration; (12) burn color, burn location, the presence and type of heat-induced crackling or exfoliation; (13) type of processing (cuts, abrasions, impact breaks) and processing location; and, finally, (14) comments. Unfortunately, recording each fragment using this system proved to be infeasible given the many thousands of small pieces of heat-shattered bones to be analyzed. After recording a single bag containing the postcranial remains from “skeleton” 34 with over 1,400 pieces of bone, it was clear that a more efficient approach was needed. Although the computer coding would have produced a large and detailed database, it would have required a great deal of time, and understanding the burning patterns from the coding would have proved difficult. Instead, I took a different approach, recording the elements and portions present and degree of burning on visual recording forms developed by Buikstra and Ubelaker (1994) for skeletons of infants, children, and adults. The extent of the element and the types of burning were coded on these forms, and copies of Shipman’s analysis forms were also annotated. As Shipman noted, most (35.7 percent by his count) of the specimens contained more than one individual. Previous analysts had separated the “skeletons” into different individuals; thus, part of my analysis included evaluating those divisions. All bone assigned a skeleton number and the unassigned bones from grid excavations were examined. Burial forms, grid excavation forms and photos, and photograph and slide documentation were consulted for the archaeological context and integrated into the findings. Although a good amount of effort has been invested in this study, much more could have been done—as is usually the case in archaeology. Earlier researchers had reconstructed some parts of elements, and this study would have benefited from more extensive efforts

142

Human Remains from the Tower Kiva to find additional matches between the specimens and material collected by grid. Nevertheless, given the time allotted, a considerable amount of data was collected.

Number of Individuals and Age Distribution None of the previously defined specimens was found to be a complete individual, although some were nearly complete, and parts could have deteriorated through intense burning, from poor preservation, and from excavation and handling. To avoid confusing the field-defined “skeletons” with separate individuals, I will refer to the burials or skeletons as “specimens” in the remainder of this chapter, using their field-assigned numbers. Two specimens (13 and 17) could not be relocated by any of the analysts. Individual 16 was also missing but was recorded in a photograph. Mixing between the specimens was noted by Shipman (2006), who matched broken pieces from Individuals 37 and 41 located in grids 7E 1N (at a depth of 2.8 m) and 8E 0N (at 3.36 m). He also noted that cranial fragments from Individual 9B were found in Individuals 8, 10, 11, and 12 (Shipman 2006). His forms also indicate matches between the lots of bone recovered from Grids 10E 3S and 10E 4S. In summary, it was rare to find a specimen that did not have bones from another individual, a matter further complicated by the considerable number of bones recovered from grids or quads. The original test trench bisecting the Tower Kiva encountered a large mass of bones from as many as nine individuals in 9E 2S and 9E 3S that was not separated into specimens. Tables 8.1 and 8.2 summarize data on the subadults in an attempt to arrive at a more accurate estimate of the number of individuals. Body part distributions were compared for each possible combination of individuals within the estimated age ranges, taking the amount and degree of burning into consideration. Several of the specimens and grid collections were so fragmentary that the parts could have come from several individuals. This process resulted in a count of about 21 subadults. Ages range from newborn to at least one 9- to 11- year-old child. Most (13) were between about 3 and 6 years of age, and only one was nearly complete (­Table

8.3). Poor representation is the rule, with over half the individuals less than a quarter complete. The two older children are among those with the fewest body parts. When the ages of the specimens and grid collections are plotted on a map of the structure (Figure 8.1), the age groups cluster somewhat. This is especially true on the west side, where very few bones comprise defined “skeletons.” The distribution of adult bones was more difficult to quantify. Table 8.4 lists the parts by specimen and grid. Duplication of postcranial elements suggests at least two males and two females. Identifying two individuals of each sex rests on a single duplication of elements for each and two different ages for females. The males are represented by two left proximal femur shafts, and the females by left scaphoids (carpals or hand bones). It is possible that one of the scaphoids is from a large juvenile, but the ages also suggest two females. In all likelihood, more adults are present, but extreme fragmentation from the burning and poor preservation, especially of the cranial remains, makes this difficult to determine. At least one of the mature individuals is older, based on a piece of the pubic symphysis and a lumbar vertebra with osteoarthritis (from Individual 34). An arthritic female sacrum (also from Individual 34) suggests that at least one older individual is a female but does not preclude the presence of another older person. Teeth in one of the mandible fragments (Individual 40), which is comparable to one of the type collection (La Plata area) females in size, have relatively light wear on the first molar, suggesting a younger individual. Well-worn teeth that could not be attributed to either sex also suggest an older individual. One of the males is quite large. Bones identified as male or female were widely scattered, with females having a wider distribution to the south and east (Figure 8.2). The age distribution for the Tower Kiva remains is substantially different from the pattern characteristic of most human groups, with high infant mortality followed by healthy young adolescence. Between 30 and 70 percent of the juveniles generally die before age 15 (Buikstra and Mielke 1985:​399). For the Tower Kiva, the proportion may be as high as 88 percent, assuming only four ­mature

143

Figure 8.1.  Distribution of subadult remains in the Tower Kiva at Salmon.

Figure 8.2.  Distribution of burning for subadult remains in the Tower Kiva at Salmon.

Age

0–6 m

12–18 m

@ 18 m

18 ± 6 m

18–24 m

@2y

2–3 y

@2

Individual

1

2

2?

3

4

4?

4?

4?

A&D

29

18

43

(A)

11A 9A

8A

Skeleton No. or Shipman designation

10E 5S

10E 4S

8E 2S

9E 5S

8-9E 5S

9E 3S TT

10E 3S 10E 3S

10E 3S

Grid calcined except scorched medial left clavicle

Burning

calcined black with exterior caramelized or black caramelized interior with an unburned core and exterior mottled brown-gray 3.50–3.60 cranial fragments, mandible frag- cranial black caramelized or burned inside and edges; mandible ments, teeth, humerus, C & T verts, rib fragments; possibly some fragment scorched inferior, black case fragments not assigned to an exterior, calcined interior; post­ cranial unburned or calcined individual 3.95 moderate cranial, teeth, C,T,L cremated calcined except 2 verts, scapulae, ribs, humerus, in- scorched lumbar bodies nominates, femurs, tibia, fibula 3.69 sparse cranial, frontal, parietal, unburned to almost calcined; ­occipital & mandible fragments some unburned exterior, scorched inside 3.45 sparse cranial, tooth C & L verts, unburned, scorched with black inrib terior, calcined with black interior, gray/brown with black core 3.5–3.6 temporal and sphenoid fragments, graded to cremated calcined teeth, long bone fragments graded scorched to calcined; 3.7–3.8 cranial: frontal and mandible ribs are black or calcine, almost a fragments, C vert, scapula, ribs, dry burn; cranial case fragments sternum, tibia, fibula, long bone scorched to light calcine; tibia and fragments fibula black

sparse cranial, C&T verts, sternum, clavicles, ribs, ulna, femur, fibula, talus, hand phalanx 3.50–3.53 tooth, clavicle, T verts, ribs 3.5 sparse cranial, parietal, temporal, occipital fragments

3.5–3.6

Elevation (m) Parts

Table 8.1.  Subadults Remains By Age from the Tower Kiva at Salmon

Dry? Temporal fragment is unburned and very eroded; could be child or mature.

Dry?

Vent shaft includes burned and unburned turkey bones.

Temporal is larger and may be from 9B; exterior is black but interior scorched and sooted.

Comment

@2y

@2 @2 @2

2.5–3.5 y

@2y

@2y

@3

5?

5? 5? 6

7

7?

7?

4,7,8?

@2 2y ± 8 m

@2

4?

4 or 5 5

Age

Individual

A&B

C

8C

34 32A

42B

37 & 41

Skeleton No. or Shipman designation

9E 3S

10E 3S

9E 3S TT

10E 3S

6-7E 1S 8E 1S 10E 2S

6E 0N

NW quad 7E 1N 8E 0N

9E 3S

Grid

Burning

Comment

3.70–3.81 cranial fragments, C vert fragment cranial: dark scorch, vert mottled brown/black 2.60–3.20 parietal along sagittal suture black black, most caramelized Sphenoid and clavicle fragments 2.8 sparse, cranial: left maxilla and match between 37 & 41; large pit 3.36 mandible; left clavicle, scapula, in deciduous maxillary canine. ribs, humerus, sternum, C,T,L S verts 3.36 sparse cranial, humerus and tibia mostly black, humerus graded fragments black to calcined 3.36 hand phalanx calcined 3.56 cranial base fragment; 2 teeth black and caramelized 3.5–3.6 temporals, scapula, clavicle, ribs, black with some caramelized or C,T verts, femur fragments calcined 3.5–3.6 moderate cranial scorched to calcined, heavily caramelized when burned black; black posterior, base, mandible maxilla, left side of frontal, scorched left and center frontal, calcined lower right frontal; teeth caramelized 3.50–3.60 sparse cranial -temporal, occipital black, some with brown portions; & mandible, ribs, C 1, vert frag- some cranial fragments burned ments inside; scorched outside may be from A mostly black, some caramelized 3.5–3.6 cranial, teeth, C2, C, T radius, or calcined; some vertebrae and ulna, clavicles, ribs, sternum, calcaneus graded ­femur, tibia, fibula, talus, ­calcaneus, metacarpal, phalanges 3.70–3.81 cranial case, C vert mottled brown/black

Elevation (m) Parts

Table 8.1. (cont’d)  Subadults Remains By Age from the Tower Kiva at Salmon

12

2.5–3.5

3–4

3–4

?

8

10

9?

9

11B

2+ y

?

10A

15

3–4

@3

4±1y

8B

35

@5

young child 3–4

27

4±1y

33

Age

Individual

Skeleton No. or Shipman designation

9E 3S TT

9-10E 3S

10E 3S

10E 3S

8E 0N

7E 1N

8E 1-2S

10E 3S

10E 3S

9E 1S

7E 0N

Grid sparse cranial, orbit and case ­fragments cranial: right mandibular m, cranial fragment, long bone fragment sparse cranial, teeth, C & L verts, ribs, sternum, humeri, innominates, tibia, calcaneus

mottled black to gray, black, sooted exterior, dry burns unburned

Burning

Comment

cranial: unburned, partial, black; teeth gray, others unburned; black to calcined; femur unburned and burned 3.5 cranial, maxilla, T & L verts, ribs, cranium – unburned to calcined, tibia, shaft fragments some caramelized; postcranial black to calcined 3.45 sparse cranial, case, maxilla, cremated calcined Active porotic hyperostosis. ­occipital, C vert, rib fragments calcined 3.36 sparse cranial, teeth, parietal, temporal, mandible, frontal, ­occipital, long bone end fragment, vert fragments 3.41–3.51 case fragment, R tibia shaft calcined Unburned animal bone. ­fragment, L ischium Some could be parts of 8C. 3.5 sparse cranial: mandible, possibly mandible and C vert unburned, possible case gray/brown with some of the 8C cranial, C & black interiors, T verts, rib, T verts, rib, prox. femur, hand ­phalanx calcined, femur graded phalanx scorch to black 3.5 sparse cranial: frontal, parietal, frontal black, others unburned and case, teeth black – 1 burned after breaking 3.65 very sparse cranial, L vert, cranial unburned; postcranial ­humerus, scapula, ribs, sternum black to calcined 3.5–3.6 mandible fragments and teeth unburned Large hypoplasia on R PM1.

3.5–3.53

3.2–3.5

3.49

Elevation (m) Parts

Table 8.1. (cont’d)  Subadults Remains By Age from the Tower Kiva at Salmon

4–5 ± 1 y

3–4 y 4±2y

11 or 12 12

13

3–4 y

11 or 12

4±2y

11

4± 2 y

3–4

10?

11 or 12

3±1y

10??

3–4 @3

3–5

10?

11 or 12 11 or 12

Age

Individual

7

38 25

20

19

24

36

A,B,C

Skeleton No. or Shipman designation

9E 2S

5E 2S 6E 2S

8E 4S

8E 5S

8E 3S 8E 4S

6E 1-2S

10E 1S

6E 0N

10E 3S

Grid

3.5

3.65 3.4

3.48

3.55

3.5–3.6 3.5–3.6

3.4

3.5–3.6?

2.89

3.5–3.6

sparse cranial, maxilla, occipital and case fragments, tooth partial mandible sparse cranial: parietal, maxillary teeth, frontal, sphenoid; T & L verts, scapulae, ribs, sternum, femur considerable cranial, humeri, scapula, ribs, vertebrae, pelvis, tibia

case, sphenoid, ribs, T verts sphenoid, tooth, T verts, rib, fragments sparse cranial: face, case, tooth, rib fragments, long bone

cranial: temporals, maxilla, ­occipital, base, C, T verts, ribs, metatarsals, phalanges sparse cranial: occipital, parietal, tooth cranial: parietal, temporals, ­occipital, base, sphenoid maxilla fragments, teeth, small cranial and long bone fragments

Elevation (m) Parts

Table 8.1. (cont’d)  Subadults Remains By Age from the Tower Kiva at Salmon

No similar age and burn in vicinity.

Comment

cranium: black right mandible fragments and left parietal, unburned or scorched posterior and right parietal, occipital edge burned – broken before or during burning; postcranial black except for unburned left ribs and scorched sternum and right rib 1

cranial and 1 rib unburned, rib mottled gray to black, long bone shaft fragment calcined unburned except for a spot of burn on the probably unerupted tooth unburned unburned

unburned unburned

maxilla unburned, others calcined Probably not all the same individual; hypo on maxillary canine.

black

black and caramelized

charred or calcined

Burning

@5

4–5± 1 y

4–5 ± 1 y

4–5 y 5–6 ± 1 y

4–6 4–6

4–5

13?

14

15

15? 15?

15? ?

13, 15, 16?

4–5 ± 1 y

5–6 y

13?

16

Age

Individual

5

C

10B B&D

14

1B

32B

Skeleton No. or Shipman designation

10E 2S

10E 3S

10E 5S 10E 5S

10E 3S 9E 3S TT

9-10E 3-4S

9E 1S

9E 2S

8E 1S

Grid

3.5

3.5–3.6

3.7–3.8 3.7–3.8

3.5 3.5–3.6

3.65

occipital fragment and tooth, ­fragments of most postcranial except right foot temporal and tooth mandible, humerus, fibula, ­possibly some of the cranial and long bone fragments teeth clavicle, ribs, tibia and humerus fragments cranial, teeth, T & L verts, ­metacarpal considerable cranial; T, L, S verts; humeri, radius, pelvis, hand or hands, femur, tibia, patella

sparse cranial, case, base, teeth, T verts, ribs

Burning

Comment

Two right humeri the same size. cranium: mostly black with mandible and posterior parietal calcined, part of one frontal piece and some interior surfaces less or unburned; postcranial: black to calcined, much caramelized, patella scorched

unburned, black, and calcined

calcined calcined

black mandible black inside, scorched outside, fibula black

dark brown to calcined, some caramelized

cranial: black and caramelized, calcined; vert body unburned, ribs unburned, smoked, and caramelized 3.20–3.50 cranial case fragments, teeth, long unburned and calcined bones, ribs 3.66 sparse cranial: parietal, occipital, unburned Occipital fragment may be older or mandible fragments even adult.

3.56

Elevation (m) Parts

Table 8.1. (cont’d)  Subadults Remains By Age from the Tower Kiva at Salmon

Age

3–4

5–6 y

5–6 ± 1 y

@5 4.5–5.5

5–6 ± 1 y

5–6

5–6

@5

Individual

16?

16?

17

17? 17?

18

18?

18?

18?

1C

9B

3B

C

6

Skeleton No. or Shipman designation

9E 2S

9E 1S

10E 3S

10E 2S

9E 2S 10E 2S

9E 2S

9E 0S

10E 2S

Grid

3.2–3.5

3.66

3.5

3.5

3.2–3.5 3.5–3.6

3.5

3.5–3.6

pelvis and femur fragment, and unidentifiable pieces cranial fragments, teeth

considerable cranial, teeth; scapula, C & T verts, proximal ribs, ulnae, radius, hands; humerus from 9A

rib and humerus fragments humerus, sacrum, ilium, femur, tibia, fibula, calcanei sparse cranial case, innominate, femur, tibia, fibula

cranial, temporals, femur, tibiae, talus fragments cranial, teeth, C, T, S verts, ilium fragment, rib fragments considerable cranial and thorax, ulna, radii, ilium, femur, fibulae, hand phalanges

Elevation (m) Parts

Table 8.1. (cont’d)  Subadults Remains By Age from the Tower Kiva at Salmon

cranial: unburned and calcined, teeth unburned

black and caramelized except for the left side of cranium, which is scorched and even unburned superior, and the left side of the C verts, left rib 1 and some of the right left scapula calcined black and caramelized except for feet that are calcined cranium: black to calcined with unburned cores and interiors; postcranial: mainly calcined but femur partial black, distal tibia and posterior black, interior light or unburned, fibula calcined, interior unburned cranium: mandible and parts of right side calcined, right parietal and left malar black, center frontal and parietal scorched to unburned; postcranial scorched or calcined mostly calcined

black, some caramelized or calcined calcined

Burning Shipman’s B.

Comment

5–6 @5

6–7 ± 1 y

7–9

7–9 6–8 6–8

4–8 5–6 ± 1 y 6 ± 2y

19? ?

20

20?

20? 20? 20?

20? 20? 20?

9–10

5±2y

19

20 or 21

Age

Individual

E

28 30 32B

20B 23

3A

2B

27B

4

Skeleton No. or Shipman designation

10E 3S

8E 2S 8E 1S 8E 1S

8E 4S 8E 4S 6E 5S

10E 2S

10-11E 2S

8E 1-2S 9E 1S

9E 2S

Grid considerable cranial, C, T, L, S verts, scapulae (2 right) ribs, pelvis, femora, fibulas

3.5–3.6

3.45 3.56 3.56

3.6–3.7 3.48 3.34

T & L verts, ilium, femur

left temporal teeth and rib fragment ribs, T vert body

femur fragment metatarsal 5 fragment maxilla fragment with teeth

3.45 left temporal 3.20–3.50 deciduous molar, long bone ­fragments (human?) 3.5 sparse cranial: frontal, parietal, temporal, base, maxilla 3.5 rib, T vert, femora

3.5

Elevation (m) Parts

Table 8.1. (cont’d)  Subadults Remains By Age from the Tower Kiva at Salmon

unburned unburned ribs: proximal charred exterior, interior unburned, scorched or sooted; vert unburned black

cranium mostly unburned, basilar black to calcined and small pieces of the parietals (this individual?); postcranial: C verts and femur and ischium scorched to calcined, some right ribs scorched and lateral T verts and one of the clavicles graded scorch to black; rest black with the left ribs and left pelvis caramelized unburned molar unburned, long bone scorched at end unburned except for right temporal fragment right femur unburned, rib smoked with unburned interior grading to burned and caramelized, the rest black and caramelized unburned unburned unburned

Burning

Hypoplasia lines on maxillary canine.

Matches 3A femur.

Hypoplasia lines on maxillary incisor and mandibular incisor.

Comment

9–11 ± 1 y

@9

8–10 9–10

8–9 7–11 10–12 @10

@10

21?

21? 21?

21? 21? 21? 21?

21?

Age

21

Individual

1C

B&E

34 21

31

Skeleton No. or Shipman designation

10E 4S

9E 3S TT 9E 2S 9E 1S Quad 1

6-7E 1S 8E 5S

9E 0N

8E 1S

Grid

3.5–3.6

metatarsal 1, long bone shaft ­fragments

One cranial case fragment, ilium fragment, long bone fragment C, T, S verts, innominate ­fragments, scapula, ribs 3.36 right proximal femur fragment 3.53 rib, femur fragment, small ­fragments 3.50–3.60 sparse cranial, mandible fragment, 3.50–3.60 scapula fragment 3.66 ilium and phalanx 3 of metatarsal 1 3.20–3.50 case fragments and L vert arch

3.53

Elevation (m) Parts

Table 8.1. (cont’d)  Subadults Remains By Age from the Tower Kiva at Salmon

metatarsal black; long bones cremated calcined

black to calcined unburned? calcined unburned

calcined unburned

case fragment calcined, ilium and long bone unburned calcined

Burning

Case fragments could be from a younger individual.

Comment

2y

2–3 y

3–4

3–5

6

7

8

9

18–24 m (18) (29)

4

2y

18 ± 6 m (43)

3

5

12–18 m (9A) (11A)

2

(27) (35)

(11B) (12)

(8C)

(37) (41) (42B) (34) (32A)

(8A)

0–6 m

1

8E 1-2S 7E 1N

10E 3S 10E 3S

10E 3S 10E 3S 9E 3S

10E 2S

9E 5S 8E 2S 10E 4S 10E 5S 9E 3S 7E 8E 0N 6E 0N 6-7E 1S 8E 1S

10E 3S 10E 3S 9E 3S 8-9E 5S

10E 3S

3.45 3.36

3.5–3.53 3.5

3.5–3.6 3.5–3.6 3.5–3.6

3.5–3.6

3.69 3.45 3.5–3.6 3.7–3.8 3.7–3.81 2.8 3.36 3.36 3.36 3.56

3.5 3.5–3.53 3.5–3.6 3.95

3.5–3.6

? ?

? ? ?

? ? ? ?

?

Excavation Comments

mostly black and caramelized, 37 single disarticulated cremated adolescent phalanx calcined 41 single disarticulated cremated mostly postcranial 42 single disarticulated cremated adult 34 single disarticulated cremated adult, with rock and adobe casts 32 single disarticulated cremated sparse head to knees mostly black and caramelized, some of left side and spine calcined 8 single cremated; beads, matting, basketry; moderate head to waist, head varies scorched to soil unburned to 3.6 sparse to toes ­caramelized to calcined; ­postcranial black, some ­caramelized to calcined 11 multiple cremated; partial disarticulated moderate head to ankles cranium unburned to calcined, some caramelized; 12 single cremated postcranial unburned, black, calcined moderate cranial, very sparse cremated calcined 27 single disarticulated cremated ribs, vertebrae, pelvis, leg 35 single disarticulated cremated adult

moderate cranial; sparse postcranial – more of left side

calcined

Burn Pattern

8 single cremated; beads, matting, basketry; soil unburned to 3.6 moderate cranium to elbows unburned to calcined 9 single cremated; partial cranium on right side 11 multiple cremated; partial disarticulated moderate head to ankle; no calcined 43 single disarticulated cremated; soil right arm unburned moderate cranial; sparse cranial unburned to calcined, 18 single disarticulated cremated partial postcranial to ankles some dry burned; postcranial cranium, just above the floor; burned graded to cremated 29 single disarticulated cremated; partial cranium calcined, some dry burned

sparse head to foot

Skeleton No., Grid, and Elevation (m) Representation

Age

Individual

Table 8.2.  Summary of Immature Remains Recovered from the Tower Kiva at Salmon

Age

3–4

3–4

4–5

4–5

4–5

4–6

5–6

Individual

10

11& 12

13

14

15

16

17

(6)

(5)

(14) (10B)

(1B)

(20) (19) (7)

(36) (24) (25) (28)

(8B) (10A) (15)

9-10E 3-4S 10E 3S 9E 3S 10E 5S 10E 2S 10E 2 S 9E 0S 9E 2S 9E2S 10E 2S

9E 1S

10E 3S 10E 3S 9-10E 3S 9E 3S 10E 3S 10E 1S 6E 0N 6E 1-2S 6E 2S 5E 2S 8E 3S 8E 4S 8E 4S 8E 5S 9E 2S

?

? ?

most of the cranium; ­considerable to ankles

moderate cranium through pelvis; sparse to ankles

moderate to ankles

3.65 3.5 3.48 3.7–3.8 3.5–3.6 3.5–3.6 ? 3.5 3.2–3.5 3.5–3.6

moderate crania to waist; sparse to mid tibia

Excavation Comments

5 single cremated, partial cranium

6 single cremated, side by side with B. 7, top cranium unburned top, of the head west scorched left side, black and caramelized; right side; mostly black and caramelized, some scorched upper left side, tarsals calcined

black to calcined, some caramelized

black on left side of cranium, 7 single cremated, beside 6, head west on right mandible, and postcra- left side, body extended east nial except for sternum and left ribs; some burned after breaking unburned 1 single cremated; head right side, torso extended graded scorched to calcined, 14 single cremated some caramelized 10 single cremated, partial cranium

24 single disarticulated cremated 25 single disarticulated cremated 20 single disarticulated cremated, ventral 19 single cremated, partial cranium facing down on sand

unburned to calcined, some 8 single cremated; beads, matting, basketry; graded and some dry burned soil unburned to 3.6 10 single cremated, partial cranium 15 single cremated, head straight, body on right side

Burn Pattern

unburned, unidentifiable two individuals – one ­moderate cranial, the other calcined fragments with 24 moderated cranial, sparse postcranial to left mid femur

sparse cranial

? ? ?

? ?

moderate cranial, sparse to waist, femur fragment

3.66

3.5 3.5 3.65 3.5–3.6 3.5–3.6 3.5–3.6 2.89 3.4 3.4 3.65 3.5–3.6 3.5–3.6 3.48 3.55 3.5

Skeleton No., Grid, and Elevation (m) Representation

Table 8.2. (cont’d)  Summary of Immature Remains Recovered from the Tower Kiva at Salmon

5–6

5±2

7–9

9-11

18

19

20

21

(34)

(1C)

(31) (21)

(20B) (28) (30) (32B) (23)

(2B) (3A)

(4) (27B)

(3B) (9B) (1C)

8E 1S 8E 5S 9E 0N 9E 1S 9E 2S 9E 3STT 10E 4S 6-7E 1S Quad 1

10-11E 2S 10E 2S 10E 3S 8E 4S 8E 4S 8E 3S 8E 1S 8E 1S 6E 5S

9E 2S 8E 1-2S

10E 2S 10E 3S 9E 1S 9E 2S

2 single cremated 3 single cremated 20 single disarticulated cremated, ventral 28 single disarticulated cremated, bone not femur fragments from 3A recoverable and 8E 4S match 30 single disarticulated cremated 32 single disarticulated cremated 23 single disarticulated cremated partial cranium, burned soil with roof casts at 3.37 28 single disarticulated cremated, burned soil sparse cranial to proximal calcined, black, or unburned 31 single disarticulated cremated femur, metatarsal, phalanx l 21 single disarticulated cremated 1 single cremated; head right side, torso extended 34 single disarticulated cremated adult, with rock and adobe casts

? ? ? ? ? ? ?

3.53 3.53 ? 3.66 3.5–3.6? 3.5–3/6 3.5–3.6 3.6 3.2–3.5

spotty cranial and post­ cranial

? ? ? ? ? ? ? ?

3 single cremated 9 single cremated; partial cranium on right side 1 single cremated; head right side, torso extended 4 single cremated, on the right side, head east, torso extended 27 single disarticulated cremated

3.5 3.5 3.5–3.6 3.6–3.7 8E 4S 3.45 3.36 3.56 3.34?

cranium largely unburned, upper torso scorched with spots of black burns, left ribs and pelvis caramelized; calcined cervical vertebrae and parts of right femur and pubis unburned, some charred, some sooted possibly dry burns

largely calcined with spots of black burns and upper cranium unburned to black

Excavation Comments

much of cranium and torso, few long bones to mid fibula

?

partial cranium; moderate to mid tibia

Burn Pattern

3.5 3.45

3.5 3.5 3.66 3.2–3.5

Skeleton No., Grid, and Elevation (m) Representation

Note: ? = association less certain.

Age

Individual

Table 8.2. (cont’d)  Summary of Immature Remains Recovered from the Tower Kiva at Salmon

Nancy J. Akins Table 8.3.  Estimated Age and Completeness of Subadult Remains from the Tower Kiva at Salmon Age (years) 75%

25–75%

1

1 2 5

1

8